Full text of "Exploring the Internet, A Technical Travelogue"

See other formats

oe:

CARL MALAMUD

EXPLORING
THE INTERNET

—=— A Technical Travelogue —=
CARL MALAMUD

Exploring the Internet is a travel book in the
classic sense. In 3 trips around the world in
6 months, Carl Malamud undertakes a voy-
age of discovery, a look at the Internet and
the emerging global village. In 21 countries
and 56 cities, here are just a few of the
stops in this technical travelogue:

i TIE TIS 1 i ABTS =

¢ In Tokyo, Malamud takes a behind-the-
scenes look at the world’s most expensive
building, the Tokyo City Hall, and intro-
duces you to Jun Murai, the Internet
Samurai, one of Japan’s leading
researchers and the architect of the coun-
try’s national network.

¢ In Singapore, he digs into the govern-
ment’s highly ambitious plans to make an
Intelligent Island, a digital port for the net-
work highways.

¢ In Silicon Valley, he meets the man with
the bike, a digital nomad who pedals
around the country on a 500-pound bicycle
equipped with a half-dozen computers and
a head mouse.

¢ In Geneva, Malamud wages battle with the
international standards cartel in an effort to
make standards documents more widely
available.

Designed to give professionals in the com-
puter and communications fields an in-depth
look at worldwide networking, this intriguing
narrative helps you gain a unique technical
perspective on the Internet and how it is
built.

About the Author

Carl Malamud is the author of seven profes-
sional reference books including STACKS
and a three-volume series that includes
Analyzing DECnet/OS! Phase V, and
Analyzing Sun Networks. He has written for
publications ranging from Communications
Week to the Bangkok Post.

(Continued on back flap)

Digitized by the Internet Archive
in 2022 with funding from
Public. Resource.Org

https://archive.org/details/exploringinterneOOcar!

Exploring the Internet

A Technical Travelogue

ers ere}
ee 4 <7) ths : re |
se : JA /

WN - i are a i y 7 ‘ : . 7 : a ss) To : Cn
; i art ro t q
| Ae a wkd ‘ae
i] ale wt

a a : —_
‘,
ae

siieainiieslbalialiatiniaiieal ' F alae :
| x "2
sigoloverl \ecietoat A |

5 ‘ s {
i

Exploring the Internet

A Technical Travelogue

Carl Malamud

PTR Prentice Hall
Englewood Cliffs, New Jersey 07632

Library of Congress Cataloging-in-Publication Data
Malamud, Carl, 1959-
Exploring the Internet: a technical travelogue/Carl
Malamud.
cm.

Includes bibliographical references and index.
ISBN 0-13-296898-3 : $26.95

1. Internet (Computer network) I. Title.
TK5105.875.157M36 1993
384.3—dc20 92-16450

CIP

Acquisitions editor: Mary Franz
Jacket design: Bruce Kenselaar
Cover design: Joe DiDomenico
Prepress buyer: Mary E. McCartney
Manufacturing buyer: Susan Brunke

Published by Prentice-Hall, Inc.
A Simon & Schuster Company
Englewood Cliffs, New Jersey 07632

The publisher offers discounts on this book
when ordered in bulk quantities. For more in-
formation, write: Special Sales/ Professional
Marketing, Professional & Technical Reference

Division, Prentice Hall, Englewood Cliffs, NJ ii
07632. 2 ISBN 0-13-29b858-3

All products are the trademarks of their respec- | 90000
tive manufacturers. Carl Malamud is a trade-
mark of Carl Malamud.

be reproduced, in any form or by any means,

without permission in writing from the publisher. 9°78013296898

Printed in the United States of America
iy GP TS 7h Te Sy eh sh pe

ESB Ni Bebe rel bore one

Prentice-Hall International (UK) Limited, London
Prentice-Hall of Australia Pty. Limited, Sydney
Prentice-Hall Canada Inc., Toronto

Prentice-Hall Hispanoamericana, S.A., Mexico
Prentice-Hall of India Private Limited, New Delhi
Prentice Hall of Japan, Inc., Tokyo

Simon & Schuster Asia Pte. Ltd., Singapore

Editora Prentice-Hall do Brasil, Ltda., Rio de Janeiro

Contents

Foreword and Preface, in which we secure funding for a voyage which will
take us three times around the world in six months and give new meaning to
the phrase “many fine lunches and dinners.”

Prologue: The Document Liberation Front, in which we delve into the
abyss of international politics and wade through hordes of Jehovah’s
Witnesses to examine a Bulgarian FEP.

Geneva ¢ Prague e Zurich °¢ Geneva e Paris ¢ Boulder

Round 1: From INTEROP to IETF, in which we meet the Internet Samurai
and the Uncle of the NSFNet and go to the zoo to witness the birth of
EBONE.

San Jose ¢ Honolulu. ¢ Tokyo e¢ Fujisawa ¢ Akihabara

Hong Kong e¢ Macau e Singapore ¢ Dublin) ¢ Amsterdam
London e¢ Tampere ¢ Paris ¢ Geneva ¢ Nice ¢ Geneva
Ithaca © New York ¢ Washington, D.C. ¢ Santa Fe e¢ Boulder

Round 2: From Christmas to Cleveland, in which we encounter a massively
parallel bicycle in Mt. View, find magic boxes in Oz, and go swimming in the
Sea of Acronyms in Europe.

Berkeley ¢ Mt. View @ San Francisco e¢ Moffet Field

Wellington ¢ Dunedin e¢ Auckland ¢ Melbourne ¢ Sydney
Canberra e¢ Adelaide ¢ Singapore e¢ KualaLumpur e Bangkok
Amsterdam e¢ Utrecht ¢ Bonn ©e Brussels @ Paris
Washington, D.C. ¢ Cleveland ¢ Chicago e¢ Boulder

Round 3: In Search of a Standards Haven, in which we meet the world’s
most intelligent building, try to convince people to do the bloody obvious,
and learn the origins of the Internet.

Marina del Rey ¢ San Francisco ¢ Tokyo e Seoul
Taipei ¢ Hong Kong ¢ Bombay e Madison e¢ Boulder

Index

™~ es ie 7

ne ther o ew:
‘Soke ye

0 > an we 7
i ite ™ ik
ee

» peal

ois meena putea enttiindiiie |
. . WH Lips 7 Sie! a a :
i ’ an NYT raat i Tv ales - ee
~. : , { } f 7 7 *
rie *, se vas

Hiw ditto aREVONG tot grits it jetta ‘ow Panes ni
oO) gninkom Woe svig int atiniann vie wv. Bihsowe gielt be
‘ardently: Gis aodortul ani y

SF a
a : -

sloth WlSb-genik ithw et jn’ suveredth SesenuacKl &
oa ais! Ve oot frail “T RAN in? oie , bine eoiliog ot
| Cee lve gid) e-store

wan
ued. Be hems) .« AoOHaA * augers er: 2¥E
fe

Pas =
= ”

wend taetciodl aab een, oer Asi ni APA at BO CSV
~ ; < Syren J4o) S. i af’ Oo! < OF ee 38 wien ni

clea "i 4 é a of i = k Hy AUT + uiulonoH es
ison leinterrath 4 twa. SQgeyale + veieM 7
BSASe) = = iA y e'Pitiaicl arie4 eg SI meh 5.
“4 <

Woijoh * oa Alre?-* 2)O 9 aninenW . Pia
ae tagle-ryy E teh) ve Sy Sate ae Bin. a a a ra q
a) wa) 2; ee oe : a

, — ) ny
® K ¢ ) Po |
- s ry chan

OUEEieith & nee "Wait MeN?

cual © lofty oo Abie * > pibsnu oe
“Rodafsh » wey! slew © mogsgale « sibisisbA | a
bt abeeuig not JSiostU. * meby

[Dh Qe ds a cowipid« basleveld « o reate

» 4
ra

baw Mi! jagin ay) eihiad pont chrveboanei2 an — ;
jp AMIR O YBOGid ith, qb) av ent iL

, ; ee uf ( 3
an yy . nu ae
ras , 7 » i‘ sy fe

ae 4 { isi « rt =
& in ; mia > omer 4 ke “ Ne vs ei we

Foreword

In July 1991, Carl Malamud flew out from Boulder, Colorado, a
place he describes as “closer to Kansas than I’d like.” He had just
finished writing STACKS, which we were distributing as The IN-
TEROP Book to conference attendees, and wanted to present us his
latest proposal.

He met me at breakfast and presented me with a pith helmet
plastered over with Interop logos. “What Interop Company really
needs,” he explained, “is an Official Internet Explorer.”

His proposal was to fly three times around the world, visit as
many sites as possible, and write what he was calling “a technical
travelogue.” The catch, of course, is he wanted me to foot the bill
for travel expenses.

Exploring the Internet is the result of this odyssey. This book
demonstrates what many of us have long felt: the worldwide net-
work is here. Interoperability is not some imaginary goal at vendor
briefings, but a concrete part of networks all over the world.

This book is more than just a series of case studies—it is truly a
technical travelogue. As we read about the worldwide spread of the
Internet, we get to appreciate its diversity and its usefulness to mil-
lions. This is not some experimental research environment, some
academic toy, it is a real tool used by real people.

At that breakfast in 1991, Carl confided his hidden agenda. Af-
ter going on about “global infrastructures” and other nice phrases
designed to impress my senior managers, he turned to me and gave
me the executive overview. “Actually, this is a very simple project,”
he explained. “Buy my airplane tickets and I'll try to get into as
much trouble as I can. Then I'll write a book.”

Here it is.

Dan Lynch
Founder and President
Interop Company

Vil

a pbswist aabbuod due ous
fetsi wrt oH oa b'l sugttt sian ot sop
pil: oAT “es: Reidtutintel sow ay sloiviw MONS
aint “i taser t) byltiw! brie anaes ener? ot

Horsi] ry dit, A; itive oe hetnieern; bre sascha 4 36 mt
ilnes pteqnieD coral jaf wogal qorotel Aliw 3 iy
‘anvolgxttoasant tetiiO mms ei* Bb site na
di tiely Dhow om Dagorg eoenit sori wi ot aew Ip a wr " a -
tusicirioet 8° woille daw of detiw whew bins ldiazog eB ia. ¥
id arlt jag). od arty btn Sdial wean Yo ens oT
4) Ee ea he _renage ie
Adod ettT. wapeybo ei ig Hiravs wt at vorestend any
ton sbivehiew ort. tie} pabh-wasd ep Yo reset taskwn ee 20.
vabrey is baie eaphigee! orthoe Dewi xi Cilkdiseqoisial ”. 7
Lihnow af? v0 Hs edzowien to hag MIIIAG a hed §
8 eu at it—2olhnde sess 16 Rete 6 tend aed svar. a
gat iat ¥ RIT fe abi aby hic ei eid Ipods feet ow ah. oar polovedl tg
‘Sint Of eénpi oes aif bas ‘tieio vib: ati Sshetqqs al + a
Sree. Jnanenivng dorseest Istasmizsqus smo Jon am
atqaood Lesx yd beet loot [sat s ef iY
4A \shboage debbict aid bobfires dos Teer ai ecole
diane soln vllo bre “séurionientbal lacolg™ iwode #
ores bas ato? bam! orl ategenaits reinsert easton
“Jouponcy oleprie yey sel ere wilawntaa meni
ar Ghat dos GF wa il bie pistol ITs
| “dood & ahty

ae, | a

1
VY

OP an

ey oon ,

er ee

Preface

The reason I visited 21 countries in six months, embarking on a
technical voyage of discovery, can be traced back to an encounter I
had with a platoon of bureaucrats. In June, 1991, I struck a deal
with the Secretary-General of the International Telecommunication
Union (ITU), coordinators of the very formal process which eventu-
ally results in standards such as the Blue Book: 19,000 pages of in-
ternational recommendations which define the operation of data
networks, telephone systems, and other aspects of communication.

The deal was that the ITU would give me a copy of the stand-
ards in the antiquated online format they used for text processing
and that I would convert the standards into something the rest of us
could deal with and publish the standards on the global Internet.
Instead of charging obscene amounts of money for this information,
the standards would be available to anybody at no charge.

Organizations, rhetoric notwithstanding, don’t work for such
lofty goals as the dissemination of knowledge. The real reason for
this breakthrough in standards distribution was that the ITU
couldn’t figure out how to convert their own data from the proprie-
tary box they had built around themselves. My offer to do the con-
version seemed an easy way out of a job they had estimated at U.S.
$3.2 million. In return for converting the data and giving them a
copy, I could publish on this Internet thing, this academic toy.

The ITU gave me half the data they promised and conveniently
lost half the documentation to their Byzantine (both in age and in
complexity) formatting system. Then, after a mere 90 days, the ITU
abruptly cancelled distribution on the Internet. The academic toy, to
the bureaucracy’s horror, turned out to have over seven million peo-
ple. They were shocked to see hundreds of thousands of ITU docu-
ments being accessed by thousands of people in dozens of
countries.

Despite their abrupt cancellation of the project, the ITU still de-
manded a report. They wanted a certified bureaucracy document

Exploring the Internet

they could input into their process, analyzing possible scenarios and
deliberating impacts.

This is that report.

At first glance, this report looks like it would get the bureau-
cratic seal of approval. It is thick, for example. However, this book
is something different than the bureaucrats might expect. My re-
search methods were unorthodox: I spent most of my time talking
to people.

In fact, I talked to lots of people. In three trips around the
world, I talked to people on the front lines: people who make com-
puters, people who create networks, and people who use them. To
explain to the bureaucracies facts that are painfully obvious to the
rest of us, I had to leave the confines of meeting rooms and fancy
hotels and go into the research laboratories, network operating cen-
ters, and bars and restaurants where real work happens.

This report is thus a travel book, a book of exploration in the
tradition of classic writers ranging from Captain Sir Richard Francis
Burton, author of over 50 books and master of 29 languages, to Dr.
Hunter S. Thompson, author of careful documentaries about politi-
cal campaigns and police conventions.

Instead of exploring the Nile or Las Vegas, I looked at the
emerging Global Village. This isn’t a comprehensive atlas, nor is ita
definitive history. Exploring the Internet really is a technical trave-
logue, a narrative description of the people and networks I encoun-
tered during my travels. In my visits, | saw something the ITU and
the rest of the standards bureaucracy seem to have missed. The In-
ternet is here and it is not an academic toy.

While I saw many people, I should add that even in three cir-
cumnavigations, I couldn’t see everybody. This is one walk through
the forest, and there are many trails one can take. This book is a
selective look at some of the people, laboratories, and institutions
that help illustrate the diversity and the scope of the Internet.

This voyage would not have been possible without the support
of Interop Company which helped finance my travel expenses. Be
aware, though, that this book is mine, not theirs. There are many
opinions which are not shared by Interop Company and it is a great

Preface

credit to Interop Company that they felt it important to finance this
kind of work.

This book would not have been possible without the help of the
hundreds of people I talked to and particularly those who opened
their laboratories, homes, and CPUs to me. The hospitality I en-
countered was truly gratifying and I am thankful for the time given
to me by the people you will read about throughout this book.

This book owes a great deal to logistics, and I want to thank the
Boulder office of United Airlines for their infinite patience. Infinite
patience is an understatement when it comes to David Brandin,
Stephanie Faul, Mary Franz, Ole Jacobsen, Martin Lucas, and Yvette
Ozer, the editors who worked on this book, and it is to them that I
give the biggest thanks.

Carl Malamud
Carl@Malamud.Com

on sui “ie he

oe acatine wins Qc «eh

ors i ate * ‘on Py verorklive ‘Adie ee ae jon bier we ak

Seiial ad ee eRGeS Wralvaiined has.onbediet. 1,2 |
a ‘ Cilstiqer (iol, opt ot aI Aaa gota} » ee
eG 3 sant. mts cae: bats rvth Tee i Pox aside
atovvd sist qoslgymAtaueds lysax, ripen ies

exit. aby: ety) 03 inkw | bab won agol if isah Incig 8 eowo aox
atiriinl @raibe sila aid fi ile tol eptisii ls belie! lo

my. iv. ahd Weds eta ARS
Mote 839) Span bi har. OF aly pO i r art ad Areaye aneleiol uy a sf
gHovt bin ago a is Ne 3% ste: ved] at) met | ylae

% » - “ ‘
+, (URDEE GF Gr a bes 400% ait | 9 baikic we oily BOF
|. é Yo epeodtin a tema
Ms!
1 4 * i
cy
5
dt
- apt
}
bi phe!
4 i 74 17%
* ‘
f - ‘ @:
4,
{
she ha - i? *
r <9," OP
= +. ad
f 52 ‘- he a a J ag
:
1
eK id anh wit
4 & é J i; Oey
. or

FF: ene 4 ey ead hd tint envvnty
se), 3 ie eo CeeHhy Ths we comms Wi
ihe! tres 4 GA hee, Onmety) eee ee emt. oy
' | | Pie ct abel santa, isheotaheninels
) ee mp meee
ae i a koe pest promi Pe
= ae . itt ier? i voy pe Nase by ey
re ae oe st eka, 5 be oe pie

Prologue

The Document Liberation Front

eee 4! =

a =
er | as
t = |
aii
+ =
t
’ ‘
= v
ve ts

te + , Mis ie a De
eat aa ee mae

signi ee

aie to +.

Ynovl noitmadil inemuned oft

q
=}

Geneva

It was August 1991, and I was sitting in a café in Geneva, Switzer-
land, having lunch with Tony Rutkowski, a man who was a senior
official in the International Telecommunication Union (ITU) but
whom nobody could seem to pin an exact title on. While many
called him the General Counsel of the ITU (for reasons that will
become apparent), Tony prefered to refer to himself as “Counsellor
to the Secretary-General,” a title with enough ambiguity to serve his
purposes.

Tony and I were celebrating a minor palace coup, a year in the
making and the result of a bizarre partnership. How all this came
to be can be traced to my propensity to flame.

For several years, I’ve been complaining about the high cost of
international standards documents published by the ITU and other
international standards groups. After all, I write professional refer-
ence books for a living and these documents are my raw materials.

At first the flames started as simple whining, but soon they
grew to full-fledged diatribes in publications running the gamut of
technical sophistication from Data Communications to ConneXions. I
got lots of quotes from illustrious people like Jon Postel, Editor of
the Internet Request for Comments (RFCs), who explained that
keeping international standards away from “random people” was
hurting their acceptance. The gestation time for software based on
Open Systems Interconnection (OSI) serves as ample illustration of
Postel’s point.

The fabled power of the press produced an effect about as pro-
nounced as, say, an editorial in favor of truth and beauty. An occa-
sional nod from a passing reader, but no cries of outrage.

As a sanity check, I posted a note on the Internet asking if any-
body else thought that the inaccessibility of international standards
was hindering technical progress. This hit a nerve. In fact, the re-
sulting firestorm surprised even survivors of historically notable
dialogues such as “should pornography be distributed on the net?”

Exploring the Internet

The Internet has a wide community of users, but a convenient
taxonomy splits them into flamers and firefighters. Flaming is easy
to explain. You send a message and bingo, you’ve made 1,000 peo-
ple all hit the delete key in their mail readers. The power to make
so many people twitch is seductive and some people occasionally
succumb. In fact, some succumb all too often.

The firefighters have a different motivation. Their happiness is
inversely proportional to the size of their electronic inbox. Not eve-
rybody was terribly happy when their inboxes started filling with
“me too,” “T agree,” and “another minor yet trivial point” variants
on the first wave of reaction.

Dan Lynch, the founder of Interop Company, is a firefighter. He
didn’t want to see this situation continue forever, so he sent me a
copy of a memo written by Tony Rutkowski at the ITU. Turns out
Tony was doing the same sort of complaining I was, but was direct-
ing his comments internally to the huge ITU bureaucracy.

Tony and I exchanged a few e-mail messages and found we
were in violent agreement: standards have to be widely available or
the standards are irrelevant. We started trying to figure out what
could be done about it.

Turns out that we agreed on another point. We both feel that
the ITU has a fairly dubious basis for asserting copyright protection
on standards, since a standard spends months (often years) wander-
ing around the public domain as a working-group document before
it ever becomes blessed as an official standard and is “published.”

You can’t make a speech to a crowded room and then, one week
later, tell people that your speech contained valuable trade secrets.
You can’t publish a newspaper and then unpublish it. Once you
give something away, it stays given away.

The lack of valid copyright protection is one argument for mak-
ing standards more widely available, but there is a much more im-
portant one. If you believe, like Tony and I fervently did, that the
ITU does valuable work, then the standards are an essential public
good. Standards are laws and laws must be known to be observed.

Failing to make the standards widely available was, in the long
run, going to make the work of the ITU irrelevant. Other standards-
making bodies were promulgating standards, and despite its official

Geneva

United Nations treaty status, the ITU had to compete with other
groups for the attention of implementors and users.

I sent Tony a note and suggested that since we both agreed that
the ITU really had an indeterminate legal basis for asserting copy-
right, I would simply take all 19,000 pages of the Blue Book, scan it,
run it through OCR software, and post it on the Internet for distri-
bution by anonymous FIP, a service that allows a site to give the
public access to file archives using the File Transfer Protocol to any
user who logs in as “anonymous.” (The Blue Book is the massive
set of 1988 standards developed by the ITU’s Consultative Commit-
tee on International Telephone and Telegraph to govern the opera-
tion of everything from telephone signalling systems to packet data
networks to ISDN to high-speed modems.)

Now, there are a couple of points worth keeping in mind. Tony
was a senior lawyer for a powerful United Nations group that made
lots of money selling these documents. While Tony certainly sym-
pathized with my goals, I wasn’t quite sure how he was going to
react to this form of standards terrorism. Putting a lawyer on notice
that you plan to relabel his corporate assets with a $0 price tag is
kind of like putting Honda stickers on the motorcycles parked out-
side a Harley bar.

The other point is that the Blue Book is over 19,000 pages long.
Scanners and OCR are great technology, but it would take an aw-
fully long time to carry out such a project on my little home PC. In
fact, in a world of finite resources, you might even call my plan an
idle threat.

Idle threat or no, this message caught the ITU’s attention. One
year later, in June of 1991, after endless messages back and forth to
Tony, I had received a fax from the Secretary-General asking me to
post all ITU standards on the Internet on an “experimental” basis. |
promptly booked a plane ticket to Geneva for August to pry the
data out of the cold, clammy hands of the bureaucracy. Next thing I
knew, I was in Geneva, ready to go to work.

Exploring the Internet

I was sitting in Tony’s office reading a book while he went off to
clear political barriers. In the other room, Julie Butterfield, his assis-
tant, was busy typing away when the phone rang. Julie raced in
with a pad of paper and answered the phone.

“Mr. Rutkowski’s office.” She scribbled some notes on a pad.

“T’ll give him the message,” Julie said, hanging up the phone.

Something wasn’t right with this picture. As she walked out, I
inquired why she didn’t just have the phone roll over to her desk if
Tony doesn’t pick it up.

“We have nothing that sophisticated here at the ITU,” she re-
sponded with a smile and a strong dose of sarcasm.

A few minutes later, Tony came back and we started talking
while Julie ran some errands. Her phone rang and Tony sprinted
out to the outer office to take the call.

The ITU was using a very old donated Siemens PBX. This PBX
was actually quite advanced for its time, but its time had long
passed. This was a feature-rich PBX, but most of the features were
so difficult to use that people didn’t use them. Tony brought in his
AT&T phone from home just so he could have a few stored num-
bers.

Until recently, in fact, the entire ITU had one fax machine for
over 900 employees. A single fax machine was more the result of
bureaucratic desire to limit outside communications than a cost fac-
tor. The fax machine was in the Secretary-General’s office. This
meant that sending a fax could require approval right up to the Sec-
retary-General. To get the Secretary-General’s approval, you had to
go through the Deputy Secretary-General, which would require re-
cursive approval right on down the food chain. Needless to say, to
communicate through this medium you would have to be pretty
strongly motivated.

What is really funny about all this is that ITU, as part of its ena-
bling treaty, has the right to free telecommunications throughout the
world. This perk is meant to allow easy communication to the 160
member countries, and even goes so far as to allow the ITU to set
up terminal clusters at off-site meetings with X.25-based links back
to Geneva.

Geneva

I’m not quite sure what I expected at the ITU, but I kind of fig-
ured a modicum of computer literacy. Nothing fancy, mind you,
but at least things like basic phone tricks, fax machines, and maybe
even electronic mail. After all, this is the home of X.400, the mes-
sage handling system to end all message handling systems. Was the
ITU using X.400, though? Not a chance! They had Digital’s proprie-
tary Message Router and the infamous All-in-(N)One as a user in-
terface. X.400 connectivity was achieved by a poorly functioning
gateway.

This lack of computer literacy at the ITU was what made Tony
Rutkowski such an odd duck. When he came to Geneva he saw
that none of the international bureaucracies had Internet hookups.
So, Tony did what any self respecting technocrat would do and
called up the nearest Internet provider, in this case the CERN phys-
ics laboratory, and asked for a guest account. Several years later,
Tony was still the only ITU employee with access to the Internet.

Tony had taken a rather strange career path to his position at the
ITU (a position he left at the end of 1991). After studying engineer-
ing at the Florida Institute of Technology, he became quite active in
civil rights. Eventually, he landed himself a seat on the city council,
getting further involved in policy and politics.

From there, through a roundabout road, his interest in telecom-
munications drew him to Washington, D.C., where he went to work
for the chief scientist at the FCC. The chief scientist was also an
Internet old boy, and the FCC was promptly installed as part of the
Internet. While working at the FCC by day, Tony put himself
through law school at night at American University.

Eventually, Tony went up to Boston to become the publisher of
Telecommunications. In Boston, he alleviated the drudgery of punch-
ing out a magazine by working as a Research Associate at MIT's
Media Laboratory.

Altogether, this mix of lawyer, regulator, and engineer was just-
the combination needed when Pekka Tarjanne took over the helm of
the ITU. A physicist by training, Tarjanne was determined to bring
the ITU into the telecommunications age and asked Tony to work as
his personal assistant.

Exploring the Internet

My first face-to-face meeting with Tony, after countless e-mail mes-
sages, was in August when I took the train in from Paris and went
to the twelfth floor of the ITU tower. As I got off the elevator, a
short man with a beard darted in after me, posted a piece of paper
in the elevator and jumped back out again just as the doors
slammed shut.

This was Tony. He was trying to get into every elevator to post
a notice of my lecture the next day. The lecture was news to me, so
I read the notice. Along with the obligatory hype, the notice ex-
plained that I would be lecturing about a host of topics ranging
from the future of networking in the civilized world to how more
bandwidth would lead to world peace.

After catching the last elevator, Tony and I wandered across the
street to the International Organization for Standardization (ISO). It
is ironic that the acronym used by this global standards organization
is a dyslexic variant of its real name.

ISO is a rather strange group. Unlike the ITU, ISO is a private
organization. Like ANSI in the U.S., ISO’s lack of real legitimacy to
carry out its quasi-governmental functions makes the organization a
bit insular and very conservative. Located in a squat, drab building,
ISO occupies the dimly lit corridors of the third and fourth floors.
The halls are almost always empty, but occasionally one can see a
functionary dart out of one office and flit into another.

Tony and I wandered around the halls to find people to invite to
my lecture. The purpose of our visit was to make sure everybody at
ISO knew about the ITU experiment. I was a bit worried that put-
ting ISO on notice would be considered confrontational, but Tony
assured me that there was “nothing wrong with pushing forward
the state-of-the-art a bit.”

My lecture, by the standards of public bureaucracies, was a stun-
ning success. Aside from the obligatory ITU contingent trying to
figure out what their management was up to, there was a large
crowd of functionaries from diverse UN contingencies interested in
the Internet. Only a few people fell asleep, and a few even asked

Geneva

questions. This was described by several people as a “dynamic”
performance.

I described how the standards would be posted on the network,
how different access protocols would be used, and how making
standards widely available had been so successful for the TCP/IP
protocols. I explained how, once the data was digital, we could all
start using advanced services, write better code and how, when all
was said and done, we would enter a state of standards equilib-
rium, a nirvana of documentation.

Unbeknown to us at the time, the audience included an opera-
tive from ISO, sent over to keep tabs on the radical goings-on across
the street. When my own spies informed me of this the next morn-
ing, I placed a call to Dr. Zakharov, head of the ISO computer
group. I explained who I was and said that, at the request of ITU
management, I would be happy to come over and give them a cour-
tesy briefing on our little project.

“That sounds mighty exalted. Come over Thursday morning.”

Two days later, I donned my suit and dashed through the rain to the
other side of Rue Varembé. I bypassed the obligatory visit to ISO
reception and went straight to the third floor and back through a
warren of offices, where I was ushered into the inner sanctum of Dr.
Zakharov.

Walking in, I saw an older man behind the desk wearing the sort
of safari suit that the British favor. “You are Malamud?” he asked.
“I have asked McKenzie to join us” he said, beckoning to a young
man wearing a shirt and tie who nodded at me with a glum gaze.

“McKenzie is our database expert. I have asked him to be here
because what we have here is a database issue, not a network is-
sue.”

“Let me first say that I know all about this Internet of yours,” he
continued. “I used the Internet even before it was the Internet. In
fact, before it was the ARPANET. This Internet of yours may be fine
for researchers, but here at ISO we have real clients to serve, so
enough about this Internet.”

Exploring the Internet

I was still standing. A pause introduced itself, so it seemed an
appropriate moment for a few formalities.

“Glad to meet you,” I said, holding out my hand.

McKenzie continued to glance at the floor and Dr. Zakharov
grunted in what I took to be an amiable fashion so I gathered my
courage and made a little speech about how my efforts with the ITU
were experimental, provisional, something. we could learn from,
temporary, and any other non-threatening platitudes I could think
of.

My little speech contained a fatal flaw. I had used the pronoun
“we.” Now, “we” is not totally inappropriate as this effort was
based on the efforts of several interested parties, but I must admit
that the corporate offices of Me, Inc. do not actually teem with cor-
porate retainers.

“We?” Zakharov pounced. “Who is ‘we’?” he demanded.
“How many members are on your team, anyway?”

Like many corporate types, my lack of a formal institutional af-
filiation did not sit well with him. I sidestepped his question and
explained that, in my view, the question of publishing standards
was, rhetoric notwithstanding, just not that hard. After all, there
were hundreds of ways to get data on-line, ranging from Wordstar
to bitmaps to SGML (an ISO standard).

To say that ISO viewed my little project as a threat was an un-
derstatement. Zakharov let slip that the Secretary-General of ISO
had sent a letter to his counterpart at the ITU protesting the experi-
ment. Zakharov made it quite clear that he viewed my approach as
simplistic, misguided, and naive.

I gently tried to make the point that there were positive aspects
to this project that should not be overlooked, such as increasing the
public awareness of the vital work that groups like ISO were under-
taking by making primary resource documents available to a wider
audience.

Once again, my diplomatic efforts seemed less than successful.
“People don’t need to read the standards,” Dr. Zakharov snapped.
Indeed, he felt that ISO standards were basically unreadable and
that normal people shouldn’t even bother to try. What was appar-
ently needed was some guru to write a book, and that these second

Geneva

or third-generation digested versions of standards were more appro-
priate for the general public. Turns out that Zakharov had himself
written such a book (“a monograph, actually”) on some long-forgot-
ten CAD standard.

I’m all in favor of promoting book sales, but there is still no
substitute for the original documents. A programmer working on
an implementation should be required to RTFM (“Read the F***ing
Manual” or, in the staid company of standards potatoes, “Read the
FTAM Manual”). There is no excuse for second-hand knowledge of
something as vital as international standards.

So far, we had spent 45 minutes of a one-hour meeting about
databases, and the subject of databases still had not come up. It
didn’t look like it was going to. McKenzie still hadn’t said anything
and didn’t appear to be about to burst into action.

As | paused to gather my forces, Zakharov jumped in. “Another
thing, young man. Do the people in charge of the Internet know
about your efforts?”

He seemed convinced that when some appropriate official
learned that I would be distributing documents—“with graphics, no
less’—on the network, my renegade activities would be quickly
stopped.

(Let’s put this into perspective. The NSFNET backbone transfers
well over 14 billion packets per month. Even if my server was
wildly popular, the backbone could handle it. Just to be sure, I was
going to send the big computer companies archive tapes so they
didn’t need to FIP the entire file store into their internal networks.)

I told him that indeed my activities were supported by appro-
priate officials. The chairman of the Internet Activities Board (IAB),
for example, had offered to be one of the anonymous FIP sites for
the standards. I didn’t tell Zakharov that the incoming chair of the
IAB, who had also chaired several ISO committees, had on his home
computer many ISO documents and had half-seriously offered them
to be posted along with the Blue Book.

“It’s not that we’re against the Internet,” Zakharov explained.
“We just want to make sure that all the issues are properly consid-
ered. In fact, we are using the TCP/IP protocols here at ISO.” ISO

Exploring the Internet

had been unable to find enough OSI software to keep their internal
network going, so they used a combination of Novell and TCP/IP.

My allotted hour was nearing the end, so I once again began
mumbling conciliatory phrases about experiments, cooperation, and
similar appropriate fin-de-meeting noises.

“So what do you want from us?” Dr. Zakharov cut in, ever the
diplomat.

I pulled out my extra large shovel and explained that since ISO
obviously had concerns about the experiment, I was simply here to
explain my efforts. “By the way,” I continued, “ISO is welcome to
join in our little experiment. I have plenty of disk space left.”

Zakharov grunted. McKenzie sat.

Zakharov rose. McKenzie rose.

Our little meeting was over.

Back at the ITU, the machinery had swung into action and, with
a few false starts, my tapes were being cut.

Pra gue

With a week to kill before my plane left to go back to the States, I
decided to try and get in a little fun. Sitting in Tony’s office using
Telnet to read my mail in Boulder, I looked next to his mouse pad
and saw a disorderly stack of business cards. On top was the card
of one Jan Gruntorad, head of data communications at the Czech
Technical University in Prague.

With the opening of Eastern Europe, this seemed an appropriate
omen. I picked up the phone and gave him a call. On Sunday
morning, after an all-night train ride, I arrived in Prague. In addi-
tion to the usually high ratio of tourists to infrastructure, I had
timed my visit to coincide with an international congress of 100,000
Jehovah's Witnesses. Every hotel was full. Each clerk in an endless
parade of hotels grunted the same explanation: ““We have Jehovah
Congress today.”

“They've always caused me problems, too” I cracked to one
surly reservations clerk, attempting to inject a little levity into my
fruitless tour of hotels. The clerk stared at me, giving me the same
blank look she would give a moldy potato. My sense of humor was
evidently not appreciated.

Most hotels in Prague are privately run, but Cedok, the state
tourist agency, still books most of the rooms. In one hotel, for exam-
ple, 80 percent of the rooms were controlled by Cedok. Cedok,
which advertises itself as “the biggest travel agency in Czechoslova-
kia” (until recently the only travel agency, so it had a long head
start), doesn’t believe in reservations. When I asked to make a res-
ervation for the next day, they told me to come back the next day
and they would see what they could do.

After 8 different hotels, I found a room. No wonder. The Pra-
gue Palace, at U.S. $230 per night (payable in foreign exchange)
would be no bargain in London or New York. In Prague, it might
be considered extortionate.

I had the day off, so I slipped into the mob of Jehovah's Wit-
nesses and set off to see the sights. For lunch, I stopped at one of

Exploring the Internet

the overcrowded, spartan restaurants and ordered the house special,
a pizza topped with ham, eggs, and peas. The beverage menu also
featured “Bois Avocat,” a French drink which could be translated
either as the blood of a lawyer or some form of avocado nectar. |
had a beer.

With relief, I finished my sightseeing and went back to my
room. The number of tourists in Prague was staggering, but they
didn’t seem to have anything to do. With few restaurants, most
milled around in the streets. The Jehovah’s Witnesses smiled a lot.
It was as if Walt Disney had taken over the country, but had named
Kafka managing director.

The next day, I went over to the Czech Technical University. Jan
Gruntorad ushered me into his office wearing a Hawaiian shirt,
having come in from his holiday a day early to see me. His office
was piled high with papers, computer boxes, a huge 3270 terminal,
and books. An ancient TV was on top of one cabinet, and on the
other were dozens of empty bottles of various kinds. In other
words, my kind of office.

For many years, Czechoslovakia had suffered a technology em-
bargo due to COCOM regulations and, like other Eastern bloc coun-
tries, concentrated on reverse-engineering IBM mainframes. Jan
took me into their computer room and showed me a Russian main-
frame clone, together with a Bulgarian-made front end processor
(FEP). The MVS-like operating system had been written locally and
it had taken two years to integrate it all into a semi-working system.

The Bulgarian FEP and the Russian mainframe were right out of
~ an episode of “Lost in Space.” Full of knobs and mechanical regis-
ters, it reminded me of early computers like the ENIAC, which can
be seen at the Smithonian’s American History Museum in Washing-
ton, D.C. In a corner were a pile of removable disk packs, each the
size of a keg of beer.

Forcing designers to reverse-engineer systems and to serve as
super-systems integrators had produced one positive effect. Many
of the engineers at the school were quite sophisticated. Unfortu-
nately, however, the operating system had never quite run right and
the users never had enough confidence to invest their time learning
to do useful work.

Prague

In 1989, with the revolution, things changed dramatically.
Czechoslovakia was able to put in a 9,600 bps leased line to Linz in
Austria and join the European Academic Research Network
(EARN), the European BITNET. In Bratislava, another school
started using UUCP links over dial-up lines to Vienna and became
part of EUnet.

An important breakthrough for Gruntorad came in 1990, when
IBM donated a 3090 mainframe as part of the IBM Academic Initia-
tive. Starting with Czechoslovakia, but later including other coun-
tries such as Hungary and Poland, plunking supercomputers down
in key locations gave an important boost to Gruntorad’s networking
efforts (you can’t build a shopping mall without upgrading the
roads around it).

EARN had expanded to include seven other facilities in Prague.
Efforts were underway to extend the national backbone to Bratslava
and other cities. From zero users in 1989, the system grew to over
1,500 users by the summer of 1991.

Gruntorad is much more than just the BITNET manager, how-
ever. He was leading the effort to install a national networking in-
frastructure. He wanted to see UUCP, NJE (for BITNET), and
TCP/IP protocols all sharing an infrastructure of leased lines.

Leased lines, as in many European countries, are quite expensive
in Czechoslovakia. A 9,600 bps line from Prague to Brno, for exam-
ple, costs 36,000 crowns per month (U.S. $1,242/month). With
monthly salaries averaging 3,000 crowns, a leased line becomes a
major line item in the budget. The PTT didn’t have the infrastruc-
ture in place to properly support digital lines, so a 64 kbps line
would be provided by ganging up six 9,600 bps line.

Although I managed to time my departure with what appeared
to be the majority of the Jehovah’s Witnesses, some other deity must
have been on my side. Not having bought a ticket allowed me,
through some mysterious bureaucratic rule, to bypass the hundreds
of people waiting to check in and report to the excess baggage win-
dow.

The burly representative from Czechoslovakian State Airlines
who sold me my Swissair ticket even smiled when I suggested she
should feel free to close the airport after I had gone—a sure solution

Exploring the Internet
to overcrowding. They even allowed me to change my crowns back

into dollars without the obligatory official receipt proving that I
hadn’t used the black market.

Zurich was a real contrast to Prague. I settled into an expensive
hotel with small rooms and crossed the street to a beer house, where
I feasted on three kinds of sausages, each accompanied by a large
frothy mug of local beer.

The next day, I went to meet Urs Eppenberger, a technical man-
ager at SWITCH who explained how the Swiss research network
was being established. SWITCH, which stands for Swiss Telecom-
munication System for Higher Education and Research, was estab-
lished six years ago by the federal government and the cantons of
Switzerland to start and maintain a university research network and
a national supercomputer center.

Urs works in the SWITCH network office, which consists of
eight engineers, an administrator, and lots of computers. SWITCH
is a network infrastructure with services like e-mail. The infrastruc-
ture consists simply of leased lines and Cisco boxes installed at 11
sites around the country, plus links into the X.25 Telepac network.
The interface between the campus and SWITCH is clearly defined at
the Ethernet plug. SWITCH maintains its own system, including
configuring the routers.

Originally, the network was going to be religiously OSI, using
services such as X.400, FTAM, VIP, and any other OSI services
available. The system quickly switched over to a more pragmatic
approach, supporting TCP/IP, X.400 over OSI, and even DECnet.

The network is very Swiss. It is carefully planned and carefully
implemented. Impeccably engineered, not too adventurous, and fo-
cusing on offering a reliable production service, SWITCH is an or-
derly addition to the carefully manicured Swiss landscape.

Geneva

I took the train back through the rolling hills of the Bernese Ober-
land and Fribourg to Geneva. There, I went to see Brian Carpenter,
head of networking for CERN, the European physics laboratory.

CERN is one of the major centers for the study of high-energy
physics in the world. The laboratory has over 5,000 visiting scien-
tists, plus another 3,000 permanent staff members. CERN is funded
by 17 European countries, including a substantial East European
membership. The laboratory is mission oriented, a place to do seri-
ous, world-class research. It has, as have many other physics labo-
ratories such as Fermilab in Chicago, been on the forefront of
networking as a way of allowing physicists to exchange data.

A typical CERN experiment in high-energy physics might in-
volve 400 scientific staff from 20 to 40 different institutions. For ex-
ample, one of the experiments on the Large Electron-Positron (LEP),
the new 27-km accelerator, involves physicists from seven CERN
members working alongside scientists from China, India, Israel,
Hungary, the U.S., and Russia. |

Because of CERN’s constituency, it has become one of the global
Internet hubs. Much of the connectivity to Eastern Europe funnels
through CERN. The CERN mail gateway routes messages for hun-
dreds of networks.

Internally, the CERN network runs the usual motley assortment
of technologies, including gigabit Ultra networks, FDDI, and over
3,500 stations on 60 different Ethernets. Over 400 gigabytes per
month are transferred on the FDDI backbone alone. Over 1,000 sta-
tions on a 4 Mbps token ring are used to control the accelerator.

WAN links at CERN are equally impressive. Over 150 gigabytes
per month are transferred out of the laboratory over wide area net-
works. The laboratory has a 384 kbps link to MIT, a T1 line to Cor-
nell, an El line to Bologna, and an 8 Mbps feed to the CHEOPS
satellite. Francois Fliickiger, deputy head of networking, estimated
that of the total 15 Mbps of international bandwidth devoted to re-
search networking, 11 Mbps terminate at CERN.

Paris

I went back to Paris to accomplish the more serious business of eat-
ing and drinking until my plane left to take me back to the U.S. In
August, Paris transforms itself. The French all leave on their annual
four- to six-week holidays. The absence of Frenchmen contrasted
sharply with the surge in the American presence.

It seemed like the entire population of New Jersey had come to
see the Eiffel Tower. They had brought with them a large portion of
Sony's Camcorder output for the past few years. This called for
guerilla tourism on my part, avoiding the Champs Elysées and con-
centrating instead on seedy cafés on the Left Bank. Before leaving,
however, there was one bit of business I wanted to accomplish.

On my last day in Paris, I called Gerard Poireau and arranged to
visit his office, located in the shadows of what used to be the Bas-
tille. Poireau is an evangelist for videotex. One of the original mar-
keting managers for the French Minitel system, he is active in efforts
to spread this underrated technology into the U.S.

When you get a telephone, France Telecom gives you the choice
of a phonebook or a computer terminal. As a result, over 3 million
people have, free of charge, one of the squat brown Minitel termi-
nals in their home. Of these 3 million terminals, France Telecom
estimates 80 percent are actually used. Even if the number is in-
flated, this was still, until very recently, the largest single network in
the world.

The core of Minitel is directory assistance, however. Poireau
had a formidable memory, rattling over a dozen phone numbers off
the tip of his tongue. When his memory failed, however, he reached
over to his Minitel terminal and tapped the two digits to connect
him to directory assistance. He typed in the last name and the city,
and the fax, voice, and audiotext numbers all appeared on the
screen. (Audiotext is the French word for phone-based services like
stock reports or voice mail.) Gerard picked the menu item for the
voice line, and the number was automatically dialed for him.

Exploring the Internet

Minitel is much more than just directory assistance. The usual
bulletin board services—pornography, dating services, IQ tests, and
sports results—were all present, but with a few additional surprises.

For example, Nynex had the New York Yellow Pages up on
Minitel. Prohibited from offering electronic yellow pages services in
the U.S. by Judge Greene’s restraining order, Nynex had no such
restraint in France and was evidently testing their U.S. system for
future deployment. Gerard and I looked up the names of all Ban-
gladeshi cafés in Manhattan and jotted down their locations for ref-
erence on subsequent trips to the Big Apple.

Most access to Minitel was based on the French X.25 network,
Transpac. With over 300 X.25 switches connected at speeds of up to
2.048 (E1), Transpac and Minitel have a symbiotic relationship. A
service like Minitel does no good if there is no access path to it.
Likewise, an X.25 network is a road to nowhere without useful serv-
ices on it.

Minitel was a case study in providing useful information serv-
ices. By the beginning of 1990, Minitel had over 12,000 service
providers. Even intra-company transactions, incuding banking, in-
surance, and order processing, occurred on the network. The direc-
tory service alone received well over 10 million calls per day and
train reservations received over 1.5 million calls per day.

Leaving Gerard Poireau, I went to catch my plane back to the
U.S. Through some fast talking, I managed to get my no-refund,
nochange, economy ticket upgraded to a first class seat. It’s amaz-
ing what you can get, if you just ask.

I sat munching my Beluga caviar and drinking my Dom
Perignon champagne and decided that travel wasn’t so bad if you
just get the right seat. By the time the single malt scotch, the vin-
tage port, and the freshly-peeled kiwi fruits had come and gone, I
even became optimistic about the weeks ahead converting ITU data.
Drink can certainly cloud your judgment.

Boulder

At the New York airport on the way back to Denver, I sat in the bar.
Over in the corner was a motley assortment of people that could
only be a rock and roll band on tour. Leather pants, nose rings,
orange hair, and a half-dozen drinks in front of each person were
just a few of my clues.

At the check-in counter, the band members stood around mak-
ing the agents nervous. Evidently, they had booked their seats late
and were not pleased with their assignments of middle seats. As I
was in the same boat, I stood with this crew harassing the ticket
agent. I surreptitiously flashed my frequent flier card and she took
my ticket and gave me an aisle seat.

Getting on the plane, I saw that I was next to one of the disgrun-
tled musicians, this one adorned with shoulder length hair and vari-
ous martial adornments on his leather jacket. He squeezed over me
and sat down.

“You have to be with a rock and roll band,” I ventured.

He grunted, presumably in confirmation.

When we got into the air, he called the flight attendent over and
ordered two double screwdrivers (“hold the orange juice”) and be-
gan methodically going at them. I drank a beer and started reading
The Chinese Screen.

To my great surprise, he leaned over and asked “Is that Somer-
set Maugham?” I said it was, and we sat there discussing late nine-
teenth century English literature.

His name was Wurzel and he was in Motorhead, the quintessen-
tial British loud rock band. They were on tour with Judas Priest
and Alice Cooper in a heavy metal extravaganza. Wurzel’s job was
lead guitarist.

“Basically, I try and play as loud as I can,” he said, explaining
his job.

“Well, somebody’s got to do it,” I replied.

This comment pleased Wiirzel so much he waved to the flight
attendent to bring two more double vodkas for himself and a beer

Exploring the Internet

for me, simultaneously repeating my new rationale for his existence
to his mates scattered a few rows up. The flight attendent was treat-
ing this crew quite gingerly lest they mistake the plane for one of
their hotel rooms, and promptly brought the drinks.

Before landing, Wiirzel handed me his business card and invited
me to the big show at some location he couldn’t remember. Alas, I
was still reading my mail by that evening and let this opportunity
for cultural enrichment pass me by.

Although I got back in late August, it was into the second week of
September before my tapes from the ITU finally arrived, leaving us
about 20 days to convert the tapes before the hooptedoodle began.

At INTEROP 91 Fall, on October 11, a live video link to Geneva
had been arranged. Pekka Tarjanne and Tony Rutkowski would
make the big announcement about the newly freed Blue Book to a
packed and hushed audience in San Jose.

My job was to get the standards converted and on the network
before the announcement so there would be no turning back. Until
we actually let the documents loose on the Internet, there was al-
ways the possibility that the bureaucrats would somehow gain the
upper hand and stop the experiment.

A week spent trolling the halls of the ITU had produced docu-
mentation on about half of the proprietary, in-house text formatting
system they had developed many years ago on a Siemens main-
frame. The computer division had given me nine magnetic tapes,
containing the Blue Book in all three languages. Despite repeated
tries, we had gotten nowhere on trying to also get the CCIR radio
recommendations and reports, even though the Secretary-General
and the head of the CCIR had both appeared anxious to see this
data online.

Along with the magnetic tapes, we had half a dozen TK50 and
TK70 cartridges from the ITU VAXen. The cartridges contained PC
files that were stored on a VAX using DEC’s PC networking prod-
ucts. We had two types of files, one of which was known to be
totally useless.

p44

Boulder

The useless batch was several hundred megabytes of AUTO-
CAD drawings, furnished by the draftsmen who did the CCITT il-
lustrations. Diagrams for the Blue Book were done in AUTOCAD,
then manually assembled into the output from the proprietary text
formatting system. The draftsmen were very helpful and quickly
said I could have any data I needed.

On my way out of meeting with the draftsmen, however, one of
them starting asking some questions about scanners and babbling
on about TIFF files. I was puzzled. Why should I care about scan-
ners and TIFF files when I had the diagrams in the original formats?

Turned out that AUTOCAD was indeed used for the diagrams,
with the exception of any text in the illustrations. The textless dia-
grams were sent over to the typing pool, where people typed on
little pieces of paper ribbon and pasted the itsy-bitsy fragments onto
the illustrations. Come publication time, the whole process would
be repeated, substituting typeset ribbons for typed ribbons. A nice
production technique, but the AUTOCAD files were useless.

The rationale for this bizarre document production technique
was that each diagram needed text in each of the three official lan-
guages that the ITU published. While AUTOCAD (and typing) was
still being used, the ITU was slowly moving over to another tool,
MicroGrafix Designer. There, using the magical concept of layers,
they were proudly doing “integrated text and graphics.”

The second batch of DOS files looked more promising. Modern
documents, such as the new X.800 recommendations, were being
produced in Microsoft Word for Windows. My second batch of
tapes had all the files that were available in the Word for Windows
format, the new ITU publishing standard.

To do the conversion, I was quite lucky to be working with Sun
Microsystems on a research grant. They had sent over two large
servers for a research program I was participating in and graciously
agreed to allow us to use one server to post standards on the net-
work. Without their help, we wouldn’t have had the resources to
do anything.

Exploring the Internet

Step one was to begin tackling TPS, the ITU wonder program devel-
oped years ago. I brought the tapes over to Mike Schwartz, a pro-
fessor at the University of Colorado and my partner on the Sun
research grant.

The ITU had documented the format we could expect the tapes
to be in. Each file had a header written in the EBCDIC character set.
The file itself used a character set seemingly invented by the ITU,
known by the bizarre name of Zentec. The only problem was that
the header format wasn’t EBCDIC and the structure the ITU had
told us would be on the tape wasn’t present.

Using Captain Crunch Decoder Rings, we finally figured out a
collating table for the mystery header character set and managed to
hack the files off the tape. There were large amounts of data at the
beginning and end of files which seemed useless and was simply
deleted. We crossed our fingers that the deleted information would
not be needed later and indeed, it wasn’t.

Next, we had to tackle TPS. This text formatting language was
as complicated as any one could imagine. Developed without the
desire for clarity and simplicity I had come to expect from the UNIX
operating system and its tools, I was lost with the Byzantine, un-
documented TPS. |

The solution was to take several physical volumes of the Blue
Book and compare the text to hexadecimal dumps of the files. I
then went to the Trident Café and spent a week drinking coffee try-
ing to make sense of the data I had, flipping between the four files
that might be used on any given page of text trying to map events
in the one-dimensional HexWorld to two-dimensional events in the
paper output.

Inbetween trips to the coffee house, I was trying to take care of
diagrams and the PC files. Diagrams were simple: I sat down every
morning for a few hours and scanned in diagrams. The diagrams
were saved as TIFF and EPS files, then uploaded to our Sun server.

The PC files were all unloaded onto a VAX, then moved over to
the Sun, then downloaded at 9,600 bps to my home network. There,
the files were loaded into Word for Windows, and then exported as
Rich Text Format, the Microsoft proprietary standard for open docu-
ment interchange. The RIF files were then converted to Word Per-

Boulder

fect and ASCII, and all four file formats were sent back up to the
Sun.

All told, it wasn’t unusual to be downloading and then upload-
ing 10 to 20 megabytes per day, all using a 9,600 bps modem. Still,
this was the easy part. It was TPS that almost killed us.

Finally, after pages and pages of PERL code, we had the begin-
nings of a conversion program. We had tried to use the software
developed at the ITU to convert from TPS into RTF, but the code
had been worse than useless.

The day before leaving for INTEROP, I was still working desper-
ately away on the conversion program. Tables and equations were
still not coming out the way I had wanted them to, but finally, it
came time to start hand editing. Any tables that couldn’t convert
properly were thrown out. Same with equations.

At 2 AM., with the SuperShuttle coming at 6 A.M., it was finally
time to pack for INTEROP (and the six weeks of travel that would
immediately follow INTEROP). The data wasn’t perfect, by any
means, but the Blue Book was on the Internet ready to be distrib-
uted. The Bruno project was ready to roll.

As I packed, I reflected on what had been an awful 20 days,
programming like a madman to do the conversion. It was amazing
how the ITU had been doing a conversion, with lots of people avail-
able, but had reportedly estimated that it would take a total of 10
years, and roughly U.S. $3.2 million, to complete the job. By my
calculations, if it would have taken 40 days to make the conversion
perfect, my time would have been worth, on the ITU scale of reality,
U.S. $10,000 per hour. My only hope was that I might be able to use
this data to justify an increase in my consulting rates.

As I was going to sleep, I tried to figure out how bad a case of
bureausclerosis one would have to have to in order to turn a 20 day
(40 to make it perfect) conversion effort into 10 years of agony.
Over the next few months, as I visited Geneva again and saw the
experiment run into the famed international bureaucracy, I was to
learn how it could easily take 10 years, or, more likely, never get
done at all.

Das

ae, we ay a ) Oe
Se ars . A ner oe i eid we p .
| : oe a ae ch! a
Gpaongneiean bite 6 fort
oH 1 yiters dept” | semana area et bk ie
its sini ribris bk: pol wtleoth do ein may.
vihercnghurh eqd 008.2 a gritas ee aq ¢
Che TY Gayest teeta sted
jn tyedtis vert iia enh vibe ae Yen ag ey. btie 2
otavilostiadt! oeu-dtabelesr del os. | maka
“ytaaem ure TP teieBat abot sea iT OSE,
a @ r PALI ‘i “ ih wel 1 it "3 Af Pi a veenedt meee ner : -
Lave en yoni, Utekews PORE Laotenivest incied
Sted verre pote Reeth. eae wad antes asovifngie Nn! soit ti —
do lata buds air # = hye: pear lus 1 Afusnn leat Nae RT
bik os | ee yi cag aidid eh ne reise besirh inate atte Ate? ae
toot ” act ay va Pw gees sud ace te ale, i |
Yiloritl voltit hone 3 da gitar ebtiagsiie appa cifveng sii Ay
| tlaiett toreiansawik Gildea) 9 nr TT 701
(rie wet aha aia Sony eit TOF AAR, wollojas
inte tdact, yk en korragtess onli! agi enue autlfobhd —

i 2
sed 7
's

lieth thee ob nisi honce ng bearnnanangy a ASU
eR RTs, helaetia cB eR rf tor he rsp" Kaine MN ioe dtp : BY shih sil
Dassen mee: it ibe ney wit ob of runt een § sail S aes iene
lis yh Al qoasiee abot tty yA RAEN RM (2% % BO rota ma sic
0 Jocleloi.5 iat bigow, di ted Perrin vie
fan, Vhtheh gel Thales eer pours tet.) baie aul
"pags antago pene AKAD OP ae hepesel bheowiodicg!
GNA, 20 eisna LT! ot oe Lt FMI ey adh, ieee
rig! Gxt thet Faerie jack} ae eh Ay’ view eMac ti Af ap" eo
cutpik . estes gs snitias WO? ite es species (0
it Jeph | EP WOG, BiG ere yt § Joh be itd ek: 34
wa Ket Tad eee RM ape
LApgs Jonge ehay, tea alRayno MD,
ad Neg hay Sy ERS PET om Retin AP
o) ene fr OPAC IAGO! ara ae
2 AA, ee Ua “6
wey ives wine daecied bebe
~ Rich Teat Speman, the Mien “ah
“rhant interchange.” Tie Es
> he | i ei ie yr
7 a a “te ;

Ay 4

Round 1

From INTEROP to IETF

7 ny
a el

¥ bro
LTV Re

| VN eg te

TA of GOGAT
aS : Cat Al oy athe 7

San Jose

When people ask me to explain how INTEROP is different from
other trade shows and conferences, I like to describe it as resem-
bling a circus, but not a zoo. A circus appears to be all chaos, but it
is carefully managed chaos. A zoo, on the other hand, puts the ani-
mals into cages and lets them do what they want.

Networld—referred to by one wag as “Notworld’’—is the trade
show equivalent of a zoo. Each little cage has its own independent
little show. In some cages, exuberant marketing types reach out to
touch you with their brochures. In others, they stand listlessly
around. ‘The visitors to Networld stroll from one cage to another,
feeding their name tags to the animals, hoping that one of them will
do something interesting.

INTEROP, on the other hand, is a show. Granted, there are also
cages, but those are a small part of what is going on. There are
tutorials, a conference, and, most importantly, a huge, operational
network.

When I arrived on Saturday, October, 5, 1991, people were just
beginning to descend on San Jose. After checking into the luxurious
Holiday Inn, I wandered over to the Fairmont Hotel. In a meeting
room, about 50 volunteers, engineers from all over Silicon Valley
and from around the world, were gathering.

All were quite bleary-eyed. Most of them had been up all night
stringing cable in the Diamond Pavilion, one of the auxiliary exhibit
halls being used by INTEROP. This gathering was getting ready for
the real work that lay ahead that night: installing the network at the
San Jose Convention Center.

All of the vendors (with a couple of minor exceptions like my
publisher) were connected to a real operational show network,
which in turn was connected to the Internet. This was no coaxial
cable down the center of an exhibit hall. The network used over 35
miles of cable and connected 300 vendors together with what one
observer estimated would be the equivalent of the components
needed to wire a 20-story high-tech skyscraper.

Exploring the Internet

The organizing principle for the network was a series of ribs.
Each rib ran through a physical area, such as an aisle in the conven-
tion center or the Diamond Pavilion. Each rib had an Ethernet
based on unshielded twisted pair and a 16-Mbps token ring. Each
of the 25 ribs had a 19-inch equipment rack at the end with the
electronics to drive the subnets and routers to be connected to the
backbones.

Two different backbones connected the 50 subnets, one based on
FDDI, the other on Ethernet. A T1 line connected the Diamond Pa-
vilion to the Convention Center. A microwave link went to two ter-
minal clusters nearby in the Fairmont Hotel. Yet another T1 link
linked the show network to the NASA-Ames Research Center,
which in turn provided links to the Bay Area Research Network and
out to the NSFNET backbone.

This entire network was put in by a semifanatical team of volun-
teers, overseen by two INTEROP staff members frantically trying to
keep some sense of order. The challenge to this network is that you
don’t get convention centers a month ahead of time to put in your
network. In fact, as we met Saturday afternoon, they were still tear-
ing down the booths from the previous week’s Seybold Publishing
Conference.

At midnight, the Convention Center would be turned over to
the Interop Company. At 8 AM, just eight hours later, dozens of
18-wheel tractor trailers would roll onto the convention floor to start
setting up the vendor exhibits. By that time, all the cable needed to
be off the floor: try telling a teamster on a forklift to take the long
way around because he might crush the fiber.

Between midnight and 8 A.M., 35 miles of cable had to be rolled
out and hung from the ceiling with cherry pickers. Equipment
racks had to be moved safely into place, and equipment for the Net-
work Operations Center (NOC) had to be moved up into the control
booth overlooking the convention floor.

To make this all happen, a core team of a half-dozen volunteers
met with Interop all year to plan the network. In July, the company
got the convention center for a day and hosted a cable laying party.
All the cable was laid out and connectors added. Then, the cables
were tied together and carefully rolled onto drums and moved into

San Jose

a warehouse. Before the show, there was a hotstaging in the ware-
house, where the cables were connected to equipment racks to be
tested. Then, everything was packed up onto pallets to await the
teamsters.

At midnight, a group of about 50 had gathered in the lobby of
the convention center. Each member of the core team had a differ-
ent colored shirt, each with the words “Do Not Disturb” stencilled
on the back.

I was assigned to the teal team, under the leadership of Karl
Auerbach. Karl was one of the founders of Epilogue Technology
and a long-time participant in the INTEROP ShowNet. He’s also a
lawyer, which makes him a formidable rabble rouser at IETF meet-
ings.

With a loud bellow, we were all called over for a briefing by
Stev Knowles. (The ending “e” in “Steve” got left off of a mail mes-
sage once, and Stev decided he preferred it that way.) Stev is vice
president of engineering at FIP Software and is widely acknow-
ledged as the loudest member of the ShowNet team.

Stev’s briefing was, as usual, direct and to the point. “Do what
you re told and if you have a question, ask.”

Stev is an interesting character. Rumor is that he got involved in
the very first ShowNet because he couldn’t read his mail. The net-
work wasn’t working, so he marched into the show and comman-
deered the ShowNet team until things started working.

By sheer force of will, he and the other core team members do
this every year, staying up for several days straight to get the net-
work up and running. For this, everybody gets a t-shirt. Of course,
the core team got put up by INTEROP at the Fairmont, but even the
impressive bar bills they ran up didn’t quite explain why they did
this.

A few minutes after midnight, the doors opened and we all
stood on the cavernous convention floor, clustered around our lead-
ers. Two tractor-trailers were driven into the center of the floor and
began to dump their contents. Equipment racks, spools of cable,
and various other network paraphernalia were all hustled to their
proper locations.

Exploring the Internet

The barrels of cable began to be unspooled. Teams of volun-
teers, spaced every few feet, would march a string of cable across
the hall. A few ties holding cable were cut and the feeds that would
hang down from the ceiling were separated from the main rib.

Then, five cherry pickers started a slow march down the conven-
tion floor. At each rib, the cherry pickers descended, then in unison
(or some ragged semblance of unison) lifted the cables to the rafters.

Hanging down at intervals were coils of cable. These feeds were
left about 16 feet off the ground, high enough to clear the semis
coming in the next morning but low enough to reach without a
cherry picker.

By 5:30 A.M., most of the ribs were successfully up, so I went
back to the luxurious Holiday Inn for two hours sleep before I
started my day of meetings. The volunteers were still working
when I left and would continue to do so for the next two days
straight. Getting the cables up was the most time-critical task, but
plenty of work still remained. Cables had to be connected to equip-
ment racks, connectors tested, the backbone had to be tested, ven-
dors had a million questions, the Internet link needed to be
initialized, and a million other details had to be finished before the
show opened on Wednesday. |

The volunteers that put in this network are an amazing bunch.
Some of the best network managers from all over the country come
just to help out. When they are done, they have a complex internet-
work up and running and connected to the rest of the global mesh.
Three days later, it is torn down.

While the ShowNet is operational, it supports some fairly heavy-
duty applications. Groups of vendors get together to demonstrate
the interoperability of standards such as Frame Relay, SMDS, X.400,
SNMP, and many others. In marketing-speak, these are called Solu-
tions Showcases. As you walk through the exhibits on the conven-
tion floor, different booths have little signs indicating which
showcases they are part of.

The Solutions Showcases are certainly useful as a marketing
tool. Lots of users come to the show to see what works and who
sells the equipment. However, the showcases are equally important
to the engineers that design and make the technology.

San Jose

I spoke to one engineer who says he gets more bugs worked out
in one week at INTEROP than he can in six months in the lab. By
testing his implementation with those of other vendors, he can
quickly hone in on ambiguities in the standards and figure out what
to do to make the standard an interoperable reality.

The INTEROP week passed by quickly. The first two days, I sat in
on a few of the tutorials, hearing people like Craig Partridge talk
about gigabit networks or MIT’s Jeffrey Schiller talk about network
security and Kerberos. On Wednesday, the show officially began
and the pace picked up. The plenary address was via T1 video link
from Geneva. I was particularly interested in this address since I
was going to need the same satellite link on Friday.

Things went smoothly Wednesday morning, mostly because the
previous two days had been spent in frantic preparation. The video
link was donated by Sprint. Coordination of this part of the show
at INTEROP was handed to Ole Jacobsen. Ole had worked long
hours with Sprint technicians, watching the line go from Geneva to
Atlanta until, the day before, it had reached Kansas City. By Tues-
day night, video was finally making it all the way from Geneva to
San Jose.

Ole was the perfect choice to handle anything telephone related.
Officially, he is editor and publisher of the ConneXions journal. Un-
officially, he is a phone junkie. For example, he has a PBX in his
house. Of course, the PBX is a small one with only 6 lines and 16
extensions, but how many houses do you know with their own
PBX? His wife Susan still can’t get used to dialing “9” for an out-
side line.

After the plenary, I dove into the exhibition and conference.
Most people were in high-speed data acquisition mode, trying to
work out optimal patterns for navigation of the floor, or flitting
from one conference session to another hoping that time division
multiplexing would enhance their information intake.

My personal favorite, as it had been the year before, were the
SNMP demonstrations. In 1990, John Romkey had developed the

Exploring the Internet

Internet toaster. The toaster was hooked up to the ShowNet with a
PC running TCP/IP and SNMP software. Workstations around the
show floor had network management software, complete with a
toaster Management Information Base (MIB). By setting variables
on the toaster MIB from a network management station, people
could make the toaster start toasting.

The problem everybody immediately saw with the 1990 demon-
stration was that you had to put the toast into the toaster. This was
suboptimal from the point of view of the network engineer, who
wanted to stay in bed while breakfast was made.

This year, the FIP Software booth sported the new, enhanced
Internet toaster. Using the latest Lego technology, they built a little
crane that would pick up a piece of bread and deposit it into the
toaster slot.

What made this demonstration especially funny is a bit of an
inside joke. If you read Marshall T. Rose’s The Simple Book, you will
see that he feels strongly that people should make better use of the
get-next operator for efficiently using the network and the remote
management agent.

To get his point across, Marshall always refers to the get-next
operator as the “powerful get-next operator.” The INTEROP audi-
ence was truly impressed that the powerful get-next operator was
being used to make toast.

Friday morning was the official unveiling of the activities of the
Document Liberation Front. I got up early and went over to the
Center for the Performing Arts, the biggest hall at the INTEROP
show, seating 2,701 people. This hall is so big that it even has
enough room to hold the hundreds of students that show up for the
tutorial taught by Doug Comer of Purdue, a researcher well known
for his explanations of TCP/IP fundamentals through his tutorials
and books.

The members of the panel started to show up. Dr. Vinton Cerf,
chairman of the Internet Activities Board, was there to lend his sup-
port to the announcement. Richard desJardins, head of a training

San Jose

group called The GOSIP Institute, was there to provide his commen-
tary on the different models of standards making.

We finally got through to Geneva. Although neither Tony
Rutkowski nor Dr. Tarjanne was present yet, we held our breath and
got ready to start the session. Looking out into the audience, which
contained no more than 50 people who had bothered to get up for
the early morning session, I was very glad that the cameras couldn’t
beam back to Geneva.

We made our announcement and the one member of the press to
show up furiously took notes. The announcement went well, but I
was worried that nobody would care and that all this effort had
been in vain.

Over the next two months, I was to discover how wrong I was.
Leaving INTEROP and travelling around the world, I got more
news at every stop on how popular Bruno and the Sons of Bruno
had become. I thought the choice of Giordano Bruno (1548-1600) as
the namesake for my server in Colorado had been an appropriate
one.

The Greeks, in those days before the printing press, used a mne-
monic method for remembering verse or other forms of knowledge.
The Dominican order had kept alive these secrets during the dark
middle ages, but had kept them tightly guarded. Bruno joined the
Dominican order in 1565 and mastered the Dominican secrets.
Then, he revealed the secrets to the rest of the world in his classic,
Shadow of Ideas (1582).

Bruno was expelled from the Dominican order, and was later
denounced to the Inquisition in Venice for acts of heresy, including
telling jokes in poor taste about God. He was burned at the stake in
1600.

Would Bruno’s secrets help the world? Would the ITU Inquisi-
tion kill our server? These and other questions were on my mind as
I boarded a plane Sunday, bound for Hawaii and then onwards,
three times around the world in the next six months.

Honolulu

I came to Hawaii to learn more about packet radio. Really.

The fact that I spent most of my time on the beach drinking Mai
‘Tais was an accident.

Honolulu is the home of Torben Nielsen, a man who has played
an important role in helping to spread the Internet into the Pacific
Rim. The University of Hawaii provides the Internet link for several
Asian countries, particularly Korea and Japan.

He is also known for starting up the University of Hawaii net-
work. Networking was not originally considered to be a high prior-
ity at the University. In fact, it was such a low priority that nothing
was happening. Torben enlisted a few of his colleagues (including
four department chairmen), rented power tools, and started digging
trenches.

For a total of U.S. $1,500, lots of hard work, and the extreme
displeasure of the facilities management group, the renegade profes-
sors managed to get four buildings linked together in a network.
The project was done so cheaply that they even salvaged refueling
lines from military aircraft to use as conduit.

Nielsen’s laboratory is a windowless room lined with foam pad-
ding and absolutely stuffed with computers. The Blue Book, Sun
documentation, and various other books line the wall. Sitting inside
this room, it is hard to imagine that nearby is a world of beaches
and drinks decorated with paper umbrellas.

I arrived at 8:30 A.M. and Torben had already been there for two
and a half hours. When he mentioned this, I made the appropriate
polite noises of amazement. Torben looked at me like some shirker.
“T’ll be here till 9:00 tonight and we do this seven days a week,” he
patiently explained to me.

Needless to say, a leisurely lunch was probably not going to be
in the works here. I realized that my time slot in Torben’s schedul-
ing algorithm was going to be limited, so I tried to learn quickly
what I could about his work.

Honolulu

Getting lines up and running is what Torben is best known for,
but he is rapidly tiring of keeping pipes open and bits flowing. In-
stead, his interests are higher up the stack in areas like videoconfer-
encing, the Andrew File System (AFS), interactive books, and a wide
variety of other applications.

The topic of the day appeared to be SGML and ODA, so we
talked about revisable form document architectures. In Torben’s
view, the Office Document Architecture is basically useless. To him,
SGML tagging is the appropriate internal format for a document.

As evidence he cites a commercial program that uses SGML as
the internal representation of documents. The SGML document is
then moved through TeX, which processes the tags. TeX is thus a
formatting engine for SGML, which in turn produces output in a
language such as PostScript for final form representation.

In Torben’s view, ODA fills the same intermediate role as TeX.
You can use SGML source and use it to produce ODA, which is then
sent to some ODA reader for display on the screen or the printer.

After this brief flurry of discussion on revisable form formats,
Torben dismissed me and switched context to some other work. |
borrowed a terminal to read my mail.

When I left the cocoon of Nielsen’s lab, it was still only 10 AM,
much too early to return to the Waikiki tourist ghetto. I swung my
car over the mountains to the other side of the island.

After driving around the sugar plantations, I arrived in the little
town of Haleiwa. There, I spotted a sign saying “Best Lunch in
Haleiwa.” Afraid that this might also well be the only lunch in
Haleiwa, I pulled onto the back lawn of the building, almost ran
over a dozen cats and parked next to the outhouse.

After a wonder lunch of Mahi Mahi and eggs smothered in a hot
sauce made with lilikoi flowers, I felt much better. I spent the rest
of the day in exhaustive research looking for the perfect Mai Tai.

Tokyo

Arriving at Tokyo’s Narita airport ahead of schedule, I waited two
hours for my bus into town. After an hour on the bus, we were still
wending our way through the Tokyo suburbs, passing Disneyland,
the docks, and threading through ever larger mazes of freeways.

Looking down, I realized the freeway was built on top of a river.
Next to us, only a few feet away, were rows and rows of office
buildings. Looking inside each, I could see rooms and rooms filled
with row after row of desks all crowded together. Below us, but
above the river, ran the railway tracks, above us another freeway.

The next morning, I ventured forth on the subway system to
find the University of Tokyo, known as Todai. There, I met Profes-
sor Haruhisa Ishida, a professor at the Computer Centre. Professor
Ishida proceeded to give me an excellent introduction to networks
in Japan.

Japan, like the U.S., has many different networks. BITNET, as in
most countries, was initially funded by IBM, but is now member
supported. What membership fees don’t cover is provided by the
main sponsor, the Science University of Tokyo. Japan’s BITNET has
a 56 kbps link to CUNY and provides tail links to Korea and Tai-
wan.

N-Inet is an older, proprietary network to tie together main-
frame systems with services like remote job entry and remote login.
N-Inet was managed by the National Center for Science Information
Systems (NACSIS). NACSIS, a research institute funded by the
Ministry of Education, also maintains a link to FIX-West, with two
other links continuing on to NSF and in Washington and the British
Library.

NSF uses a portion of the line to search several large databases
maintained by NACSIS. A small part of the line is available for BIT-
NET and TCP transfers and is used mostly for mail exchange from
their X.400 messaging system. NACSIS plans to upgrade the U.S.
link to 192 kbps in 1992, at which time it will become one of the key
international links for Japan.

Tokyo

A third network is the Todai International Science Network
(TISN, pronounced “Tyson”). TISN uses the DECnet protocol suite
and is used by physicists and chemists. TISN maintains a 128 kbps
link between Todai and the University of Hawaii. Due to political
walls between organizational fiefdoms, Japan also maintains HEP-
net connectivity at the High Energy Physics Laboratory (KEK) to the
Lawrence Berkeley Laboratory in California.

One last network is JAIN, a TCP/IP-based university network
that links the university LANs together. This whole plethora of net-
works is tied together by JUNET, based on UUCP, and WIDE
(Widely Integrated Distributed Environment), the Japanese Internet.

There are two paths between the Japanese Internet and the rest
of the Internet. WIDE maintains a 192 kbps link from Keio Univer-
sity in Fujisawa to the University of Hawaii. In addition, the 128
kbps link between Todai and Hawaii, used primarily for DECnet
traffic, acts as an automatic backup in case the 192 kbps link has
problems.

WIDE is an interesting network in the Japanese system. All the
other networks are funded by the Ministry of Education or another
group. Officially, WIDE doesn’t exist. Even more amazing, in the
tightly segmented world of Japanese politics, commercial and edu-
cational users are all mixed together.

I asked Professor Ishida how such a situation could come to be.
His answer was quite simple.

“Jun Murai.”

Jun Murai used to be a research associate at Todai, working with
Ishida, but recently moved south to his alma mater, Keio University.
In the staid world of academics, Jun is a fairly remarkable character.

Just for starters, he wears blue jeans. Failing to dress in the
regulation dark blue suit has caused no small amount of comment
among senior faculty members. Jun gets away with it because he
really knows what he is doing. He commands a loyal following
among students, has the respect of all his peers, and has even won
the grudging respect of his seniors.

The WIDE network is based on donations of money from corpo-
rations and labor from graduate students. The network lives almost
hand to mouth. Money is funnelled into Murai’s “research pro-

Exploring the Internet

gramme” and is used to pay for the network. It was not unusual at
times to have the coffers get down to 1 or 2 months of operating
costs, forcing Murai into perpetual fundraising.

After this introduction to Japanese networks, Professor Ishida
gave me a tour of the Todai facilities. The main campus is wired
with three FDDI backbones, one for TCP/IP, one for DECnet, and a
third for administrative computing. A fourth 400 Mbps backbone is
used for video. Fanning out from each of the backbones are UTP-
based Ethernets. These local networks form the point of connection
for terminal clusters, workstations, and even the supercomputers.

Professor Ishida led me past a peopleless room stuffed with Hi-
tachi mainframes and supercomputers into a terminal cluster. The
cluster was divided up into cubicles and was dead quiet. At the
entrance was a color video display with a map of the cluster. The
occupied cubicles had red dots inside and the empty cubicles were
marked in green, allowing people to find an available workstation
or terminal without disturbing the people already working.

Off to the side was a glass-lined room with 9-track tape and car-
tridge drives. Each drive had a terminal in front of it with a menu
system to help users do their own tape work.

Another room was filled with printers. When a user prints a job
at Todai, it automatically spools to disk. Each printer has a card
reader attached to it (magnetic card, not punch card, that is). A user
walks up to the reader, slides an ID card through, and the job is
retrieved from disk and printed. A terminal in front of each printer
indicates how long the current queue is.

Leaving the cluster, we went down to the first floor where Pro-
fessor Ishida pointed to a large electronic signboard on the wall.
The board displayed the current status of the mainframes, including
the number of jobs and the expected delay before a new job would
begin processing. No need to walk upstairs and log in if the system
is slow.

Walking outside, I felt a little whir under my feet as the auto-
matic brushes on the doormat came to life, cleaning my shoes. I
headed down to the subway.

Tokyo

Coming off the subway, I ducked into a tiny, dark noodle shop
where a line of salary men all sat hunched over the counter slurping
noodles. Hoping my neighbor hadn’t chosen chicken lips, I pointed
to his dish and was promptly handed a steaming bow] of delicious
miso soup.

Feeling refreshed and refurbished, I went back to my hotel, en-
tering at the same time as LaToya Jackson and a very large entou-
rage. I fell in with the entourage and smiled graciously at the hotel
reception committee.

On the way to the elevators, we were waylaid by a group of six
American tourists all bearing labels to identify their tour group and
all armed with cameras. Figuring LaToya could handle this one
alone, I slipped into the elevator that three hotel staff members had
been guarding for her. Before they could say anything, I punched
my floor number and the startled attendants went jumping out in
search of another elevator.

I spent the afternoon in the Pub Misamu, waiting for an evening
dinner engagement with Tomoo Okada, general manager of Fu-
jitsu’s Value Added Group. My appointment with Tomoo Okada
was at 7 PM. At precisely 7:01, my phone rang. I descended to the
lobby to meet what was obviously a very senior manager, a distin-
guished-looking executive in his early 50s. Okada supervised more
than 700 people, including Fujitsu’s Value Added Network division,
two wholly owned subsidiaries, and a collection of divisions of
other subsidiaries. His main responsibility was running a network
used by both Fujitsu and its customers.

The network is essentially an X.25 network with 144 local access
nodes spread throughout Japan. Over 250 leased lines running at
speeds from 128 kbps to 6 Mbps form the network backbone. Run-
ning on the network are classic protocols for asynchronous termi-
nals, such as X.28 and X.29, and synchronous protocols for 3270 and
Fujitsu terminals. As a general rule, full protocol stacks aren’t run
on the network, although an insurance company has begun deploy-
ing an OSI CONS-based service over X.25. International links in-
clude 64 kbps lines to Sydney, Singapore, Korea, Hong Kong, and
Malaysia. Higher-speed links are available to the U.K., the U.S., and
Germany.

Exploring the Internet

Perhaps the most interesting project Okada supervised is
NiftyServe, a wholly owned subsidiary that acts as the licensee to
the U.S. CompuServe. NiftyServe preserves the famous Com-
puServe user interface, but the software was entirely rewritten to
run on a UNIX platform and to support Kanji characters.

To make NiftyServe start off with a bang, 32,000 Fujitsu employ-
ees were given accounts. Over 70 percent of those accounts are ac-
tive users. The reason for this high utilization is quite
simple—important meetings, such as promotion reviews, are posted
there.

Some of the most devoted users are overseas Japanese employ-
ees. In addition to providing things like daily Japanese language
news, the service has proved important in another respect. In Ja-
pan, when family or friends die, it is considered very important to
immediately express condolences. With Fujitsu or any large corpora-
tion, of course, fellow employees are family. Before the bulletin
board service started, it could take days for the postal service to
deliver the news overseas, forcing Fujitsu employees into the un-
willing position of appearing impolite.

NiftyServe has added a few other interesting twists to the classic
bulletin board. For example, you can instruct the system to redirect
your mail to a fax machine. You can even go to a pay phone and
have your mail read to you.

All this information about Fujitsu was imparted to me with
rapid-fire delivery over a seemingly infinite parade of dishes in a
dim sum restaurant. After polishing off a half-dozen large bottles of
beer, Okada suggested we switch to a Chinese aged wine, similar to
sake, but darker in color (and at least as potent).

Even with all these drinks, Tomoo Okada kept up a steady de-
livery of information on Fujitsu. With my head swimming and my
stomach stuffed, I stumbled back to my hotel to await the next day’s
pilgrimage to Fujisawa to visit Jun Murai, the Internet Samurai.

Fujisawa

The next morning I was up early, having heard horror stories of
how hard it would be to navigate the succession of subways and
trains in the middle of rush hour.

Yes, there were packers to get more people into the subways.
Since the density was only about equivalent to a crowd leaving a
football game, however, the packers weren’t yet needed. They sat
watchfully by, wearing white gloves and ready to jump into action
when the time came.

The density of the crowds was amazing. Passing the Tokyo sub-
urbs, I expected the crowds to disperse, but the Yokohama commute
brought even more people onto the trains. Even outside of Yoko-
hama in Fujisawa, the density remained constant.

Arriving at Tsujido station, I waded through huge crowds until I
got outside the station. There, I started waving Jun Murai’s card
around and mispronouncing “Keio,” my destination, until a kindly
old gentleman in an indeterminate uniform pointed me to the stop
for bus number zero.

I queued up and boarded. We passed through Fujisawa and
started to head into the countryside. Throughout the ride, every
time we passed a cluster of university-like buildings, I would accost
some hapless passenger, mumble “Keio,” and wave the business
card around. One by one, the passengers waved for me to sit down.

Finally, we reached the end of the line. Rising out of the coun-
tryside, sitting on a hill, was a futuristic clump of buildings, sur-
rounded by empty fields. I made my way up a long set of stairs
into the central courtyard. In marked contrast to the subways,
trains, and streets, constantly teeming with people, the Fujisawa
campus of Keio University, Japan’s best private university, was ab-
solutely deserted.

Feeling like I was in some futuristic ghost town, I wandered
around random buildings until I chanced upon a group of adminis-
trators hidden away in a back office. Waving Jun’s business card
produced a flurry of whispered consultations until a map of the

Exploring the Internet

campus suddenly appeared and I was pointed towards building
number zero.

Going to the third floor, I knew I was in the right place when I
saw an office with a pile of empty computer boxes piled outside. I
parked myself on the boxes and waited for Jun. It was an eerie
feeling, sitting there in this deserted building on this deserted cam-
pus listening to the building creak and groan. Occasionally, a door
would open and a person would scoot out of a room and scamper
away, leaving only the echoes of their footsteps.

I was to find later that the deserted campus was merely an illu-
sion. Being a self-employed idler, I had arrived well past 9, by
which time everybody was already hard at work.

Jun Murai’s secretary soon arrived. Known worldwide as “Jun-
sec, her e-mail address appears on all WIDE literature, posters,
stickers, and the like. Junsec let me onto an X terminal to read the
several hundred mail messages that had accumulated since Hono-
lulu. Most of the mail was about Bruno, the standards server.

Evidently, word had gotten out. Bruno was running 24 hours a
day, with load averages of 35 packets per second not uncommon.
So many mail requests had come into the infoserver software that
the batch queue had over 150 unfulfilled jobs. The amount of FIP
traffic was so heavy that we had become a serious user of band-
width on the transatlantic links.

This posed a delicate problem. On the one hand, we wanted
lots of people to use the server in Colorado so that we could prove
to the ITU that the service was needed. On the other hand, being
good network citizens meant setting up mirrored servers, putting
replicas of the data on other machines to avoid having too much
redundent traffic cross the backbone.

Offers to provide mirrored servers were pouring in from all over
the world. Comparing notes with Tony Rutkowski, it appeared that
every country in Europe had volunteered to maintain a set of the
standards. Several hundred hosts had already received files from
Bruno and it looked like we were getting FIP traffic from over 25
countries.

Fujisawa

When I was almost done plowing through my mail, Jun Murai came
bursting in. Disheveled, animated, and wearing his trademark blue
jeans, he greeted me enthusiastically. Taking me into the obligatory
conference room, we began talking while Junsec brought in cups of
tea (she had already brought me two cups of coffee).

I presented Jun with a copy of STACKS for his boss, Dean Aiso.
He presented me with a handsomely bound copy of the WIDE an-
nual report, several hundred pages of kanji with gold-embossed let-
tering on the cover. My manners being better than my Japanese, I
thanked him while mentally trying to figure out how to get this
heavy and, to me at least, cryptic tome back to the States.

Formalities duly dispensed with, we started talking about
WIDE. Jun is only 36 years old and I was curious how, in a strati-
fied society like Japan, somebody like him had ended up running
the Japanese Internet.

In 1984, NTT still had a throttle on all telecommunications. Put-
ting alien devices (e.g., modems) on the telephone lines would be
considered about as proper as greeting the Emperor by slapping
him on the back. However, it was widely known that in April, 1985,
NTT would deregulate.

All the senior researchers had been debating how to take advan-
tage of deregulation to put in networks. Meetings were held to de-
bate the subtleties of various OSI architectures. To Jun, this was a
waste of time. As he puts it, “I was young and that was boring.”

He took two modems, scammed a phone line from university
administrators (no easy feat), and started running UUCP transfers.
That was the start of JUNET. While the establishment continued to
attend OSI meetings, JUNET continued to grow. Links were set up
to a machine named mcvax in Amsterdam (the precursor to EUnet)
and to seismo in Washington, D.C. (the precursor to UUnet). By
1986, a domestic IP network had started and by 1989, Jun Murai and
Torben Nielsen had established a link to Hawaii. Larry Landweber
helped to hook up Japan to CSNET.

Meanwhile, the powers that be continued to debate OSI. When
they finally looked up from their deliberations, Jun already had sev-
eral hundred nodes on his network. As in much of the world, while
committees waited for OSI, a few people turned TCP/IP networks

Exploring the Internet

into a reality. Japan, like other countries, had many people trying to
legislate networks into existence while a few people rolled up their
sleeves and installed cable.

By the 1991 school year, WIDE had grown to the point where it
had an annual budget of U.S. $1 million raised totally by donations.
WIDE was officially a research project, involving 57 researchers,
roughly half from the university community. The operational re-
quirements of WIDE became so demanding that Jun was actually
turning away requests to be connected until a suitable infrastructure
could be set up to run the net.

During our conversation, Jun repeatedly referred to a desire to
stop running networks and get back to his real research. Even with
the demands of running WIDE, he has compiled an impressive re-
cord.

When JUNET was first being put into place, Jun noticed “we
had a network but nobody was using it.” Networks are like a fine
dinner: the whole effort is wasted if nobody comes. The reason in
the case of JUNET was quite simply that e-mail and USENET used
the Roman alphabet and Japan uses kanji characters. Jun changed
that. He and others added kanji support to the X Window System,
kanji character handling for RFC 822 mail, and multibyte character
handling for the C programming language.

He also helped design a font server. This software lets users
start spelling out characters in romanized Japanese. As the words
begin to be formed, Kanji characters start appearing on the screen.
When the right character shows up, the user points to it and pro-
ceeds to the next word.

When I visited Fujisawa, Jun was involved in fascinating pro-
jects to support mobile hosts and mobile people. One of the more
interesting projects was the Phone Shell. The Phone Shell is a fully
functional UNIX shell which takes input from a 12-key telephone
pad instead of a keyboard or a mouse. Typically, users execute
scripts, such as having mail messages redirected into a voice synthe-
sizer.

As Jun explains it, “I can go to the bar and drink beer. I go toa
phone and ping my routers, and if they are still working, I go back
and drink more beer.”

Fujisawa

Another project brought ISDN into the TCP/IP protocol suite.
Unlike other countries, which are still demonstrating the viability of
demos, Japan has deployed ISDN as an operational service. In fact,
WIDE uses ISDN lines to supplement leased lines in case of conges-
tion or failure.

The WIDE ISDN module is fully integrated into TCP/IP. When
a datagram for an ISDN-reachable source is encountered, a call is
placed. The delay to set up the circuit ranges from 800 ms to 3
seconds. Once established, the line stays up until it has been idle
for a user-configurable period, at which point it is taken down.

ISDN can be used for more than just routers, however. Jun has
TCP/IP on his laptop. Tokyo has ISDN pay phones. He can bring a
laptop into a phone booth and be a fully functioning member of the
Internet. Even Superman would be jealous.

Another project close to operation when I visited was the use of
satellite circuits for home PCs. Japan has deployed satellites as an
alternative to cable TV. Dishes cost as little as U.S. $100 each. Satel-
lites for TV signals are a one-way data channel, but they operate at
high bandwidth. A PC with an ISDN card can use 64-kbps “B”
channels to send commands. Data coming back to the PC can use
up to the 8 Mbps wide-area bandwidth of the satellite dish.

Ethernet cards are used to give the PC a 10 Mbps interface in a
LAN environment. Jun and his staff had linked the Ethernet card to
the satellite receiver to give the home user the ability to do WAN-
based multimedia, large file transfers, and other operations requir-
ing large amounts of bandwidth.

We went from Jun’s laboratories to tour the campus. Fujisawa
was a brand new campus of Keio University. In sharp contrast to
other Japanese (and most U.S.) universities, senior researchers at
Keio teach freshmen and sophomores. Jun delighted in telling how
he would instruct a class of freshmen, “now we will ping Switzer-
land.”

At the Fujisawa campus, all entering freshmen are required to
learn UNIX, to operate workstations, and to acquire other basic
skills. Although laptops are not required, they are very strongly en-
couraged.

Exploring the Internet

We walked into the library where students run an ID card
through a reader to enter. Of course, Jun forgot his card, so we
crawled over the barriers. We walked into a lounge area where
groups of students sat clustered around HDTV sets watching assign-
ments.

Around them, students were at workstations, each equipped
with a cassette tape drive and a headphone, doing language exer-
cises. Other students were debugging programs.

Finally, it dawned on me.

“Jun, there are no books in here.”

“You still have books?” he asked with a smile. The books were
up one floor, but it still gave the library a strange feeling.

Walking outside, we spotted Junsec patiently waiting for us ina
car. We all went into Fujisawa where we were ushered into a pri-
vate tatami room for an exquisite lunch of many courses, each beau-
tifully arranged and delightfully prepared.

As I was gulping down the tenth course, a delicate broth with a
single tiny mushroom and a little piece of fish, Jun started explain-
ing to me how the spindly little mushroom I had just inhaled was
handpicked in the forests and only available for a few weeks each
year. Each mushroom cost as much as 5,000 yen (U.S. $35). I puffed
out my cheeks as if I were still rolling the mushroom in my mouth
to savor the delicate flavor, while Jun went on to explain how rare
the fish was.

Akihabara

Akihabara is a district of Tokyo devoted to consumer electronics,
where building after building is stuffed with everything from 2 by 2
foot stalls specializing in diodes or resistors to full-fledged, multi-
floor shops, each so full of devices that Marconi would have
thought he was hallucinating.

I knew I had reached Akihabara proper when I looked down on
the sidewalk and saw a street vendor selling oscilloscopes and mi-
crochips. In a stall behind him, 60 different kinds of laptops were
on sale, ranging from plain vanilla clones to full-fledged 80386, 60
Mbyte, 6-pound notebooks.

Thinking I must have died and gone to geek heaven, I knew that
this district was going to take a while to visit. I ducked into a noo-
dle stand and used my two-word Japanese vocabulary to order a
beer and a bowl of miso soup. As I noisily slurped my noodles, I
could hear dozens of stereo systems, all playing different brands of
disco at full volume.

Suitably fortified, I spent the next two hours trooping up and
down stairs in building after building. I had to keep reminding my-
self that I had many more weeks on the road, lest I yield to the
temptation to buy a home satellite dish (only $100) or a personal
computer that weighed only 980 grams and ran on AA batteries.
The computer was only available in a kanji model, but I figured I
could always learn. Even more tempting were DAT drives, high-
definition TVs, global positioning system receivers, and three-inch
color televisions.

Feeling somehow unsatisfied, yet still solvent, I tore myself
away and headed back towards the subway. On the way, I passed a
group of young women all dressed up in natty uniforms bearing the
somewhat cryptic label “With Me” and all walking in a line. The
women at the head and tail of the column carried banners which
matched their uniforms. The rest of the “With Me” girls each jaun-
tily carried the new two-pound Hitachi notebook, shamelessly dis-
playing their VDTs for all to see.

Exploring the Internet

Every ten minutes or so, the conga line would go out into Aki-
habara and walk for a few blocks, dragging along a sound system
guaranteed to attract lots of attention. They would wend their way
back towards the With Me store and disappear inside, luring a por-
tion of the mob in with them. Somehow, I suspected the short skirts
had more to do with the crowd’s loyalty than any desire to purchase
Hitachi's latest and greatest.

Hong Kong

I arrived at 11 PM. at Hong Kong’s Kai Tak airport, thinking that the
God of Aviation must have been guarding me. A new gate opened
up at customs just as | walked up, and my luggage was the first off
the chute.

But, this apparent good fortunate was only to set me up for a
cruel joke by the Aviation Deity. Once outside the airport terminal, I
saw a line of several hundred people at the taxi stand. It was the
end of the long weekend that marks Ching Ming, the Grave Sweep-
ing Ceremony, and all the planes from the mainland had just re-
turned.

] jumped the barrier to get out of line and walked up the road to
where taxis were letting off customers. Standing with my baggage,
the temptation proved too much for one driver. We both looked
furtively around for police and I jumped in.

There were only two problems. He knew no English and he
looked like he had been driving for at least 48 hours. We circled the
freeways for 20 minutes until it was established that I was going to
“Number One Hotel, Happy Valley.” Actually, Happy Valley had
only one hotel, so number one didn’t mean much, but it did get us
heading the right direction.

The sleep problem was a bit scarier. Every time his head would
start to nod, he would jerk up suddenly and then wave his arms
around in spasmodic circles, presumably to get the circulation go-
ing. I know it certainly kept me awake.

The next morning, I strolled over to see Edwin Yeung, an MIS man-
ager at the Royal Hong Kong Jockey Club. The Jockey Club is to
Hong Kong what the Louvre is to Paris. Gambling is the national
pastime and the Jockey Club is the national icon.

Officially, the Jockey Club is the only place you can bet in Hong
Kong. In practice, of course, all sorts of other gambling takes place

Exploring the Internet

ranging from bookies to mahjong parlors. In theory, though, if you
want to gamble you go to the neighboring island of Macau. If you
want to bet you go to the Jockey Club.

Founded in 1884, the club is run as a non-profit organization. It
makes so much of a non-profit that it is, by far, the largest charity in
Hong Kong. In the 1990-91 season, for example, the club handed
over HK $1.04 billion (U.S. $138 million) to charity.

The club conducts 67 racing days a year, either in Happy Valley
or at the mammoth new Shah Tin racetrack. A Diamond Vision sys-
tem, like that used in U.S. football stadiums, shows the race at the
track not being used, allowing overflow crowds to go to the alter-
nate location to watch the races and, of course, place bets.

This is what you might call a serious transaction processing ap-
plication. In the 1990-1991 season, over 187 million bets were placed
for over HK $47 billion (roughly U.S. $8 billion). On a typical race
day, 2.8 million bets were placed. To make this especially challeng-
ing, the vast majority of bets were placed within the last 10 minutes
before a race, yielding a transaction processing rate of over 600 per
second.

Like a bank, this is a system that better not make mistakes. Un-
like a bank, this was a system that you really don’t want going
down (or even slow) right before a race. A bank can afford a few
minutes’ delay, but bettors don’t appreciate having to wait until the
race has started to have their money refused.

To funnel these bets into their coffers, the Jockey Club uses two
main systems. Cashbet is for people handing over cash, and Telebet
handles people who maintain accounts with the club.

Cashbet consists of terminals that accept money, or debit cards
which lift money right out of a bank account using electronic funds
transfer. The Jockey Club operates 127 off-course betting counters.
These have roughly 2,900 terminals staffed by operators. Another
2,900 operator-staffed terminals are at the two race tracks. In loca-
tions throughout the city, people can place their own bets at 144
Electronic Funds Transfer (EFT) terminals and 700 cash-based termi-
nals.

Hong Kong

Off-course centers use 275 leased lines to connect to a series of
12 PDP 11/44s that act as front-end processors. Another 16 PDPs
handle the terminals at the race tracks.

The Telebet system allows customers who maintain an account
with the club to place their bets remotely. Bettors can place bets

using a telephone or using a special custom terminal developed by
the club.

The telephone system is quite simple. You place a call and are
routed to one of 1,600 operators at Sha Tin or 388 operators at an
off-site facility on Tsing Yi island. The incoming lines are all con-
nected to a bank of over 2,000 voice recording lines to resolve any
later disputes. An average of ten times each race day, a particular
conversation has to be located.

Each operator on the Telebet system has a terminal which is con-
nected to a series of VAX 3100s and PDP 11/44s. Operators enter
the bets into the system and the bets make their way through the
Jockey Club network.

Rather than continuing to add operators to service increased de-
mand, the Club decided to design a custom portable betting termi-
nal, known as the Customer Information Terminal (CIT). Bets are
entered using a series of menus on a touch screen. Then, the device
is plugged into a nearby phone jack where it autodials the Jockey
Club network, using banking-standard cryptographic security to
protect the line. The bet is placed, the customer’s account debited,
and the line cleared, all in a few seconds. A voice transaction could
easily take a minute.

The amount of time per transaction is a real issue for this appli-
cation. Ten minutes before race time, the number of calls completed
on the phone network spikes to 20 to 30 times over normal levels.
In fact, the telephone system was so overloaded that it was esti-
mated that only one in eleven call attempts actually got through to
an operator queue. In other words, to handle the Jockey Club load
at race time would require a telephone network designed to handle
more than 200 times the normal load.

To alleviate the load, a dollar minimum on bets was imposed
five minutes before race time, but even this didn’t solve the prob-

Exploring the Internet

lem. The key was reducing the connect time per transaction to fun-
nel more transactions into the available capacity.

The typical CIT customer could be described as a Hong Kong
Yuppie. The devices requires a deposit (the interest from the de-
posit covers the unit cost in five years) and a service fee. Over
30,000 CIT terminals are in the field. Higher numbers had not been
shipped, the Jockey Club said, because of a bottleneck in manufac-
turing which had limited production to only 400 units per month.

Once a bet makes it through the front-end systems, it hits the
Jockey Club back-end network, based on lots of VAXen. In fact, the
Jockey Club is Digital’s largest customer in Asia. It has struck quite
a few people as ironic that such a major customer for the straight-
laced Digital should be a major betting operation.

The club maintains 10 VAX Clusters at the two race tracks, with
a total of over 30 large systems, 22 HSC controllers, and 100 Gbytes
of disk space. In addition to the VAXen, special data security mod-
ules on the network provides authentication for Telebet based on a
6-digit PIN code. Other systems control the video display and pro-
vide links to disseminate race results to other networks, including a
paging system, newspapers, X.25 networks, and Hutchison Mobile
Data’s packet radio network.

On race day, the two sites share processing and back each other
up in case of failure. Transactions are logged at both sites. Connect-
ing all this together are dual Ethernets at each site and an FDDI
backbone that spans more than 20 km, using single-mode fiber pro-
vided by the telephone company.

The club has an aggressive policy towards vendors and technol-
ogy, often telling the vendors what to make instead of being content
to choose from glossy brochures. When their first network went in,
John Markwell who would later become the Head of Information
Technology (the senior MIS position), designed a custom data link
protocol. For the CIT terminal, the Club participated heavily in the
design.

While this heavy involvement caused occasional grumbles from
vendors and caused the staff to have a bit of NIH (Not Invented
Here) syndrome, you couldn’t argue with success. The system

Hong Kong

worked, and worked well enough that it has been licensed to simi-
lar operations in eight countries.

After my briefing, I was taken on a tour of the facilities. In the
machine room, I was asked to put on a white lab coat. Lots of tech-
nicians with windbreakers, baseball caps, and tennis shoes were
walking around inside the machine room, making my medical outfit
a bit superfluous, kind of like smoking a cigarette in the clean room
of a semiconductor factory.

After the tour, I suggested that we all go out to lunch. I sug-
gested (rather strongly in fact) that Chinese food was in order. A
worried look crossed my hosts’ faces who had evidently been plan-
ning on taking me out for Western food just to be safe, but they led
me through the lunchtime throngs to a place jammed with people.
We had a nice lunch of noodles, sesame shrimp cakes, and Peking-
style chicken, and I grabbed a pair of chopsticks and dived in, to the
evident relief of my hosts.

After lunch, I left the Jockey Club and boarded the subway to Sham
Shui Po to visit the Golden Shopping Center, the red-light district of
the computer industry. The Golden Shopping Center is like a sleazy
version of Akihabara, all packed into three floors of an immense
building crammed with warrens of shops. There are three reasons
to go to the Golden Shopping Center: cheap hardware, pirated soft-
ware, and counterfeit books. Whenever I go, I make three passes
through the place.

First, I look to see if any of my books have been pirated into
Chinese. Figuring that my descriptions of DECnet Phase V (a book
I like to think of as “paperware about vaporware”) wouldn’t be any
more popular in Chinese than in English, I searched instead for my
Novell book.

To tell the truth, I always kind of hope to see my books. Van
Nostrand Reinhold, the legitimate publisher, had kept my books
such a tightly guarded secret that I would have welcomed having
somebody read them, even at the cost of no royalties.

Exploring the Internet

I soon became convinced that I was not a celebrity in this dis-
trict, so I made my second pass to look at software piracy. Stopping
in stalls with names like Dream Maker and Ultimate System Com-
pany, I perused the lists. Each shop maintained a master list, a
menu du jour of illicit data, if you will.

Software here was priced by the disk. Low-density disks were
HK $15 (U.S. $2) each, high-density disks are HK $25 (U.S. $3). To
see how much your favorite software would cost, simply count how
many disks it is distributed on. Each list had several hundred soft-
ware titles, each title with the number of disks written next to it. In
the U.S., I had just purchased Microsoft Word for Windows under
an extra-special rebate program for U.S. $129. The Golden Shop-
ping Center was having a special sale for U.S. $12.50.

Particularly impressive were programs like AUTOCAD (15 high
density disks) or even a full distribution of SCO’s Open Desktop, a
wonderful combination of UNIX, Ingres, Motif, and a DOS emula-
tor, available on 46 high density disks for U.S. $138. In the U.S., the
price was U.S. $1,000 and was, quite frankly, a great deal even at
that price. In Hong Kong, it was literally a steal.

As I walked from shop to shop, I watched locals and foreigners
queuing up. Business was booming, despite the Hong Kong gov-
ernment’s assurances that piracy had been stopped. A year ago,
there had been a crackdown of sorts. Then, you could buy your
software at the flat rate of U.S. $2 per disk, but the goods had to be
delivered to your hotel in a brown paper bag. Things were back to
normal now, and you only had to wait for “copy a: b:” to complete
its magic.

Going from stall to stall, I turned the corner to face a 3 by 9 foot
stall called Macrosoft. It featured a sign that caught my attention:
“Licensed Software Available Here.” This was highly unusual. In
fact, it was downright miraculous.

Crowded in this booth were three young men, several dozen
PCs, and an attractive display of shrinkwrapped software. The spe-
cialty of the stall was “Booky,” a three-pound computer with a 40
Mbyte drive and a 9 inch VGA monitor, all packaged in a box small
enough to fit under an airplane seat. It was a desktop system, but
the unit was smaller (and much more useful) than my DECnet book.

Hong Kong

There was a bit of a language barrier, so I spent 10 minutes es-
tablishing the fact that I didn’t want to make copies of their legiti-
mate software. I wanted to know if they sold many copies, and
they kept insisting that I couldn’t make copies, sorry, please.

Finally, the boss, who couldn’t have been more than 18 years old
got off his mobile phone. We talked for a while, and I delicately
brought up the fact that he was selling software for an order of
magnitude more than his neighbors.

“Do you sell much software?” I asked.

A long pause.

“No,” he said.

“Why don’t you make copies like other stores?”

Another pause.

“Some things are right, some things are not.”

As he said this, I noticed a beatific glow arising from him and
realized that I was looking at one of the first computer engineers
destined for sainthood.

Leaving the hallowed grounds of Macrosoft, I turned the corner
to find myself at the administrative offices of the Golden Shopping
Center. Inside, two very large men sat smoking cigarettes and
drinking tea. Outside the office were posted the financial reports of
the center, detailing gross rentals, overhead, and other items of in-
terest to the pirates that ran the stalls. I took out my camera and
started snapping pictures of the financial reports, feeling a bit like a
Spy.

As I was snapping the last picture, I noticed one of the secretar-
ies who had come back from lunch, standing there and staring at
me. She scurried inside, yelling something I couldn’t understand
(but could guess). It was evidently time to explore other areas, so I
made myself as inconspicuous as a foreigner wearing a suit can be
in such a situation. I felt a bit like a marketing executive trying to
hide at an IETF meeting.

I then started my third pass through the center, looking at hard-
ware. This is always the most enjoyable part. I put my self right-
eousness aside, put away my investigative cloak, and became just
another digital tourist.

Exploring the Internet

The selection is not as unusual as in Japan, but the prices cer-
tainly were wonderful. This is the place to come if you are looking
for the latest 486 motherboards, fax modems, cheap disk drives, or
other components. My favorites are the little shops that sell com-
modities at good prices, perfect for people building clones.

One shop, for example, featured a box of XT motherboards for
U.S. $12 each. Inside the store was every conceivable part from so-
lenoids to keys to keyboards. Definitely the place to come if you
wanted to make your own generic PC.

Leaving the bowels of the Golden Shopping Center, I decided it was
time for a beer. I threaded my way through the noodle stands and
incense vendors and bought a can of San Miguel in a nearby phar-
macy.

I tried to sit at a table next to a noodle stand, but quickly discov-
ered no noodle, no chair. I perched atop a crate of empty bottles
and watched the vendors cut up chicken giblets and toss them over
bowls of steaming noodles. It reminded me of when I lived in
Bangkok, where I would have one of these bowls everyday for
lunch.

Wearing my going-to-meeting clothes and sitting on a box furi-
ously scribbling notes, I fit in about as well as an Elvis impersonator
in the Senate Caucus Room, but I met the bemused stares of the
noodle cooks with my own noncommittal look. Finishing my beer, I
went back up Fuk Wing street, past the snake shop (great for soups)
to the subway. One interesting attribute of the Sham Shui Po area,
where the Golden Shopping Center is located, is its proximity to the
Hong Kong airport landing pattern. As you walk along, it is not at
all unusual to look up and see a Boeing 747 appear to be landing on
the building in front of you.

I took the subway to Tsim Sha Tsui, right on the edge of the
Kowloon side of the bay. It was still too early for my dinner ap-
pointment, so I strolled by the bay and watched dozens of boats
scurrying about their business, then went to the lobby of the Penin-
sula Hotel to await my friends.

Hong Kong

The Peninsula is one of those grand old hotels of the Orient. It
is the kind of place that sends a Rolls Royce Silver Shadow to meet
you at the airport and has the chauffeur phone ahead with your
cocktail order so it is waiting for you when you alight.

The lobby is a place where you can nurse a gin and tonic (or a
beer if you happen to be uncivilized) and watch the beautiful peo-
ple act genteel. The thirty-foot ceiling is ornately decorated with
golden ceiling ornaments, angels, and other remnants of the 1920s.

Going up to the concierge, I asked if I could borrow the display
copy of “The Peninsula, Grand Old Lady of the Orient,” explaining
that I was a journalist and wished to include some information
about his establishment in a story I was writing.

“Could I borrow the book for a few minutes for a story I’m
working on? I'll be at that table right over there,” I said, pointing
three feet away.

“$160,” he said without looking up. My journalist disguise evi-
dently was not working.

Tuesday morning, my assignment was to meet Dr. Nam Ng (pro-
nounced just as it looks), director of Hong Kong University’s com-
puter center. The university is one of six in Hong Kong and I had
gotten Dr. Ng’s name out of John Quarterman’s classic guide for the
digital tourist, The Matrix.

To get to Hong Kong University, you go up winding streets
steep enough to make San Francisco look like Kansas. Not an inch
of space is wasted in Hong Kong, a city with one of the highest
population densities in the world and built on islands that are noth-
ing more than the tops of submerged mountains. The university is
no exception, carved into a hill that most countries would call a cliff.

After a 45-minute expedition from my hotel, I ended up spend-
ing only 20 minutes with Dr. Ng. Yes, they had a 9,600 bps BITNET
link. No, they didn’t connect to the Internet. The PCs were based
on Novell’s Netware.

All this was obviously useful to the students (the microcomputer
labs were well equipped and full of intent-looking undergraduates),

Exploring the Internet

but this didn’t fulfill my dual aims for the day: finding material
worthy of a technical travelogue and reading my mail, which was
accumulating at the rate of fifty messages per day back in Colorado.

It was still early in the day, so I called over to the Chinese Uni-
versity in the New Territories. I started with the operator, and after
several false starts, ended up connected to Michael Chang, a lecturer
in Information Engineering (lecturers are the equivalent of U.S. as-
sistant and associate professors). He offered to let me dial in from
Wanchai on the main island, but I didn’t relish the idea of reading
200 messages at 2,400 bps on my prehistoric excuse for a notebook.
Besides, I’d never been over to the New Territories, having confined
myself on previous visits to Kowloon and Wanchai.

I tumbled down the hill from Hong Kong University, crossed
under the bay on the MTR subway, and boarded a train for the New
Territories. I passed the huge Shah Tin racetrack and got off at the
stop for Chinese University. There, I took a shuttle bus up an even
steeper hill, finally stopping at the Lady Shaw building, right across
from the Run Run Shaw Building. I asked a couple of students on
the bus who the Shaws were, but nobody seemed to know. (I found
out later that Run Run Shaw produced kung fu movies and was the
man who brought Bruce Lee to America.)

Michael Chang turned out to be young, competent, and very
friendly. He ushered me into a room full of DECstations and PCs,
and started up a guest account for me. As I plowed through my
mail, I noticed students around me confidently working their way
through the Internet, reconfiguring systems, and otherwise acting
like normal undergraduates. I found out a little later over lunch
that Chinese University had only received its Internet link two
months previously.

After lunch, I crossed back over the bay to find Waleed Hanafi,
Managing Director of Hutchison Mobile Data. Hong Kong has the
highest concentration of mobile phones in the world. It is not sur-
prising, therefore, that this is where the world’s first public radio
data network sprang up.

Hong Kong

Waleed started the company, a subsidiary of the Hutchison con-
glomerate, in 1987, and by 1989 Hutchison had fielded its first net-
work. When Federal Express contracted with Hutchison for
communication with its trucks, the network achieved a firm finan-
cial footing.

Hutchinson’s network is based on 35 transmitters, strategically
posted around Hong Kong using 20 different channels of the radio
spectrum. The architecture of 20 channels and 35 stations allow ex-
pansion by stacking. In the financial district, one site already had
two transmitter/receivers. There is no reason why 20 couldn’t be
stacked up at one site.

When a radio modem powers up, it checks the 20 different chan-
nels. Each transmitting site continually transmits a signal and the
modem scans until it receives a signal with an acceptable bit error
rate, set at the time I visited at 0 percent.

At that point, the modem sends a “here I am” message to the
transmitter/receiver, which logs the unit ID. All units sharing a
channel with a particular station use an arbitration scheme to decide
when to send data. When the receiver is occupied, the transmitter
sends a busy bit. When the channel clears (i.e., a modem is done
sending a packet), the busy bit is dropped.

At this point, each modem backs off for a random period of
time, then listens again. If the signal is still free, that modem gets to
send. Otherwise, somebody else got the channel.

Collisions are theoretically possible, though highly unlikely be-
cause of the short packet transmission time. The key to this arbitra-
tion scheme is a very fast attack time by the modem—f the line is
available after the backoff, the modem sends very quickly.

One thing that makes radio modems much simpler than cellular
phones is the fact that handoff from one cell to another is not an
issue. A packet is sent to one station. When necessary, the modem
simply gives a “here I am” message to a new station, and the ID for
the unit is moved. There is no need to keep a continuous circuit
going as there would be for isochronous voice traffic.

Once a packet hits a transmitter/receiver, it is then routed into
the Hutchison computer systems, which are simply two pVAX 3500s
running the Ultrix operating system. The two pVAXen keep track of

Exploring the Internet

individual units and routed packets out to other sites. For example,
Federal Express packets go back out via an X.25 link to the Federal
Express host system.

A unit called a Develnet acts as a sort of LAN bridge, taking a
radio packet and translating it into X.25, Ethernet, or Asynch-based
formats. The Ethernet packets could go over a Hughes WAN bridge
into a Novell LAN; the X.25 packets out over the Hong Kong public
data network called Datapak, or even out via X.75 links from
Hutchison’s network to reach Sprint’s X.25 network.

The biggest problem with a service like this is convincing people
that it is worth using. “Great features” doesn’t usually convince
procurement agents. To spur demand, Hutchison took an 80386 and
made it into a fancy BBS. Waleed founded the first Hong Kong BBS
in 1982, so this was certainly a natural tack to take. The system on
Hutchison’s network links to several information sources, including
weather servers and the Knight-Ridder news service.

Using your radio modem and a laptop, you could sit in your car
and read the news wires. More importantly, in Hong Kong at least,
you could check out stock prices.

One other link on the network which has proven quite popular
is connection to the Royal Hong Kong Jockey Club’s Telebet system.
Customers hooking a CIT portable betting terminal up to a
Hutchison radio modem can place bets from anywhere in Hong
Kong.

For the outlying islands, this is often how bets are placed on
race day. One old gentleman on the island of Ping Chau habitually
sets up a unit next to the ferry depot. Islanders come to him to
place their bets. At construction sites, the foreman often sits in a
minivan and places bets on race days, thus keeping his crew from
disappearing.

Mobile radio networks are big business. Hutchison is a world-
wide conglomerate and hopes to expand the data network into
other countries. Hutchinson is putting a network into England and
several other companies are putting in systems in other countries.
In the U.S., IBM and Motorola have joined forces to put in the Ardis
public data network. In Sweden, Ericson is putting its own Mobitex
system in.

Hong Kong

Although the service has potential, it is taking a while to get
started with consumers. Hong Kong, with the high concentration of
cellular phones and small area, is certainly an ideal place to start.
Yet, even in Hong Kong, less than 1,000 radio modems are being
used.

I left Waleed’s office and took the subway back to Happy Valley.
Actually, that’s not quite true. I took the subway to the wrong stop,
walked around lost for 30 minutes, and then, after asking directions
several times, took a taxi to Happy Valley.

There, I strolled through the market past piles of smoked oys-
ters, dried mushrooms, fish heads, and jackfruit, and finally parked
myself on a stool at a duck stand. After a few minutes of panto-
mime, a cold beer was produced and I spent the next few hours
eating duck giblets smothered in chili sauce and drinking San
Miguel. The cooks couldn’t figure out why I was scribbling frantic-
ally in a notebook, but decided after a while that anybody who
liked duck innards couldn’t be all bad.

One interesting thing I noticed, between bites of giblets, was that
people would come in off the street and, without asking, pick up
the phone and place a call. Hong Kong Telephone has a flat
monthly fee throughout the islands, so it is not considered at all
rude to just walk in and pick up a phone. Try that in your average
New York deli.

An outgrowth of the flat fee has been some interesting value-
added services. Datapak, for example, offered a free ASCII-Fax
service to its customers, since the marginal cost of doing so was
zero. Flat-rate pricing, if properly designed, can be an ideal way to
spur emerging industries. Commercial IP service providers in the
U.S., for example, often used this strategy to encourage customers
who might be scared of per-packet, per-hour, or other volume-based
pricing schemes. Waleed Hanafi was using the same strategy to get
his radio modem service off the ground.

Once an industry matures, more sophisticated pricing structures
become possible. At first, however, customers don’t have any idea

Exploring the Internet

what their use of a new service might be and don’t want to assume
the risk that their users might accumulate infinite service fees. Set-
ting engineers loose on a volume-priced Internet is equivalent to
leaving a dozen teenagers in a room with a list of 976 numbers
posted next to the phone.

I finished my duck guts, drank up my last beer, and returned to
my hotel.

Macau

Wednesday morning, with the day to kill before the flight down to
Singapore, I took the hydrofoil down to the island of Macau. Set-
tled by the Portuguese in 1554, Macau is now known for its casinos.
It is also one of the few places that the government of North Korea
maintains contact with the western world.

I took a cab down the stately Rua da Praia Grande to the Com-
mercial Building. On the 23rd floor, I got off at the Talented Dragon
Investment Company, Ltd., where I asked for Mr. Cho.

I ended up in this situation as a result of my dinner the previous
evening with my friend Harry Rolnick, a writer who, among his
many other accomplishments, is the music critic for the South China
Morning Post.

“What's new?” I asked.

“Nothing much,” Harry responded. He paused and took a
drink of his Mekong whiskey.

“Oh,” he added, “I’m writing a tourism guide to North Korea.”

I laughed. Harry has a great sense of humor.

“Really, I am,” he insisted.

The week before, he had noticed a small want ad in the Post
which read “Tours to DPR Korea - Inquire Mr. Gpu.” So he in-
quired.

Over a long lunch of rice wine and kim chi, Mr. Gpu had told
Harry that of course Americans were welcome in North Korea. He
could issue a visa on the spot. And, by the way, did Harry know
perchance where to get large quantities of American cigarettes?

Harry had introduced Mr Gpu to a distributor in Hong Kong
and presumably somebody is now smuggling cigarettes into North
Korea, because Harry got his visa. I figured nothing ventured,
nothing gained, so I asked Harry to tell Mr. Gpu that I would pay
him a visit.

I was ushered into a paneled conference room and presented Mr.
GPU with a copy of STACKS, dutifully inscribed “To the Democratic

Exploring the Internet

People’s Republic of Korea, with best wishes.” I had never in-
scribed a book to a country before.

I explained the idea behind Exploring the Internet and my desire
to visit Kim Il Sung University in Pyongyang to see the status of
networking under the Great/Dear Leaders. I had heard that the
complete works of Kim I] Sung, the Great Leader, were available in
digital form and wished to meet the people behind the software.

Mr. Gpu abruptly got up and left the room. I wondered if he
Was going to get my visa. He returned a few minutes later with a
shopping list of equipment which included a strange device called a
“down converter” which, as best as I could make out, translated
radio waves from one frequency to another. I confessed I had no
idea what a down converter did, let alone why anybody would
want one. I had this funny feeling that I didn’t really want to know.

My ignorance appeared to have blown my credibility as a com-
puter expert. The interview quickly ended. No offers of rice wine
or kim chi. No visa.

I strolled back on the Avenida Do Doutor Rodrigo Rodrigues,
stopping on the way for a fine lunch of Galinka Piri-Piri, a grilled
chicken Macanese style smothered in spices.

The hydrofoil on the way back to Hong Kong was almost empty,
with no tourists taking snaps of the boat floor or pointing to each
island and asking “is that China?”

The tourists must have known something. We hit rough seas
and the hydrofoil started to resemble a rollercoaster ride. I closed
my eyes and hung on tight, finally arriving at the hydrofoil dock. I
collected my bags from Harry’s friends at the Macau Tourist Office
and headed across the bay to the airport.

Singapore

My plane arrived after 1 A.M. in the morning at Changi airport and I
steeled myself for waiting in long lines while customs officials woke
up and ambled over. To my amazement, from the time I came off
the plane to the time I got into a taxi, including clearing customs
and getting my bags, a total of seven minutes passed.

My taxi hurtled into town on a wide, immaculate freeway lined
with flowers and trees. We passed a Pizza Hut and a bowling alley.
This is Asia? Cars even pulled over to the side of the road to let an
ambulance through.

Getting to the hotel, I found that the five copies of STACKS sent
over by Interop had not arrived. I placed a call to David Brandin,
the Vice President for Programs, who assured me they had been
sent over a week ago by Federal Express.

The next morning, we found out that the books had in fact ar-
rived five days earlier, but had been intercepted by the Undesirable
Propagation Unit, one of the government censors. I knew that por-
nography and communism were not welcome, but I hadn’t figured
my descriptions of SMDS would offend anyone other than Frame
Relay manufacturers. After pointing out that the books were engi-
neering manuals, Federal Express eventually sprang them loose.

I spent the morning at the National University of Singapore
(NUS) talking to Dr. Thio Hoe Tong, director of the computer center.
NUS, with 17,000 students, is the largest university in Singapore.
The university had just installed a 64 kbps link to JVNCnet in
Princeton and joined the Internet.

The campus internet had recently caught the attention of the Na-
tional Science and Technology Board (NSTB), which decided that
the experiment should be expanded. Dr. Thio’s deputy, Tommi
Chen, had been delegated to develop TECHnet, a project with the
aim of developing a national research IP network.

Technically, TECHnet was simple. For SG $300 per month, an
organization got a port on the NUS router. Politically, however, it
was a big step forward. TECHnet involved both commercial and

Exploring the Internet

educational users. As long as the NSTB determined that a company
is a “bona fide” research and development group, it could join the
network.

This may seem a bit pale to some readers, but the idea of multi-
ple users sharing a single leased line (which is what TCP/IP does),
runs contrary to a long-established tenet of international communi-
cations, the infamous Recommendation D.1.

Recommendation D.1 of the CCITT established the principle that
multiplexing multiple users over a single international line should
be the province of the “network provider,” a code word signifying
the single monopoly, the PTT. In this parlance, a “value-added net-
work” was anybody who was forced to buy services from the tele-
phone company. By defining multiplexing over international lines
as the proper function of the network, not the value-added network,
competition never even got out of the gates.

Recommendation D.1 has been defanged for most of the world.
When you think about it, D.1 and similar barriers to entry run
counter to the Treaty of Rome, the GATT, and any other principles
of fair trade and antitrust ever promulgated in international law.
While most organizations and some countries ignore it, there are a
few places where D.1 still rears its ugly head. Although the recom-
mendation itself is rarely cited specifically, it has certainly estab-
lished itself as part of the telephone company mentality.

Rumors were that Singapore Telecom was not pleased by the
idea of TECHnet. After all, if a company uses the Internet to com-
municate with the U.S., the reasoning went, they wouldn't be plac-
ing phone calls.

This reasoning had been applied before to networks. In Singa-
pore, BITNET was not available to undergraduates. The rumor was
that Singapore Telecom objected because if undergraduates could
use electronic mail to reach brothers and sisters studying in the U.S.,
they wouldn't place international voice calls.

In the case of TECHnet, the government had exerted enough
pressure so that the system moved forward despite the objections of
more conservative elements. However, as a safety check, CEOs of
companies joining TECHnet had to sign a certification that the em-

Singapore

ployees of a company would only use the network for R&D, strenu-
ously avoiding commercialistic or even (save us) personal purposes.

After describing TECHnet, Dr. Thio explained the rather un-
usual approach to campus networking at NUS. A mere 18 months
before, several thousand PCs and Macintoshes had been isolated is-
lands with no connectivity. Rather than phase in networks over a
long period of time, NU.S. adopted a “big bang” theory of network
design.

An RFP was issued which was won by a consortium that in-
cluded Hewlett Packard and Fibronics. In two stages, six months
apart, 3,000 machines went on the network. Most of the network
was based on HP9000s and PS/2s acting as LAN Manager servers.
An FDDI backbone and Ethernet subnets tied it all together.

Needless to say, with several thousand users going live at the
same time, training and support became, as Dr. Thio put it, “an is-
sue.”

That night, I was scheduled to give a talk to the local IEEE chap-
ter. I gave my usual rambling discourse on “our friend, the In-
ternet.” Afterwards, I was presented with an attractive pewter dish
and we went out for a dinner of hairy crabs and garlic prawns.

I had agreed to be a guest lecturer Friday morning at the Informa-
tion Communication Institute of Singapore (ICIS). ICIS, a coopera-
tive program between AT&T and the Singapore government, offered
a postgraduate degree in telecommunications.

I was picked up by Don Stanwyck, an ICIS faculty member and
AT&T employee. Don had been one of AT&Ts representatives to
various CCITT and ANSI study groups and had served as editor for
several key ISDN recommendations.

Don took me down to Little India, where we had a breakfast of
roti prata, Indian fried bread which you dip into a bowl of spicy
mutton curry. Don explained how he had ended up on the expatri-
ate circuit in Singapore. As a CCITT delegate, he had logged well
over a million frequent flier miles, travelling 70 to 80 percent of the
time. He finally decided “enough is enough,” and told AT&T he

Exploring the Internet

wanted to live in Singapore. In addition to Singapore, Don had
lived in Japan, and speaks proficient Mandarin and UUCP.

With a love-hate relationship to Singapore, he was typical of
many of the expats who live in Asia. He knew more about Singa-
pore than many locals, yet remained very much a visitor here. After
two years, he had kept an intellectual detachment and alternated
between wild praise and cynical criticism.

At ICIS, I lectured to a packed room with several classes meet-
ing together. It was a typical lecture in some ways. The students
laughed at my jokes and seemed to enjoy the talk. Yet, when I was
done and asked for questions, not a single person raised their hand.

I had experienced this in the past, so was not as disconcerted as
I might have been. In many Asian countries, it is considered bad
manners to ask a question. To do so implies either that the teacher
did a bad job explaining or that the student wasn’t paying attention.
Risking either implication would be a loss of face for somebody.

Instead, questions are best asked in a more informal setting. Af-
ter the lecture, sitting in the lounge, I spent a half-hour answering
detailed questions on everything from namespace administration to
exhaustion of the 32-bit address space.

I spent Friday afternoon with Singapore Telecom. Singapore Tele-
com: is one of the classic PTTs, controlling every aspect of telecom-
munications, as well as the postal system. For many years, local
telephone service in Singapore had been essentially free, with a flat
annual equipment rental of SG $190 for homes and SG $250 for
businesses. (The Singapore dollar was trading at 1.694 to the U.S.
dollar at the time.)

When I arrived in Singapore, the flat rate was about to be abol-
ished in favor of a straight usage-based system of 1.4 cents per min-
ute during peak periods and 0.7 cents per minute during off-peak
periods. “Rentals” would decrease to SG $100 for homes and SG
$150 for businesses.

This change was causing such an uproar that the Prime Minister,
travelling in Zimbabwe, had called a press conference to discuss the

Singapore

subject, and his comments were front page news in the Singapore
Straits Times. For Singapore, this was a major political flap. A
newspaper columnist felt compelled to write a piece debunking the
theory that the entire social fabric would disintegrate because chil-
dren would place fewer calls to their mothers. Singapore Telecom
issued a study saying that 66 percent of bills would go down and
that the scheme would be revenue-neutral. Of course, recent year-
end results by Singapore Telecom of profits of SG $1.2 billion didn’t
help matters.

In order for a residential bill to go down, a residential customer
had to make less than 107 hours per year of peak, or 214 of off-peak,
calls. It boiled down to a situation where families with teenagers
and modems were going to face higher bills.

Singapore Telecom contended that the previous flat rate policies
had made it difficult to sell some new services. ISDN was a key
example. Singapore was quite proud in 1989 of being the first coun-
try to have nation-wide availability of ISDN.

ISDN cost 10 cents per minute in Singapore. With variable pric-
ing on regular phone lines, Singapore Telecom hoped to increase the
demand for data services. I asked Chew Mun Chun, manager of the
Network Technology Group at Singapore Telecom how successful
ISDN had been. He smiled bravely and said, with all the positive
spin he could muster, “today we have close to 300 subscribers.”
Needless to say, most businesses had found it simpler to use high-
speed modems on dialup lines.

While ISDN had not been a raging success, Singapore Telecom
had been more successful in other areas. It teamed up with British
Telecom and the Norwegian Telecom to offer Skyphone, a service
that allowed airplanes anywhere in the world to place calls. Each of
the three telephone companies provides a ground station that cov-
ered a certain geographic area: Norway covered the Indian Ocean,
Singapore the Pacific, and British Telecom the Atlantic.

Skyphone is built on top of a network of eight satellites run by
INMARSAT, a consortium headquartered in London and sponsored
by 60 member countries. INMARSAT provides a generic satellite
service which members then use to build value-added services such
as Skyphone. An airborne user communicating with an INMARSAT

Exploring the Internet

satellite indicates which service she is trying to use. When a call
reaches one of the earth stations, such as the one that Singapore
Telecom operates on Sentosa island, it is routed through the local
telephone network and then out to the International Direct Dial
(IDD) links. Since Skyphone is a telephone link like any other (al-
beit, an expensive one at U.S. $6.70 per minute), it can be used for
fax and data as well as voice traffic. It is even possible to place
ground-to-air calls, although the caller would need to know the air-
craft ID in order to do so.

The first customer for Skyphone was the Sultan of Brunei. Sin-
gapore Air was the first commercial airline to install the service.
Other airlines quickly followed suit.

Another interesting Singapore Telecom project I saw was Teleview, a
videotex system that went into field trials in 1988 and went public
in 1990. The government had very ambitious plans for the system,
but by late 1991 there were only 6,500 subscribers.

Teleview was an unusual videotex system in that it supported
high-quality photographic images. Commands to Teleview went in
via normal asynchronous dialup lines. Photos returned at 2,400 bps
would not be very effective, however, so a return data path was
provided via a television channel dedicated to Teleview.

Each frame on the television channel had a subscriber ID on it,
directing it to the appropriate system. The signal had a 4.5 Mbps
data throughput rate, allowing frames for lots of users to share the
single television channel.

Your average PC is not set up to handle either photographic
quality displays or to receive RF signals. To provide these capabili-
ties, the original Teleview terminal was a specially developed device
which could be seen in many shopping centers and public facilities.

Dedicated terminals, especially fairly expensive, single-purpose
terminals, were not an ideal way to penetrate the home market. PC
users were able to access Teleview by adding an RE card for receiv-
ing data, a modem for sending commands, and a photo display card

Singapore

to drive a monitor at 480 by 280 pixels with 18 bits per pixel for a
total of 256,000 simultaneous colors.

While photographic display was a unique feature of Teleview, it
was possible to bypass the feature and access the system simply
with a modem for an 80 by 40 character display. It appeared that
many, if not most, of the users were content with the character-only
displays and used the system for things like stock quotes. |

As videotex systems went, Teleview had a fairly good selection
of information available. In addition to the obligatory stock prices,
it also had the Reuters news service, airline information, real estate
listings, and even record and book shops.

Some of the more unusual applications reflected the influence of
a strong government involvement. Mandarin tutorials, for example,
proved quite popular as an instructional tool. Always sensitive to
the cultural makeup of the island, the government added Malay les-
sons soon after the Chinese.

The initial architecture for the system was a single proprietary
computer system with X.25 links to service providers and asynchro-
nous and RF links to users. The next generation system was an at-
tempt to vastly expand the functionality (or at least the capacity).
Teleview officials were still in the middle of the procurement proc-
ess and could not discuss details, but interviews with an anony-
mous source established that the contract for the system had been
won by Digital.

The front-end to the new network was going to be a series of
modem racks, presumably connected to VAXen, which are in turn
connected to a series of front-end Ethernets. Each front-end Eth-
ernet would then be connected, via a bridge, to an FDDI backbone.
Other bridges would connect to a back-end Ethernet. On the back-
end Ethernet would be another series of VAXen, which would con-
tain a cache of recently accessed frames.

The cache VAXen would be connected by a mystery net (pre-
sumably Ethernet and LAT) to a series of VAX Clusters, which
would serve as data repositories. Other VAXen would act as X.25
links to other information sources.

As one of my anonymous sources said, “they’re throwing VAXes
around like water.” Singapore had, as of October 1990, invested SG

Exploring the Internet

$50 million in the system and it was likely that current investment
was an order of magnitude higher.

The government placed great importance in Teleview. When it
was deployed commercially, Yeo Ning Hong, the Minister for Com-
munications and Information said “beginning today, Teleview will
bring the Information Age into every home and workplace.” Like
France, Singapore saw videotex as the way to bring widespread
computing into every home.

Saturday morning, I went back to Science Park and entered yet an-
other ultra-modern office building, this one housing the offices of
Singapore Network Services, developers of the Tradenet EDI system.

I was met by two individuals wearing the international market-
ing uniform of business suits, business cards, and business-like
smiles. They whisked me into the board room and sat me down at
a table the size of a golf course. They sat down far away on the
other side of the table, yet still close enough that I could see them.

I knew I was in trouble when the slide projector came out. The
junior of the team, bearing the title of “Executive Marketing” got up
and started smiling broadly and talking about “increasing produc-
tivity” and similar concepts that we could all agree on as being posi-
tive.

Singapore Network Services (SNS) was formed with very strong
government support as a joint venture between the Civil Aviation
Authority of Singapore, the Port of Singapore Authority, the tele-
phone company, and the government.

Although Tradenet, designed to clear cargo through customs by
using EDI, was the first application, SNS was branching out into
other EDI applications. The slide presentation was meant to illus-
trate a grand vision of the world, and words like “network hub”
and “information services” kept getting tossed out.

A bit of inquiry established the fact that the “network hub” was
in fact an IBM 9000 which traders accessed via dialup or leased
lines. Not being familiar with the IBM 9000, I naively asked what
operating system was being used.

Singapore

“That is a very technical question,” my presenter responded.
“We were not told you would want such technical information.”

It is the bane of marketing the world over that successful pro-
jects, such as a massive application of EDI in Tradenet, gets papered
over with superficial nonsense like “information services for pro-
ductivity.” Instead of simply describing the success and how it
came to be, which would certainly be very impressive, marketing
feels compelled to invent silly catch phrases.

Tradenet had in fact been quite a successful system. Traders
used the EDIFACT standard to format documents which were
uploaded to the IBM 9000. SNS didn’t use X.400 to transfer the
document, although Executive Marketing agreed with me that X.400
was a good thing. Once the files were uploaded OfficeVision mail
was used to move the documents between different groups.

The system linked traders in Singapore to customs, the Trade
Development Board (TDB), and to what was euphemistically re-
ferred to as “other controlling agencies.” An example of the latter
was the Undesirable Propagation Unit, the government agency that
had held my books for inspection.

Typical transactions were import/export declarations, certificates
of origin, and other documents. Many of these operations, such as
customs clearance for a shipment, were semi-automated. An an-
swer from customs on whether clearance is automatically granted or
an inspection is needed, could be returned in a few hours instead of
the few days required in a manual system.

Tradenet was launched in 1989 and at the time the government
had hoped that by 1993, 90 percent of the traders would be using
the system, at which time it would become mandatory. By 1990, 90
percent of traders were already on the system and it became manda-
tory in 1991.

After Tradenet, Singapore Network Services started branching
out into other EDI applications. Medinet was used for automated
processing of medical claims. $Link was an electronic funds trans-
fer application. OrderLink was a system developed to allow the
government to issue Requests for Quotation and receive bids from
suppliers.

Exploring the Internet

Executive Marketing finished his slides and I escaped the plush
confines of the Singapore Network Services board room into the re-
freshing humidity of the street. I flagged down a taxi, which man-
aged to decelerate from 50 mph to 0 in a little under 1 second. I
jumped in with a bit of trepidation and the tires squealed as he took
off.

“Where you from?” the cabbie asked.

“Colorado, U.S.A.,” I replied.

“Ah, Chicago!” he replied. He seemed to find this hilarious and
laughed furiously. After laughing for a few minutes, he would
quiet down, then repeat the word “Chicago” again and start laugh-
ing. I occupied myself with pressing my foot against an imaginary
brake pedal.

Saturday night, Michael Yap from the National Computer Board
picked me up for drinks. After a couple of beers in a generic pub in
one of Singapore’s huge entertainment developments, we went for a
ride and ended up in Chinatown.

There, we went into a traditional tea house, Tea Chapter. After
taking off our shoes, we walked up narrow stairs two stories into a
large room filled with low tables and pillows.

“This is a Singapore Chinese Yuppie joint,” Michael explained.
It had been started a few years ago by a few friends as a place peo-
ple could sit and relax. Around the room, people were sitting at
tables. At a corner, a young couple sat on cushions on the floor
while their two young children bounced off the walls.

Tea Chapter was part of a growing trend in Singapore to redis-
cover ethnic roots and to allow ethnic diversity. Originally, the
country pursued an aggressive melting pot strategy, but it had since
taken an about face. Malays, Indians, and Chinese all have distinct
ethnic communities in Singapore.

Tea arrived with an elaborate array of devices. A big pot of boil-
ing water was used to feed another tiny pot where the tea steeped
for a few minutes. The tea from the first batch was thrown away
and more water added. Then, a few batches from the tiny pot were

Singapore

used to fill a medium-sized serving pot. From there, the tea was
poured into teacups so small they held only a swallow. The tea was
accompanied by eggs which had been preserved in green tea leaves,
a delicious snack.

Along with the teapots was a container filled with mysterious
implements: miniature shovels, tongs, picks and the like. I asked
Michael to explain the purpose of all these devices.

“I’m not quite sure,” he confessed sheepishly.

Michael Yap is typical of the technocrats who run Singapore.
His education at the University of Maryland was paid for by the
government as part of an elite scholars program for fast-trackers. In
return, he is bonded for eight years of government service, although
it is likely that he will stay much longer.

Michael is passionately devoted to the joint causes of technology
and Singapore. He continually steers the conversation back to one
of the two topics (and preferably both at once), always accentuating
the positive aspects of Singapore society.

Michael's job is Programme Manager for IT2000, Singapore's ef-
fort to build an “intelligent island” by the year 2000. The scope of
IT2000 is vast, one of the most encompassing visions ever formu-
lated by a government to build a network infrastructure for a soci-
ety.

First and foremost, IT2000 is a symbolic commitment by the
government to use technology. It is an effort to find a way to main-
tain the stunning economic growth that Singapore has managed in
the last 20 years through shipping, finance, and hi-tech manufactur-
ing.

As a vision, it starts with high-level goals, such as increasing
productivity and well-being. At the next level down, it moves on to
the applications that can help achieve these goals. Only at the very
end is specific technology considered.

Thus, unlike many national plans, fiber to the home or use of
ISDN are not the goals. Instead, the vision of IT2000 is to maintain
productivity increases of 3 to 4 percent per year for a population of
2.65 million people for a sustainable period of time.

Singapore is one of the few countries that could seriously ad-
vance such a goal. Known as Singapore, Inc. to the financial press,

Exploring the Internet

the country is run much like a big corporation. It is not unusual for
senior management (i.e., government ministers) to give direction to
corporations, citizens, and the civil service.

Throughout its history, Singapore has been run with this combi-
nation of broad consensus and top-down direction. IT2000 is not
the first large scale plan. In 1984, there was “Vision 1999,” in 1986,
the “Agenda for Action.” In 1990, the government issued a beauti-
ful coffee-table book produced by Times Editions called Singapore—
the Next Lap. The book contained a series of broad goals, such as the
following:

“we must encourage more Singaporeans to
marry, to do so earlier, and to have three chil-
dren, and more if they can afford it.”

Pictures of smiling children and happy families were accompa-
nied by specific goals, in this case a target population of 4 million
people. These goals were then used to guide government policies,
such as tax policies, or even housing assignments.

In the area of information technology, which the Singapore gov-
ernment refers to as IT, policy was spearheaded by the National
Computer Board (NCB). The NCB coordinates government comput-
erization efforts, and is also charged with creating a “National IT
Plan.” In consensual Singapore, the NCB shuns the term “create,”
preferring instead to “lead to the realization” of a National IT Plan.

In the civil service computerization area, NCB had deployed
over 800 of its staffers into government offices. An example of a
system developed by the NCB is the immigration system, which
claims to be able to check passports for foreigners in 30 seconds and
residents in 15. My own experience at the airport, even though in
the middle of the night, certainly bore this out.

Other examples of government automation efforts are an auto-
mated card catalogue for the national library, a computerized regis-
try of companies, and automated production of birth certificates.
No one of these applications was particularly unique; what was in-
teresting was the breadth and focus of the computerization effort.

NCB had been slowly trying to tie these different systems to-
gether to allow data sharing among government bodies, an effort

Singapore

known as IDNet. I met with employees of NCB to learn more about
IDNet. The meeting took place in the vast NCB headquarters lo-
cated in Science Park. On the walls were posters of people looking
very modern, sporting slogans like “Success Through Partnership”
and “Fulfillment Through People Orientation.”

I was met by an NCB public relations officer and ushered into a
conference room. “Help yourself to coffee,” she said, pointing to a
device called the Café Bar and leaving to get the people I would be
meeting with.

The strange looking contraption had different buttons you
pushed to dispense various powders (coffee, milk, soup) into a cup.
You then pressed the water dispensing bar to put hot water into
your own personal mix of soluble powders. After a few false starts,
I figured out the apparatus and made myself a cup of fresh, steam-
ing Nescafé.

Being the curious type, I had the cover off the Café Bar and my
head buried inside when the door opened and a group of NCBians
(as the posters downstairs referred to employees) were ushered in
by the PR officer. I slammed the cover on my finger and sat down
to learn about IDNet.

The network is fairly simple. It consists of an IBM 3090 at the
Ministry of Finance acting as the hub for a multidomain SNA net-
work. The basic service allows a user on one net to send and re-
ceive 3270 data streams to another net. A few custom applications
are also available.

To use cross domain services requires a special ID. One thou-
sand IDs had been granted, each tied to a user profile detailing
which services that particular user could access. IDNet has 31 sites,
90 machines, and 40 applications.

The IDNet architects didn’t seem particularly adventuresome
technically, as is often the case with people whose job is to keep a
network running. They grudgingly acknowledged that token ring
networks might play a role in the future, but more for terminal ac-
cess based on PCs than for peer-to-peer networks. Protocols like
TCP/IP and OSI were acknowledged, but with a notable lack of en-
thusiasm.

Exploring the Internet

In a distinct twist from many large organizations, NCB was cen-
tralizing its servers into four very large data hubs, each devoted to a
particular type of data. Centralizing the location of computers
would, hopefully, save on overhead and permit better control over
operations. The four centers would have data about land, people,
legal establishments, and finance. While each data center might
have many computers, the NCB staffers didn’t rule out the possibil-
ity of replacing many systems with a single mega 3090 running DB2.

The guiding principle for computerization of government was
the slogan “one stop, non stop.” This meant that a citizen ought to
be able to go anyplace and do all government business at once (and
quickly). As one official pointed out, no matter how comfortable
you make the chairs in a government waiting room, people will still
want to leave. Of course, it is worth pointing out that reality does
not always meet the lofty goals of slogans, but it was encouraging to
see that Singapore policy makers had a policy they were trying to
communicate.

Data sharing among government agencies was one aspect of one
stop, non stop. The other was delivery of services. Systems like
Teleview videotex were envisioned as the delivery vehicle for many
government services. Some services were already available on
Teleview, such as listing of government tender notices and awards.

Other systems were not on Teleview but did show the effects of
data sharing. For example, the census bureau tapped into existing
databases and preprinted census forms, allowing people to verify
data instead of reentering it.

One stop, non stop was part of the broader vision of IT2000.
Singapore had had a national IT plan for several years, but in 1990,
Dr. Tay Eng Son, the Senior Minister of State for Education unveiled
the IT2000 concept which greatly increased the scope (and the
stakes).

IT2000 is being called a “national masterplan for an information
infrastructure.” When I talked to Michael Yap in late October 1991,
the NCB had just finished the first phase of feasibility studies.
Based on that first phase, NCB had received significant government
funds to develop an architecture to make it all happen.

Singapore

The feasibility phase divided the economy into 11 sectors, rang-
ing from construction to health care to transportation. Each sectoral
study group was chaired by a “captain of industry.”

NCB staffers assisted each study group. The staffers led the
groups through top-down analyses such as trends analysis and wish
lists and through bottom-up planning procedures such as depend-
ency graphs and functional decomposition.

The goal was a set of strategic applications: things that people
should be able to do on an intelligent island. Many of the projects
that fell out of this planning are truly ambitious, ranging from a
national multi-function smart card and an intelligent vehicle high-
way system, to a queueless, counterless government services sys-
tem.

Altogether, the planning process involved 200 industry leaders
and 50 NCB staffers. This resulted in the goals for IT2000. The next
step will be to begin defining a series of target applications. Under-
lying the applications will be a set of common network services
(e.g., authentication) which in turn use the national telecommunica-
tion network. ‘Tying all this together is a series of technical stand-
ards and a legal framework.

Defining each of these pieces is the challenge. When Michael
Yap talks about a legal framework, for example, he includes such
difficult issues as copyright for intellectual property. When he talks
of common network services he means everything from transport
connections to “look and feel” to application tool kits.

Will it work? When NCB decides an item is a priority, it can
achieve impressive results. For example, it developed a ship plan-
ning system for the Port of Singapore Authority that reduces the
planning time for each ship by half. Together with the massive use
of EDI in TradeNet, there is no doubt that the port is a highly effi-
cient, automated enterprise.

The real challenges will be the common network infrastructure
and the legal framework. In the U.S., such issues are still in the
very early conceptual stages. The Corporation for National Re-
search Initiatives has attempted to address both issues in their
Knowbot architecture, but Knowbots are still at the early prototype
stage.

Exploring the Internet

Michael and I finished our tea and he dropped me off at my
hotel. I was off the next day to Dublin to learn more about a key
U.S. infrastructure, the NSFNET, which had come into being be-
cause of an Irishman.

The next morning, I stood in long lines at the Singapore airport
and read the Singapore Straits Times. I noticed a story that had an
interesting mix of old Asia and high-tech. In the Taoist religion,
when people die, paper models of important things in their life are
burned in a ceremony, allowing the newly departed to take the good
things they enjoyed in life with them. Typically, these paper models
are things like houses and cars. On Sin Ming Drive, however, paper
models of portable phones, VCRs, and even karaoke sets were avail-
able. No ISDN handsets were on sale yet, though.

Dublin

I arrived in Dublin on Monday. I had heard that it was a holiday, so
I asked the cabbie what the occasion was.

“Nobody knows,” he replied.

And there goes your tip, I thought to myself.

Checking into my hotel, I caught a bus into town to find a true
Irish pub. I must have walked five miles before I found one that
was open.

“Never thought I’d have trouble finding an open pub in Dub-
lin,” I remarked to the bartender.

“Ah, it’s the holiday, you know,” he replied.

“What's the occasion?” I asked.

“Nobody knows,” he replied.

I left him a tip. Always tip your bartender is what my mother
told me as I left home, and the advice has never failed to serve me.

The next morning, feeling a combination of Guinness and jet lag,
I met Dennis Jennings, who picked me up in an ancient BMW and
drove me across the street to University College, Dublin, where he
runs the Computer Centre.

A tall, affable Irishman, Dennis set me down in his office, filled
with mountains of paper and wall-to-wall binders on a wide range
of networking topics.

“I understand you’re the father of NSFNET?” I asked.

He laughed.

“More like an uncle,” he said. For the rest of the morning he
proceeded to tell me the story of how the NSFNET was born.

When all is said and done, Jennings certainly qualifies as the
father (or at least one of several fathers) of the NSFNET. One must
ask, of course, if the child turned out as hoped, but that is a differ-
ent issue from its paternity.

Jennings started working on research networks in 1979 when he
started the Irish Universities Network linking the two major Dublin
schools. At the time, X.25 was considered a network and for two

Exploring the Internet

years, from 1979 to 1981, the Irish Universities Network consisted of
X.25 and “higher level” protocols such as X.28 and X.29.

In 1982, he proposed a national star-based network centered
around University College Dublin. After a year of discussion, nego-
tiation, and consideration, the Higher Education Authority funded
HEAnet.

The original HEAnet was based on the JANET Coloured Book
protocols. When I visited Jennings at the end of October 1991,
HEAnet 2 was beginning to be operational. Based on X.25 with a 64
kbps backbone and multiprotocol routers, the network was rapidly
becoming TCP/IP-based.

In 1983, with funding for HEAnet approved, Jennings heard a
rumor at Yale that IBM would be bringing BITNET to Europe. Al-
though the European Academic Research Network (EARN) was go-
ing to be based on the big six European players, Jennings managed
to turn that into seven countries.

The initial EARN meetings were chaotic and, as I found out over
the next couple of days, Dennis Jennings is not shy about grabbing
the white board and jumping in. He did, and ended up becoming
the first president of the EARN board.

The EARN position started bringing him to the continent on a
frequent basis. On one trip, he attended a Landweber networking
symposium. At the symposium, Jennings and his wife had dinner
with many of the participants. Jennings’ wife mentioned to Larry
Landweber that Dennis had always wanted to work in the States.
You know, one of those polite things you say over dinner.

“Would you like to be director of CSNET?” Landweber asked,
not one to let such offers pass idly by.

Dennis had a trip scheduled to California anyway. While he
was in California, he phoned Larry Landweber. Next thing Dennis
Jennings knew, he was flying to Washington, D.C. to meet with the
head of the CSNET board.

“By the way,” Larry added, “I’m also going to Washington. The
NSF is considering a new initiative. Want to sit in?”

Landing in Washington after a red-eye flight, Dennis took a cab
straight to the NSE, where he meekly took a seat in the back while
network luminaries like Dave Farber, Ira Fuchs, and Vint Cerf came

Dublin

in. “Frankly,” Dennis explained, “I was a bit awed. After all, this
was America.” Among some Europeans at the time, America had
the reputation as some sort of super technology haven.

He didn’t stay awed for too long. In fact, he soon found himself
at the white board again.

Eventually, he was offered two jobs: Director of CSNET and the
NSF Program Director in charge of networking. He played hard to
get for a bit, until NSF upped its offer to include considerably more
money.

Back in Dublin, he was sorting out his affairs when NSF called.

“Any chance you can be English?”

Asking an Irishman to be English is a bit like asking an IBM
account executive to sell camels.

“I beg your pardon?”

In order to work for NSE, you had to be a citizen of an allied
nation. Ireland didn’t cut it. Luckily, Dennis had been born in Eng-
land. He got an English passport, and on January 2, 1985 he started
work at NSF.

At the time, NSF wasn’t thinking in terms of building a national
network. It had funded four supercomputer centers and merely
wanted to give scientists access to them. The original thinking, in
fact, was to have four networks, one for each of the centers.

A technical advisory group was formed, chaired by Dave Farber
and including such people as the ubiquitous Larry Landweber, Tony
Lauck (head of the DECnet architecture group), Frank Kuo from
SRI, and several others.

The first big issue to tackle was standards. The computer scien-
tists quickly honed in on TCP/IP. Others were not so sure. At the
time, there were several higher-performance proprietary networks
such as DECnet and MFEnet (sardonically referred to by some as
MuffyNet).

The argument was framed in terms of openness versus effi-
ciency. Protocol suites like DECnet or MFEnet were thought to be
significantly more efficient than TCP/IP. Another issue was free-
dom of choice. Some supercomputer center directors didn’t want to
be constrained. Significant populations of users, especially in the
physics community, preferred DECnet.

Exploring the Internet

Eventually, TCP/IP won out. It was decided, however, that a
network could start out as something else as long as it would mi-
grate to TCP/IP. An interesting side note was the role of OSI. Eve-
ryone envisioned that OSI was five years away and would provide
a migration path to a world of openness, truth, and beauty. More
than five years later, OSI was still five years away and had become
the canonical example of the sliding window.

Even more fundamental a choice than the protocol suite was the
structure of the network. Jennings felt very strongly that this
should be a general purpose internetwork. General-purpose meant
that it was not specifically aimed at the current supercomputer user.

The argument was made that a general-purpose internetwork
would best serve the goal of promoting science and engineering, the
mandate of the NSF. The internetwork part of the structure was
fundamental: this would be a network of networks. The lowest unit
was the campus network.

From the present vantage point, this doesn’t sound like a radical
requirement, but at the time there were many major universities
without campus networks. As Jennings explained it, “networking
was the private domain of individual researchers.” By refusing to
extend the internetwork to the desk or the department, Jennings
hoped to help fuel an expansion of the internet.

In 1985, many of today’s mid-level networks got their start.
Some, like the San Diego Supercomputer Center’s SDSCnet, were
run by the centers. Others, like the Bay Area Regional Research net-
work (BARRnet) were independently formed.

Looking back, one thing that Jennings regrets is that the NSF did
not make a long-term funding plan explicit to these regional net-
works. Funding was based on a simple year-to-year allotment. A
better model would have told the regionals that they would be
funded at 100 percent for three years, with a two-year falloff. This
is the same model, of course, that IBM used with BITNET and
EARN.

Making the funding model explicit would have had two bene-
fits. First, NSF would have been freed from continuing funding ob-
ligations with the regionals, allowing NSF to move on to other
infrastructure projects. Second, and more important, the regionals

Dublin

would have had a very strong incentive to find alternative funding,
from their members, consortiums, or the states in which they did
business. The Jennings model of infrastructure is thus to have gov-
ernment start off the project and then have the users take it over.
This leaves the government free to move on to the next building
block.

In addition to funding the regionals, NSF ended up putting ina
backbone, the NSFNET. In August of 1985, Jennings attended a
two-day meeting in Boulder, Colorado with the directors of the four
supercomputer centers and the director of the National Center for
Atmospheric Research (NCAR).

Jennings presented his model of regional networks connected by
a national backbone. The directors were not pleased. After all, this
was a way for one of their local researchers to use the facilities of
another center.

The next morning, however, the directors did an about face.
True, their local researchers could use somebody else’s system, but
the converse was also true. An Internet would make one center’s
services available to the entire country. If you think your center is
better than the others (and all the directors did), a backbone is a
wonderful idea.

As Jennings related the story, not only did the directors change
their minds, they started pounding the table asking why this wasn’t
already in place. The group adjourned to a picnic table up on Table
Mesa, in the shadow of the Rocky Mountains, and sketched out the
backbone on a piece of paper. Jennings still has the piece of paper,
dated September 17, 1985. He had been on the job nine and a half
months.

The backbone would consist of 56 kbps lines linking NCAR and
the four centers. It was immediately labeled the “interim backbone”
since everyone agreed a T1 network would soon be needed.

Events were moving very quickly. At times, Jennings was cau-
tioned by the NSF bureaucracy that he was “sailing close to the
wind.” For example, Bill Schrader and others at Cornell went ahead
and ordered all the 56 kbps lines before NSF had formally approved
the project. Cornell then took the lead in putting in the interim sys-
tem.

Exploring the Internet

In addition to lines, there was a crucial decision on the selection
of routers. Major and minor router vendors were asked to make
presentations. BBN had a wonderful system, but their routers cost
U.S. $180,000 each. Cisco and Proteon came close, but didn’t have
nationwide support structures in place.

Ultimately, attention focused on the Fuzzball, a device based on
the PDP-11 and developed by David Mills. Everybody was reluc-
tant to depend on a homegrown system developed and maintained
by one person. As reluctant as anybody to have the national net-
work based on the Fuzzball was David Mills himself.

David ended up spending the next two years maintaining and
modifying his Fuzzball. Some of the changes were simple bug fixes,
but others reflected some significant architectural flaws in the origi-
nal TCP/IP.

For example, the TCP/IP protocol suite implied a single back-
bone: the ARPANET. Now, with the addition of the NSFNET, there
were two backbones and the Exterior Gateway Protocol (EGP)
broke. Mills was forced to periodically hack up his Fuzzballs to
handle this, and other, situations.

Jennings felt that the decision to go with the Fuzzball was one of
the major successes of the NSFNET project. It forced the Internet
community into a formal examination of routing, leading to specifi-
cations of router requirements and the development of new proto-
cols such as the Border Gateway Protocol (BGP).

Common, open specifications for routing in turn led to the im-
mense growth of companies like Cisco, Proteon, and Wellfleet, as
well as substantial product lines for larger companies such as DEC
and Sun Microsystems. Low-cost routers allowed the creation of ad
hoc internetworks, which in turn led to the creation of companies
like PSI and AlterNet.

In March 1986, Dennis Jennings brought to a.close his 15 months
at NSF. Just 15 days later, the 56 kbps backbone went live. Jennings
was replaced at NSF by Steve Wolff, who began the difficult job of
formulating the solicitation for the T1 backbone and then later, a T3
backbone.

If Dennis Jennings is the uncle of NSFNET, then Steve Wolff
would qualify as the nanny, expanding the network from an ad hoc,

Dublin

interim system into a mature, production network transferring over
14 billion packets per month. Wolff deserves a great deal of the
credit for the rapid growth of the Internet by agressively expanding
the backbone, funding international links, and encouraging the
growth of the regional and campus networks.

After a morning learning the history of NSFNET and a chance to
view the tattered piece of paper that served as the original architec-
ture document, we went out to lunch with the directors of the other
two university computer centers. Over a glass of wine, the conver-
sation turned to the question of European backbones. How would
Dennis, if he had the power, accomplish the same thing in Europe
that he did in the U.S.?

Dennis quoted a classic Irish proverb given when a tourist asks
a farmer for directions: “I don’t know, but if I were going there, I
certainly wouldn't be starting from here.”

In the United States, a government agency, the NSF, was charged
with promoting research and education. Within the European Com-
mission, however, there was no such group. There was no mention
of education in the Treaty of Rome, the enabling legislation for the
European Community. Instead, funding for research networks was
funnelled through groups like the infamous 13th directorate.

DG XIII was responsible for promoting things like competitive-
ness and standards. In practice, this led to a strong focus on the
PTTIs and OSI. What it did not lead to was an operational, pan-
European backbone, either commercial or research. Everybody was
waiting for the Commission to act and the Commission was waiting
for OSI to be commercially available.

After lunch, Dennis told me about an extraordinary meeting that
was to be held the next day to try and bootstrap a European Back-
bone (EBONE). We quickly rearranged my plane tickets to route me
to Amsterdam and Dennis drove me back to my hotel for a glass of
Guiness.

Amsterdam

My wakeup call came at 5 AM. I found my way into a cab, met
Dennis at the airport, and flew to Amsterdam. We hailed a cab to
the EBONE meeting being held in the Tiger Room at the Artis Zoo.
En route, our cabbie regaled us with obscene animal jokes.

When we arrived at the Tiger Room, the meeting had started
already, with 35 deadly serious people sitting around a large square
formed by long tables. Kees Neggers, co-managing director of
SURFnet, chaired the meeting, which included most of the major
networking players in Europe.

The idea behind EBONE was quite simple. Until the Commis-
sion got its act together, people would pool existing resources and
form a voluntary, interim consortium. If and when the EC or some
other body started operating a more formal backbone, the consor-
tium would disband.

A consortium like this was an incredible balancing act. Deep
religious barriers divide Europe. Some people insisted that any
workable network must be based on OSI, the Connection Oriented
Network Service (CONS), and X.25. Others prefered TCP/IP and
leased lines. To make a workable backbone, enough of these players
had to be convinced to sign up to the EBONE concept.

Kees Neggers and a few others had formulated a memorandum
of understanding which would form the “gentlemen’s agreement”
to put EBONE into being (the word “gentlemen” being appropriate
for this almost exclusively male enclave). The gentlemen were
walking through the draft word by word.

The basic concept behind EBONE was that all organizations
signing the memorandum would make some form of contribution in
return for being able to use the backbone. Some would contribute
lines or routers, others would contribute people, and a few would
even contribute money.

All the resources would be a loan to EBONE. At the end of
1992, everything (except the money, of course) would revert back to
the contributors. The hope was that by the beginning of 1993 some

Amsterdam

formal group would be in place to handle the pan-European back-
bone and EBONE would be unnecessary.

One of the biggest items of debate was who would or should be
the parent organization that would replace EBONE in 1993. Some
wanted the Réseaux Associés pour la Recherche Européenne
(RARE), the association of research networking groups in Europe, to
play a role. Others wanted no such thing. Kees Neggers continu-
ally refocused the discussion on 1992, emphasizing that EBONE was
making no decisions about anything past the one-year interim ad
hoc network.

The EBONE backbone, like the NSFNET, would have no directly
attached users. It would be connected to regional systems, such as
SURFnet, which in turn would be connected to users. The backbone
would be a diamond, connecting four major European cities. On at
least two points on the diamond there would be links to the U.S. (A
few months later, the initial 1992 topology was finally determined to
be five major hubs—London, Montpellier, CERN, Amsterdam, and
Stockholm—connected with 256 kbps to 512 kbps links and with
three links to the U.S.)

The network, unlike NSFNET, would have no restrictions on the
content of traffic. Although targeted for the benefit of academic and
research use, commercial traffic could pass over the backbone and it
was up to each regional network to determine an appropriate use
policy. Likewise, although EBONE was envisioned primarily as a
TCP/IP backbone, it was explicitly multiprotocol. An OSI “pilot
service’ was listed in the draft memo as one type of traffic, and
others were alluded to.

Much of the wording of the memo was quite delicate. For ex-
ample, EBONE was described as providing “value-added open net-
working services.” Why these particular words?

Well, networks are the domain of PTTs in Europe. If you leave
out the word “value-added,” you tread on the turf of the PTT. Like-
wise with the word “open.” Many countries had decided that open
networks (i.e., OSI) were crucial and they would not participate in
EBONE unless the word open was used.

Other types of wordsmithing at the meeting reflected the cul-
tural differences of the 35 participants. For example, the backbone

Exploring the Internet

had been described as being “redundant.” In England, when you
fire somebody, you “make them redundant.” “Resilient” proved to
be a better word.

Lunch at the zoo consisted of frantic huddles among various fac-
tions. While I enjoyed a delightful buffet of eel and paté, others
were desperately trying to forge a consensus.

Issues like the role of DECnet proved to be especially tough.
Some people wanted to allow DECnet Phase IV traffic across the
backbone. Others said that this wasn’t necessary as DECnet Phase
V would be able to use the ISO CLNS service. It was finally de-
cided not to decide. DECnet traffic could cross the backbone if the
two ends of a link decided it was allowed and it didn’t adversely
impact the operation of the backbone.

Carrying DECnet traffic was in some ways a foregone conclu-
sion. Some of the links being contributed to EBONE were multi-
plexed into underlying lines. These underlying lines were not an
EBONE issue and could carry DECnet, SNA, and any other kinds of
traffic.

EBONE was clearly a chicken-and-egg situation. Many organi-
zations would be unable to persuade their management to join until
the consortium existed. The consortium wouldn’t exist until enough
organizations joined.

At the end of the day, with all the words finalized, Kees Neggers
went around the table and asked people to tell the group if they
would join, and, if so, what their contribution would be. To start
things off, SURFnet would donate ECU 300,000 (U.S. $369,000)
worth of lines, routers, and people.

Others joined in. The Spanish network would help the line costs
of the backbone and would donate 35 percent of a person. Telecom
Netherlands, if the EC approved, would provide a gateway to IXI,
the large European X.25 network funded by the Commission.

Dennis Jennings at UCD would use the IXI links to reach
EBONE, as would many of the smaller countries. Dennis also had
an unusual contribution of money. He offered ECU 10 per full time
academic staff member at University College Dublin, for a total of
ECU 8,000 (U.S. $9,840). Not a huge sum, but Dennis pointed out

Amsterdam

that if every university in Europe adopted his formula, the network
would have a budget of ECU 10 million per year.

Many groups were unable to commit immediately. Brian Car-
penter, head of networking at CERN, the international physics labo-
ratory, had to get approval from all his member countries and
would ultimately end up “supporting” EBONE instead of formally
signing the memorandum. The European BITNET, EARN, wanted
to commit a line to CUNY in New York, but had to wait until its
board met to commit formally.

On the opposite side of the room from Kees Neggers was Harry
Clasper, the representative from IBM. IBM ran EASInet, a large
European network. Its contribution of lines would be a significant
addition to EBONE.

Throughout the day, Harry Clasper had been a vocal participant.
Dressed in a regulation IBM pinstripe suit, he had been carefully
tracking the proceedings. When the question of contributions
reached him, he began speaking in a quiet voice and the room
strained to hear him.

“ Y'm disappointed that this is not as open as it could be, and
therefore IBM Europe is reserving its position at this time.” IBM
had substantial SNA traffic in EASInet and was clearly worried that
if they joined the EASInet, sites might have problems.

Even with IBM sitting on the fence, however, it was clear that
there was enough to make EBONE work. After IBM, groups like
EUnet and NORDUnet added substantial contributions.

Finally, the end of the table was reached. The last person to
speak was Bernhard Stockman from NORDUnet, also an active par-
ticipant in the IETF. NORDUnet had already made known its con-
tribution, but Bernhard had one thing to add.

“I will take everything you contribute and turn it into something
that works.” Bernhard would help form the Ebone Action Team,
the engineers who would try to rationalize the contributions into a
network.

With EBONE becoming a virtual reality (it would still take sev-
eral months to get all the pieces in order), Kees Neggers made one
more round around the table to see if anybody would be interested

Exploring the Internet

in being on the management committee. Harry Clasper from IBM
raised his hand.

“If IBM were to commit, I would like to be on the management
committee.”

Everybody laughed and the meeting was adjourned. A half-
dozen of us took a tram to the railway station to have a few beers
and conduct a post mortem. Kees Neggers, Harry Clasper, Dennis
Jennings, and myself sat with some beers while Kees and Dennis
worked on the IBM position. Harry Clasper, like some pin-striped
corporate Buddha, sat quietly and drank his beer.

Finally, Kees left to catch his train to Belgium and Harry, Dennis
and I all boarded the express train to the Schiphol airport. We were
cutting it a bit close, but the 18:26 train left us just enough time.

“Don’t worry,” Dennis reassured us, “you can set your watch by
the Dutch trains.”

I looked out the window as 18:26 came and went. Finally, ten
minutes later, an announcement came across the intercom in Dutch
and people start gathering things and got off the train.

A mob of 100 or so people all sprinted over to another track to
catch the local train. On the local, we began doing mental arithme-
tic. The conversation quickly went from European infrastructure to
travel horror stories then to nervous silence.

Arriving at the airport train station at 7:10 for 7:30 flights, the
three of us took off in three separate directions, each of us doing an
O.J. Simpson imitation. Despite an airline clerk who gave new
meaning to the term lethargic (although that was certainly not the
word I used at the time), I made it to my plane for London.

London

Thursday morning I strolled from my seedy hotel room through
soot-blackened streets to University College London to meet Steve
Hardcastle-Kille. Steve is one of the two guiding lights behind the
ISO Development Environment (ISODE), the other of course being
the eminent Dr. Marshall T. Rose.

While Marshall concentrated on making a public-domain OSI
implementation of the middle layers that can run over both TCP/IP
and OSI stacks, Steve focused on the application layer. Through a
prodigious string of Internet Drafts and RFCs, Steve helped to turn
the X.400 and X.500 standards into workable services.

It is somewhat ironic that some of the best work on ISO stand-
ards (and the most popular implementation) should have come out
of the TCP/IP world. By providing public-domain OSI code, Hard-
castle-Kille and Rose had pushed these standards from paper and
theory to things that people use in their day-to-day work.

If you want a standard that works in the real world, it needs to
interact with the existing base of services. Some of the early work
Steve did was to define the functionality for X.400/SMTP gateways.
Later, Steve focused on X.500.

UCL, under Steve’s leadership, integrated X.500 into ISODE in
an implementation known as Quipu, after the fringes of knotted
cords used to keep numerical records by the Quechuan Indians of
the Incan empire. It was traditional to name each country’s master
Directory Service Agent after an animal, preferably one found in
South America. Finland’s DSA is the jaguar, Germany the puma,
and Marshall T. Rose contributed the alpaca for the U.S. In a his-
toric decision, however, as a tribute to X.500’s growing maturity,
South American animal names were abandoned in favor of moni-
kers marketing could pronounce.

X.500 directories are structured as a tree. Each part of the tree is
managed by a DSA (although it is important to note that the direc-
tory and the DSA are carefully defined as separate concepts allow-
ing a particular part of the tree to be provided by multiple vendors).

Exploring the Internet

The root for the world is the Giant Tortoise at UCL. In addition to
being the world root, this DSA is also the root DSA for the United
Kingdom.

The United Kingdom had been aggressively deploying X.500.
The Joint Network Team (JNT) began funding X.500 by giving uni-
versities a Sun/4 to act as a DSA. By November 1991, 40 organiza-
tions in the U.K. had DSAs serving them with a total of 54,387
registered entries. By the end of 1991, all 55 universities in the U.K.
were scheduled to have their Sun workstations.

UCL also functions as the coordinator for the PARADISE ES-
PRIT project. PARADISE stands for Piloting an International Direc-
tory Service. (A cute acronym is one of the prerequisites for
European Commission funding.)

Under the coordination of PARADISE in Europe and similar
projects in the U.S., the global directory had grown by November
1991 to reach 1,212 organizations, 144 DSAs, and 421,552 directory
entries.

The DSA is the X.500 component responsible for some part of
the directory tree. To access X.500, users have a Directory User
Agent (DUA). A couple of dozen user interfaces have been defined
to interact with Quipu, ranging from a simple white pages lookup
utility (FRED) to a full-fledged management interface (DISH) to
Macintosh and X-Windows based graphical interfaces.

To make X.500 work in the real world, several extensions were
needed from the standards as originally defined. The X.500 specifi-
cation assumes a homogenous OSI-based network where any DUA
can set up an application layer association over a worldwide net-
work to reach any DSA.

In many places, a homogenous network is not realistic. The
U.K., for example, had long supported the Coloured Book protocols,
although TCP/IP support had recently been added. Many ISODE
implementations run on top of TCP/IP, adding another important
environment. It is interesting to note that many places that use
ISODE on top of TCP/IP still have as an official policy a “migra-
tion” to true OSI. They have yet to learn that birds migrate, not
corporations.

London

One of the first requirements for the deployment of X.500 was a
relaying mechanism. If a DUA in the TCP/IP world needs to reach
a DSA in the OSI world, it can ask a DSA connected to both worlds
to relay the request.

A second key area addressed was replication. X.500 assumes
one DSA is responsible for one piece of data. If a DUA needs to
reach a DSA in another country, it is possible that many intermedi-
ate DSAs must be contacted as the user climbs the name tree to the
root and then back down towards the target (although an intelligent
DSA implementation would presumably cache some of this infor-
mation). Replicating some of that upper-level information makes it
much easier to find target organizations and is a significant per-
formance enhancement.

In addition to replication and relaying, Steve has been especially
active in helping to forge a consensus on a common schema for the
directory. After all, it doesn’t do much good to find an organization
if the information it keeps is non-standard. A common schema con-
tains the definition for standard objects, such as a person, and
standard attributes common to those objects, such as a person's fa-
vorite drink or e-mail address.

Just before I arrived in London, the Internet Activities board had
issued a carefully drafted RFC endorsing X.500 as a strategic direc-
tion for the Internet community. If that stragtegy took hold, it
would be a significant shift away from the Domain Name System
(DNS).

The transition between older name systems, such as DNS and
the venerable WHOIS service, would certainly be the key to the suc-
cess or failure of X.500 on the Internet. DNS adherents cite two
problems with X.500: the complexity of the namespace and the
slowness of implementations.

X.500 structures names in a well-defined tree, with objects typi-
cally going from country to management domain to organization to
organizational unit to name. Some feel that a rigid hierarchy is a
key flaw. Steve argues strongly that the opposite is true.

One rigid hierarchy is indeed the basis for X.500; however, alter-
native hierarchies can be defined on a local or regional basis with
pointers into the basic tree. An alternative tree, based on the Do-

Exploring the Internet

main Name System, for example, can be used to point to X.500 ob-
jects.

A single rigid hierarchy does have some advantages, particularly
in the area of management. With a well-defined schema, people
know exactly where to put an object, making management of the
namespace similar in different areas.

Before X.500 can operate as a truly global directory, it needs
much better performance. One of the major problems for perform-
ance is that X.500 implementations must carry around the baggage
of the fully general middle layers of OSI. Although a skinny stack
has been defined, most X.500 implementations support all the fea-
tures of the session, presentation, and ACSE services, including
those developed for other networking paradigms such as transac-
tions processing. For example, X.500 makes no use of the check-
pointing and synchronization features of the session layer.

A light-weight Directory Access Protocol could go a long way
towards providing faster DSAs. The lightweight protocol would
map OSI down to a needed subset. If you need full generality, as in
the case of communicating with a DUA or DSA that doesn’t support
the lightweight protocol, a separate process can provide the transla-
tion. |

After spending a morning with Steve, I started making the
rounds at UCL. UCL is highly unusual among computer science
departments in that it emphasizes networks as a valid area of re-
search. In fact, networking takes prominence at UCL (the school
was the first international member of the ARPANET).

UCL’s prominence is due in no small part to the influence of the
department chairman, Peter T. Kirstein, known around the world by
his login name of PTK. An old joke in the networking community is
that when in Europe you have to deal with the PTTs, but in Eng-
land, you have to deal with the PTK.

No visit to UCL would be complete without a courtesy call to
Peter, so I left Steve’s office at one end of the building to pay my
visit to Peter Kirstein. Peter’s current passion (and a source of sub-
stantial funding) was the Office Document Architecture (ODA), so
we spent a half-hour talking about his testing of ODA packages, and
his cooperative projects with groups like Bellcore and the American

London

Chemical Abstracts to convert large databases into ODA to help
spur the standard forward.

UCL is also active in many other areas of networking. Jon
Crowcroft described his efforts to run video conferencing over the
Internet. One researcher showed me his work on ISDN Primary
Rate interfaces; another showed me new X.500 user interfaces run-
ning on soot-blackened terminals.

With my head swimming from networking vertigo, I left UCL to
get a drink. I stepped into an old pub near the university, with
some appropriately quaint name like The Queen’s Foot or the Tam
and Mutton. Settling down with my pint of bitter, I looked up to
see that a karaoke competition was scheduled to start soon. Feeling
that I could miss this particular event, I finished my pint in record
time and went off for some Indian food.

vaimpene

Friday morning, I headed north to Tampere in Finland. With three
hours to kill in the Helsinki airport before my plane left, I idled
away the time trying to decide between the sauteed reindeer with
lingonberries and the elk pate with rowanberry gelatin. It was 4 PM.
and the sun had just set.

I was met at the Tampere airport by Vesa Parkkari, a man I often
refer to as the Finnish Bill Gates. Of course, Vesa has a much more
effusive personality than Bill Gates, although I confess my only con-
tact with Mr. Bill has been through interviews in the trade press,
which could make anybody look boring.

The next morning, I got up early enough to watch the sun rise at
7:30, then had a leisurely breakfast of pickled herring and brown
bread. It was Saturday morning and evidently some sort of holiday
in Finland. Of course, nobody had any idea what the holiday was
for, but it was enough to close most of the shops in town.

It wasn’t enough, though, to slow down Vesa. At 10 AM., he
brought me down to the company he founded, Mikrokonsultit Oy.
The name is a bit of a misnomer, as they didn’t have anything to do
with microcomputers and didn’t do any consulting. (Since my visit,
they changed the name of the company to Relevantum Oy.)

In the late 1970s, however, Vesa was heavily involved with both
microcomputers and consulting. He took early Motorola chips,
made them into microcomputers, and then used the systems to de-
sign custom applications. One system was developed by Vesa for
NESTE, the state monopoly oil company. Involving 2 square meters
of electronics (and lots and lots of assembly language programs), the
system coordinated the weighing of all lorries coming into the refin-
ery to load oil, printing all the necessary lists and other documents.

Another, system, developed by Raimo Harju, his copartner, was
still being used. The system takes raw timber and figures out the
most efficient way to saw up the log to produce the most lumber.
In an economy where 30 to 40 percent of GNP is based on forestry,
this system proved quite popular.

100

Tampere

Vesa had the same problem with consulting that others all over
the world face. He did the programming, and thus bore the risk
when a job took an order of magnitude more time to complete than
planned. A fact of consulting, however, is that you often end up
several subcontracts deep in order to get a job. Vesa bore the risks,
but others higher in the food chain were making all the money.

Vesa decided instead to give seminars about putting microcom-
puters together. He gave a few at the Tampere Institute of Technol-
ogy, and the demand quickly overwhelmed the university facilities.
This was obviously a worthwhile business to be in.

Vesa and his colleague Raimo Harju went into business. At first
there was no company, just a couple of guys in a hotel conference
room. The hotel couldn’t handle the idea of a meeting without a
company name, so they made one up: Mikroboys.

Mikroboys became the more dignified Mikrokonsultit and busi-
ness flourished. In the early 80s, however, the prepackaged PC
came into being and Vesa saw that there would be a limited future
for very low-level courses like his. In 1982, they switched the com-
pany to an exclusive focus on data communications seminars.

Seminars are a funny business. Almost everything is a fixed
cost. You spend a month or two (or more) to develop the seminar.
Then you print and mail brochures. Then you wait.

If enough people sign up, usually something like 15 to 20, you
break even. Everything after that is gravy. Well, in 1983, the gravy
began for Vesa. Finland had just started up two different public
X.25 networks and Mikrokonsultit had a hands-on practical X.25
course. They began teaching seminars to 70 people or more on a
regular basis.

Based on the early success, the company expanded and now has
13 different seminars. Finland, with a population of 5 million peo-
ple, is a small country. Yet it is big enough to support Mikrokonsul-
tit.

There are two keys to the early X.25 seminar success, and both
factors hold for the later courses. First, courses are based on tech-
nology that is available and cost effective. In the X.25 realm, for
example, Finland has two public data networks in heavy competi-
tion with each other supporting over 3,000 hosts. Leased lines are

101

Exploring the Internet

cheap enough in Finland that it is estimated that there are another
3,000 hosts on private X.25 networks.

The second key to Mikrokonsultit’s success is that the seminars
are hands-on. In a 5-day X.25 course, for example, users not only
learn the theory of X.25 but how to configure Packet Assem-
bler/Disassemblers (PADs) and hosts and how to run applications.

Vesa took me through some of the exercises that the instructor
does in the X.25 class. We took a PC X.25 card, installed it, config-
ured it, then loaded up the X.25 and PAD software. We then set up
links to Russia, China, the U.S., and the United Arab Emirates. We
looked at a network analyzer to analyze PAD and X.25 parameters.
We looked at tariff tables and calculated breakeven points for leased
lines versus packet networks. We ran SNA 3270 and X.400 over
X.25 to real hosts.

Other courses on subjects such as X.400, TCP/IP, internetworks
and routers, and SNA all take the same hands-on approach. Drag-
ging Cisco routers, PCs, analyzers, and other equipment along
means that you can’t just slap an instructor on a plane and have him
read the words on slides to the class. You need people who can
both teach and use the technology and, in the case of Mikrokonsul-
tit, are fluent in Finnish. |

All day Vesa led me from one room to another and from one
topic to another.

“Now we will do X.25 demonstration.”

“Now we will discuss DECnet Phase V.”

“Now you can use the Internet to read your mail.”

At the end of the day, with my eyes glazing over, he had a new
assignment.

“Now we will go to my castle and do Finnish sauna.”

The training market had been kind to Vesa and he had recently
bought a castle. Much as I wanted to see this castle, I was torn. I
had visions of being thrown naked out of a scalding sauna into a
moat filled with ice.

We arrived at the Parkkari castle, and Vesa showed me around
from room to room. While his wife Pirjo Sistola and their three year
old daughter took their sauna, we polished off a bottle of cham-
pagne. Finally, the fateful moment arrived.

102

Tampere

“Now we will do Finnish sauna.”

I tried to demur, but Vesa made it quite clear that failing to par-
ticipate would be considered a grave form of national insult. After
a shower, I entered a very, very hot room.

Vesa picked up a bucket and threw ladles of water onto the
stones, producing blasts of steam. What I had thought of as very,
very hot became an order of magnitude hotter. While Vesa kept
throwing ladles of water onto the rocks, he explained the seven lev-
els of the Finnish Sauna Reference Model.

I was experiencing level 1. Level 2 seemed to have something to
do with more water on the coals. In level 3, you exit the sauna
periodically and rub ice cubes over your body, then jump back in.
Level 4 consists of whipping yourself with strips of birch bark to
stimulate the skin. In level 5, you leave the sauna and run outside
to the lake. Level 6 has you roll around in the snow. Level 7 was
some form of Finnish national secret and, quite frankly, I didn’t re-
ally feel I needed to know.

Sunday morning was spent discussing the Internet with Vesa’s staff.
We speculated on the motives of ANS and the fate of the NREN,
after which I spent the afternoon taking a walk around the Tampere
lake.

That evening, I was met by Juha Heindnen, one of the creators of
the Finnish Internet and an important technical influence for the
whole Nordic region. Juha took me to “Salud,” a Spanish restau-
rant, where we feasted on a carpaccio of wild boar and an entree of
alligator served with a coconut and papaya sauce.

Juha began setting up UUCP links in Finland in 1982. At the
time, Finland had a large X.25 network for connecting the Univer-
sity VAXen, and UUCP was used to transfer mail and news over
phone lines within the computer science community.

In 1984, having finished his doctorate, he went to the University
of Southwestern Louisiana for a postdoctoral fellowship. While
there, he helped set up a TCP/IP network for Louisiana and, as he
tells it, enjoyed many Cajun parties.

103

Exploring the Internet

Coming back to Tampere, he spent four years as an Associate
Professor at the Tampere University of Technology. At the Univer-
sity, he was asked to join the steering committee for FUNET, the
Finnish University Network.

There was a problem, though. FUNET was basically DECnet
running on X.25 and Juha had a Sun workstation. He persuaded
FUNET that their role should be to connect LANs together, not
VAXen.

At the time, there were no multiprotocol routers, so FUNET be-
came a network based on wide-area bridges. A year later, however,
Juha went to a USENIX conference and saw a couple of guys stand-
ing behind a small table.

This was the Cisco booth and the company had just added DEC-
net support to their boxes. Juha came back to Finland ready to buy
Cisco boxes, but couldn’t find a distributor. He convinced a sales-
man at a local company to start carrying Cisco and two months later
FUNET had become a multiprotocol network based on routers.

After his success with the University network, Juha became a
consultant. His clients include FUNET, as well as Telecom Finland,
the country’s nominal PTT.

Finland, unlike most European countries, has considerable com-
petition in telecommunications. At one point, in fact, there were
over 700 telephone companies. In 1991, there were over 50, each
serving a small area. In addition, there was Telecom Finland.

Telecom Finland was state owned when I visited, but was soon
to be incorporated. It had two functions. It served as the local tele-
phone company in areas where there are no others. It also func-
tioned as the long distance telephone company. Even in the long
distance market, it already had competition in data and would soon
have so for voice traffic.

In all cases, Telecom Finland received no special treatment and
competition was encouraged by the Ministry. Data networking was
especially competitive. Finland had two commercial IP networks:
Datanet, run by Telecom Finland, and Lanlink, run by a consortium
of the smaller companies.

Both Datanet and Lanlink had connections to FUNET. FUNET,
in turn, was part of NORDUnet, the regional academic network

104

Tampere

linking five Nordic countries. Datanet had another link into Infolan,
an international IP service owned by a consortium of PTTs. Infolan,
in turn, was connected to PSI’s commercial network in the U.S.

All told, Finland had over 10,000 Internet hosts. This was 10
percent of all European hosts on the Internet, far in excess of Fin-
land’s proportion of the European population.

Both Datanet and Lanlink differed from their commercial Ameri-
can cousins in a very important respect. PSInet, CERFnet, and other
U.S. networks were open networks. When you joined, you could
talk to any other IP host. Of course, that host might not want to talk
to you, but the backbone delivered packets up to the customer
router or host.

In the Finnish commercial networks, access was explicitly con-
trolled in the backbone. When a customer started up service they
explicitly identified a set of hosts. Routers at the edge of the back-
bone enforced the policy with access control lists.

Datanet and Lanlink were thus more than just IP networks.
They actually formed a sort of multiprotocol, LAN interconnection
service. When I went to talk to Juha, Datanet was in transition. It
was just beginning to provide the generic LAN interconnection serv-
ice in a different way.

The old model consisted of a border router at the edge of the
backbone, which in turn was connected to a router at the customer
site. The backbone itself consisted of a large number of E1 lines
running at 2 Mbps. As Juha explained it, “no need to scrimp on
backbone capacity. After all, we are the telephone company.”

The new model of Datanet was based on Frame Relay. The
Cisco router at the customer site, equipped with a Frame Relay in-
terface, established a permanent virtual circuit (PVC) to another
router on the edge of the network. If a user wanted to send TCP/IP
traffic onto the Internet, for example, a permanent virtual circuit
would be set up to either the FUNET or Infolan router. Traffic from
Novell, DECnet, or any other protocol could use the same PVC to
send data.

Juha saw a number of advantages in moving the public network
down from the network layer into a level 2 data-link service. First,
bridges can be supported on a multi-user network. Bridges broad-

105

Exploring the Internet

cast all packets to the other side. Since Datanet supports multiple
private virtual networks, in the old model there was always the
question of where to broadcast the packets. With permanent virtual
circuits, the customer answers that question.

Subnetworks are equally difficult to support in a general-pur-
pose Internet, like the old model. If a customer had a single class B
address split over multiple sites, an IP-based internetwork would
not be able to figure where to route the packets.

There is an even more difficult problem with a level 3 network,
the question of policy routing. Let’s say that there are two Datanet
customers, one commercial and one a researcher, wishing to set up a
connection to the same destination in the U.S. The researcher would
want to use the FUNET route and, in theory, the commercial user
would need to go over the Infolan path.

When a packet from the first user comes in, a route to the desti-
nation would be cached in a router. Subsequent packets, regardless
of the source address, would all go by the same route. Under the
Frame Relay model, the user would set up a permanent virtual cir-
cuit to the edge of either FUNET or Infolan. Note that the policy
routing dilemma may show up further down the road, but as far as
the local (or national) data network is concerned, policy routing de-
cisions have been eliminated.

Policy routing protocols are one of the most difficult issues in
the Internet. Many are advocating the solution that Juha chose of
simplifying the network layer topology in favor of a general, public
data link service such as Frame Relay or SMDS. Others would ar-
gue that moving the public network down to the data link solves
problems today, but is a solution that will not scale well as the scope
of the network increases.

For Juha’s employers, Frame Relay certainly solved several im-
mediate needs. It avoided the policy routing dilemna for TCP/IP, it
supported a variety of non-TCP protocols over the same infrastruc-
ture, and it did away with access control lists.

Monday morning, I stopped by Juha’s laboratory at Telecom Fin-
land to see the network and read my mail. Finland is so well con-
nected that I got better response time reading my mail on a Telnet
session back to Colorado than I got when I went in locally over a

106

Tampere

9,600 bps dialup line. In fact, Vesa Parkkari had demonstrated
FTAM file transfers on top of TP4 and the CLNS providing end-to-
end performance from California to Tampere of 38,000 bytes per sec-
ond.

I then spent the rest of Monday making my way from Tampere
to Helsinki to London to Paris, with the obligatory many-hour lay-
over at London Heathrow. It was a relief to arrive at my familiar
hotel in Paris.

107

Paris

Tuesday morning, I took the Paris Métro over to the offices of Com-
munications Week International. The one editor I knew let me in,
showed me his boss's office, then promptly left for the U.S. The rest
of the morning, I furiously banged away at a PC, trying to get
enough columns written to stave off my editors in New York (and
pay for some dinners in Paris).

After finishing enough pithy columns to pay the Maitre d’, I
went to the nearby bistro for a lunch of chicken Riesling, ratatouille,
and a bit of wine. Sauntering back to the office, I had only to figure
out how to submit the columns to the head office.

Of course, the menu option on the PC labeled “MCI Mail” did
absolutely nothing. I knocked on a few random editor’s doors, but
was met by frantic stares and hurried brushoffs. My problem, of
course, was that I was wearing a suit and was thus mistaken for one
of the all-too-frequent employees from the U.S. that came in to use
the office for a day in order to write off their trip to Paris on their
income taxes. 3

Needless to say, the editor with a deadline to meet has no time
to work with these corporate junketeers who come in to do some
“work.” The only work these visits involved would be for the lo-
cals. Finally, I convinced an editor that I was only masquerading as
a respectable person and actually did write for a living. I couldn’t
quite bring myself to take on the identity of “trade press” or “jour-
nalist” but it was finally established that I was a “guest colum-
nist/consultant.”

The editor became quite helpful and took my disk and disap-
peared into her office to perform some magic submission ritual into
the Atex system in New York. I asked her why any PC couldn't
perform the same operation.

She explained that a team of consultants had been working as-
siduously on the problem, but some mysterious incompatibilty
problem had prevented fully operational capabilities from being at-
tained. Since I had a couple of hours to kill while waiting for a call

108

Paris

from London, I decided to take a look at this mysterious modem
malady.

I went back to the PC I had been working on and started up
Procomm. It seemed to be installed correctly, although, frankly, it is
just about impossible to install it incorrectly. I then pulled up the
dialing menu and picked MCI from the directory.

Nothing. Next step was to pull up a simple terminal emulation
screen and type “AT.”

“OK,” the modem responded.

Hmm. Obviously there was either a fully functioning modem or
a pretty good modem emulator on this system.

Walking around to the back of the desk, my highly trained eye
spotted the problem. The modem wasn’t plugged into the wall.
Sure enough, a few minutes later, I was reading my MCI mail.

Why would an Internet explorer keep an MCI Mail account?
There are two reasons. First, the telex gateway on MCI Mail is won-
derful. It allows me to communicate with people in countries like
China, which have very limited e-mail capabilities. The telex gate-
way actually allows me to send and receive telexes.

There is a more important reason, however. There are a few
people with whom I correspond, mostly members of the trade press
who haven't figured out how to reach my Internet mailbox from
their MCI Mail accounts. If I don’t keep an MCI Mail account, I’m
reduced to sending faxes back and forth to these people.

It has often puzzled me how people who spend their life cover-
ing computers can know so little about how they work. There are
some exceptions of course. The editor of Communications Week, for
example, has both Internet and MCI mailboxes, and knows the dif-
ference between them.

People like him are, unfortunately, very much the exception in
the trade press. The features editor of one prominent networking
newspaper, for example, was unable to figure out how to use MCI
Mail, yet persisted in editing “in-depth” articles about Broadband
ISDN, NFS, and a host of other technically sophisticated issues.

So how does the trade press write in-depth pieces? All too
often, the articles are simply a rehash of brochures and briefings. I

109

Exploring the Internet

remember a time I did a feature piece on DECnet Phase V for one of
the monthly publications.

I met the senior editor assigned to “work with me” in Boston.
Wearing a yellow power tie and driving a Detroit muscle car, he
told me that the piece lacked “punch.” So he rewrote it. Close to
deadline, somebody decided I ought to have a look at it. The new
and improved lead explained how DECnet Phase V would be avail-
able on systems ranging from VMS and Ultrix to the Macintosh and
PDP.

Macintosh and PDP?

Running DECnet Phase V on a PDP is kind of like trying to
simulate the weather on a PC/AT. Running on a Macintosh, though
highly unlikely, was theoretically possible, but in my stacks and
stacks of reading on the subject, I had not heard of any such project.

In a not so gentle manner, I explained to my senior editor that
having read a stack of functional specifications five feet high, it was
my considered technical opinion that running on a PDP was un-
likely if not impossible. Did my esteemed editor have a source for
his information.

“I was briefed by marketing,” he replied.

I called up the Senior Executive Technology Editor and told him
he was free to run the DECnet piece, but I would prefer, indeed
insist, that my name not be on it.

So he fired me and killed the piece.

Technical knowledge is considered to be of minimal importance,
indeed often a liability by the computer and communications trade
press. Again, there are exceptions. For example, Kelly Jackson, for-
merly with Communications Week, is known for insightful coverage
of data communications issues. There are others, but they are too
rare.

People with degrees in journalism are often hired straight out of
school, and are given no time to acquire any technical knowledge.
Frequent contact with senior executives gives junior journalists. a
feeling that they are “part of the scene.” All too often, though, they
are simply manipulated and fed a carefully composted salmagundi
of marketing manure.

110

Paris

The cycle feeds itself. The executives get used to getting good
press based on marketing hype instead of real information. The
journalists keep feeling like they are getting the real story and feel
increasingly confident and powerful.

There are two solutions to this quandary. First, technically as-
tute people need to write more. People like Marshall T. Rose, who
devotes part of his time to writing books, are all too rare. The tech-
nology does no good if people don’t know how it works and it is
the responsibility of the developers to document their work.

The second answer is for the trade press to acquire a bit more
knowledge. Being able to use electronic mail, for example, should
be a prerequisite to getting a piece published and certainly won't
cause a writer to transform into a technogeek.

Wednesday morning, I sat outside my hotel and waited to be picked
up by Pascal Renaud, head of the computer group at ORSTOM, a
national research institution in France.

ORSTOM’s mission is to work with researchers in developing
countries. It is an unusual institution in that half of its permanent
staff are outside France, scattered over 41 sites around the world.

Twenty minutes late, early for Paris, a tiny car zipped up on the
sidewalk outside of my hotel. Pascal jumped out and we intro-
duced ourselves. As we darted through the streets of Paris, |
handed him a copy of STACKS. He was pleased.

He was so pleased, in fact, he commenced to leaf through the
book as he drove. I had visions of my own book causing my death.
While this sort of untimely morbid vision had certainly occurred to
me before, I had always assumed I would meet my end at the hands
of an OSF hit squad, not in a ten horsepower Renault. Of course,
the OSF would use different tactics, probably boring me to death.

As we passed the Louvre, Pascal finished my book and we
started talking about ORSTOM. ORSTOM was founded right after
World War II as a think tank devoted to the French colonies. While
much of the research concerns questions such as agriculture and
health, it is nonetheless not a development agency. It focuses in-

111

Exploring the Internet

stead on basic research issues, such as how to stop the spread of
diseases like malaria and sleeping sickness.

Many of the ORSTOM research centers are located in former
French colonies. For example, most of the 17 African centers are
located in French West Africa, in countries like Niger and the Ivory
Coast. A typical research center, such as the ones in Niamey, Niger,
and Lomé, Benin, employs 50 people. Others, such as the center in
Dakar, Senegal, employs 100 French and 200 African researchers.
All told, the institute has a budget of FF 900 million (U.S. $150 mil-
lion), employs 880 researchers and 720 support staff, and has offices
in Africa, South America, Asia, and Oceania.

So how do you provide computing support to such a group? In
1985, Pascal Renaud was brought in to head the computer support
group. At the time, resources consisted of a motley assortment of
systems like an HP1000 and a Honeywell mini.

“I tossed out everything,” Pascal said with a smile. “It was a bit
of a fight, but I got rid of it all.”

The basic applications needed for ORSTOM were typical of
many scientific institutions, including statistical analysis, a bibliog-
raphic database, and analysis of satellite images. There were, how-
ever, two quite unusual requirements. First, with 41 locations all
over the world, communication was vital.

To make the communication issue even more important was the
issue of support. Many ORSTOM sites were located in places with
some of the world’s worst telecommunications. In addition, most
sites didn’t have trained personnel. Setting up and maintaining
equipment in such locations was a real challenge.

After a bit of digging around, ORSTOM settled on Sun worksta-
tions. Within France, the systems were hooked together using
TCP/IP over X.25 and were part of the French Internet.

Keeping links up 24 hours a day to Africa would not be practi-
cal. Some countries turned off electricity at night. In addition, tar-
iffs were high enough that a call to Paris 24 hours per day would be
prohibitive. TCP/IP also had some problems in such an environ-
ment. Running TCP/IP over a 2,400 bps link, the wizardry of Van
Jacobson notwithstanding, just doesn’t work.

112

Paris

The solution was UUCP. In some countries, such as Togo, X.25
networks were available and were often much newer than the voice
network. A SunLink board connected the workstation to Togopac at
1,200 bps. An X.75 link hooked Togopac to the French Transpac.
The UUCP F protocol ran on top of X.25 and calls were placed once
or twice a day to transfer mail.

The normal UUCP G protocol, designed for very bad lines, is
highly resilient but also inefficient. The F protocol has one check-
sum per file and is thus more appropriate for the relatively high-
cost, high-quality X.25 links.

Other places didn’t have X.25. In Ouagadougou, for example,
ORSTOM simply placed a daily call to the Montpellier hub in
France. The UUCP G protocol was used to transfer mail over the
dialup link.

This network was called RIO, which stood alternatively for
Réseau Informatique de l‘ORSTOM or Réseau Intertropical d’Ordi-
nateurs. Connectivity to the rest of the Internet was provided
through a TCP/IP link into France’s FNET and UUCP links to
EUnet.

The service on the network (at least the UUCP portions) was
simply e-mail. A pre- and post-processor for mail allowed accents
to be included in 7-bit ASCII]. UUENCODE allowed transfer of PC
and Macintosh binary files.

When a new site was added to the network, Pascal Renaud and
local researchers would host a conference. Conferences were typi-
cally well attended and typically included government officials, PTT
staff, any local non-governmental organizations (e.g., the UN), and
university researchers.

ORSTOM had a policy of allowing any of these people to use
RIO. This open policy helped make ORSTOM part of the local com-
munity and helped spread mail access to many new parts of the
world.

Invariably, Pascal relates, somebody at one of these conferences
would want to know if they could access RIO from the PC in their
office instead of walking over to the ORSTOM offices. As far as
Renaud was concerned, he had no problem with remote access.

113

Exploring the Internet

Unfortunately, placing a telephone call in many of these coun-
tries can be a torturous experience. After Pascal would explain that
remote access was possible, but would involve a modem and a call,
most researchers would shrug their shoulders and say “thanks any-
way, I'll walk in.”

Supporting scientific computing in such a far-flung network is
quite a challenge. Only a few sites had real computer support peo-
ple. Dakar, for example, had one staffer to support the 15 Suns at
that site. For other facilities, ORSTOM had an interesting support
structure.

In France, military service is compulsory. An alternative, how-
ever is a system somewhat like the U.S. Peace Corps. ORSTOM, as
a national laboratory, receives an annual quota of young engineers
and puts them to work maintaining computers in far-away places.

With several hundred users and an open-door policy, RIO is evi-
dence that some of the most important work in the Internet commu-
nity is done far on the periphery, removed even from the TCP/IP
protocols that many use to define the core.

At the core of the Internet, the issue is one of adding new classes
of users, such as recent K-12 initiatives. RIO serves an equally vital,
if not more important role of adding new countries to the global
village.

Leaving Pascal, I went to a café near my hotel and had a fine
cassoulet of snails for lunch. There, over a half bottle of Beaujolais
Villages, I pondered the immense difficulty of spreading the Internet
into some of the places Renaud had to work.

A few years ago, I travelled to a some of these sites: Niamey in
Niger, Lomé in Benin, and Abidjan in the Ivory Coast. In the cities,
such as Niamey, there is a passable infrastructure. Phones work
much of the time and with a UPS, power regulators, and security
guards you can keep a computer system going much of the day. In
regional cities in Niger, however, things were much tougher. Yet,
these regional cities have universities, international aid groups, hos-
pitals, and research centers.

Some would argue that high technology, such as computer net-
works, are inappropriate in countries where per capita income is

114

Paris

measured in the hundreds of dollars and the basic infrastructure,
like roads, is crumbling.

In such an environment, technology should be appropriately
used, not avoided. Agricultural aid workers, for example, need to
consult their colleagues overseas. Likewise, doctors should be able
to work as part of a global community instead of being left alone in
an isolated outpost.

Voice communication over a telephone system is not an effective
technology in these circumstances. International telephone calls
take a very long time to place and are often terminated abruptly.
Calls are expensive, and you never know if your party will be there.
Voice calls are a point-to-point communications technique and do
not allow appeals to large groups.

It is precisely in these outlying areas that the Internet is most
needed. Messages can be prepared at any time and then batched for
transmission. In many countries, the X.25 networks are far superior
(i.e., far newer) than their voice counterparts.

Technologies like electronic mail are important in the developed
countries by helping us do better research or provide better educa-
tion to children. In countries like Niger, electronic mail can support
education and research, but it can also do something much more
important. It can help save lives.

115

Geneva

Thursday morning, I got up early so I could reach the Gare de Lyon
in time to take the 7:25 Train 4 Grande Vitesse (TGV) to Geneva.
The TGV is the French high-speed train, taking only three hours to
get to Geneva, compared to six hours on a normal train and four on
a plane (by the time you get to and from airports).

I soon found myself on the Rue de Mont Blanc in Geneva with
plenty of time to spare before my long-awaited audience with Dr.
Pekka Tarjanne, Secretary-General of the ITU. The purpose of my
meeting was two-fold: to keep the standards server alive and lobby
for further changes. This was the only day of the month he would
be able to see me, hence the special visit.

Calling on people like the Secretary-General is a funny kind of
affair. You typically get only a few minutes, most of which is spent
in pleasantries, and I don’t ever expect much to be decided. Yet, it
is a vital step to take in an international bureaucracy like the ITU.

First, I went up to the 12th floor to see Tony Rutkowski. While
he finished sending out some mail messages, I wandered around his
office looking at the stacks of paper lying all over the place. Occa-
sionally, I would see multiple copies of some interesting report and
grab a copy.

Meanwhile, Tony had assembled his own stack of paper for me,
including copies of the slickly produced “Friends of Bruno” news-
letter. The newsletter was rife with references to the Digital Re-
source Institute and “Project Leader Malamud.” This institute thing
was certainly taking on a life of its own. Even worse, I had been
stuck with a title. The only title I usually get is “Mister” (and even
that is somewhat rarely employed).

Finally, the time came for the meeting with Dr. Tarjanne. Tony
ushered me up to the 14th floor. Now, I’ve been in some impressive
offices before, but this one certainly took the cake.

The secretary’s office and adjacent waiting rooms could easily
have rented as a SF 15,000 (U.S. $10,000) per month apartment. The
main office had a beautiful view of Lake Geneva and the Alps. It

116

Geneva

was lined with Persian rugs and was big enough to fit all 400 atten-
dees at an IETF meeting.

Dr. Tarjanne came striding over from the other side of the office.
While he walked over, I did mental calculations trying to figure out
how many miles per day he must put in simply greeting visitors at
each meeting.

Dr. ‘Tarjanne was trained as a Finnish physicist. Active in poli-
tics, he rose to become head of the liberal party in Finland. As a
member of the coalition government, he secured the appointment as
head of Telecom Finland, the Finnish PTT and he was able, through
an adroit political sense, to parlay that into his position at the ITU.

After reviewing the progress of Bruno and the 21 Sons of Bruno
servers, we talked about the steps that would be needed to realize a
truly global network that included developing countries and had a
solid, well-managed infrastructure. Perhaps the ITU could play a
role in helping that come about?

Tony and I then paid a similar visit to the Deputy Secretary-
General to brief him on our visit with the Secretary-General. Mr.
Jipguep, a former high official in Cameroon and considered the
dean of the African telecommunications community, appeared to
support the Bruno experiment.

Returning to Tony’s office, we discussed who I should meet and
brief on my return visit back to the ITU the next week. I suggested
to Tony that he call across the street and extend an offer to have me
brief them on the “experiment.”

He called Mike Smith, an Irishman who heads up the informa-
tion systems area. Tony relayed my offer, and hung up.

About five minutes later, Mike Smith called back. Tony listened
for a minute and hung up the phone laughing.

“Larry Eicher very much wants to talk to you,” he said, “I said
you'd be delighted.” Eicher was Secretary-General of ISO. 1
laughed and asked if I should bring my bulletproof tie.

I left Tony to take care of an important personal matter. No visit
to Switzerland is ever complete without a stop at the Mercure
chocolate shop. Picking up a nice assortment of champagne truffles,
bittersweet bars filled with 90-proof Cognac, and various other

117

Exploring the Internet

forms of adult candy, I resisted the temptation to buy the 4.5 kilo-
gram bar of Toblerone.

With the chocolate safely mailed back to the U.S., I took the TGV
back to Paris, arriving after midnight at my hotel in Montparnasse.
I ordered a taxi for 6, set my alarm for 5, packed my things, and
went to sleep.

118

Nice

Friday morning, I took the first plane down to Nice to meet Chris-
tian Huitema, the first European member of the Internet Activities
Board. One parochial American had referred to him as the first
“alien” on the IAB, but in my view that title had long since been
taken by one of the American members.

In Nice, I picked up my rental car and joined the morning grand
prix doing at least double all posted speed limits. In Antibes, I took
the turnoff for Sophia-Antipolis, the “French Silicon Valley.” So-
phia-Antipolis was founded 20 years ago as a joint venture between
nine local towns, a regional government, and the Nice chamber of
commerce. The regional group, the Departement des Alpes-Mari-
times, contributes over half the money.

By 1991, the park had grown to over 700 firms, the majority in
computers and communications. Air France keeps its central reser-
vation system here, Digital keeps a major research center, and the
EEC has the European Telecommunications Standards Institute
MENS.

Sophia-Antipolis is also the site of one of the major centers of
the Institut National de la Recherche en Informatique et Automat-
ique (INRIA), the premier French computer science research insti-
tute and the home base of Christian Huitema.

Travelling too fast, I zoomed out of a roundabout and didn’t see
the INRIA sign. A few kilometers later, I saw a sign for INRA and
turned in. The place was full of greenhouses. I had arrived at the
Institut National de la Recherche Agronomique. Woops.

After a few more diversions, I found INRIA. INRIA, like OR-
STOM, is a national research institution. It specializes in computer
science and numerical methods and employs 900 staff in 5 locations.
Each location is organized into a series of research projects.

At the Sophia-Antipolis site, there were a couple of dozen pro-
jects employing a little over 300 staff members. Three projects were
in robotics, including work for the European Space Agency’s Mars

119

Exploring the Internet

mission. Occasionally, the robot would take a tour around the
building to test its navigation skills.

Christian headed a project in computer networks. When I fi-
nally made it into his office, he was explaining his work to a young
German who was considering starting a Ph.D. at a nearby university
and working at INRIA to do his research.

I spent the day with Christian, and by the end of the day I be-
gan to see why he had been named to the JAB. At the age of 22,
Christian had no definite career plans, so he became a consultant.
He did quite well, but found he was always doing the same thing
and got bored.

He then joined the Centre National d’Etudes de Telecommunica-
tions (CNET). By 1985, he had received the Doctorate d’Etat,
roughly equivalent to an assistant professor at a U.S. university. His
thesis, part of the NADIR project, dealt with transport protocols for
satellite links.

He had by then moved on to GIPSI, a French effort to build a
workstation. CNET made the hardware, there was a port of UNIX,
and the system even sported a graphical interface.

As part of the team working on GIPSI, Christian concentrated on
networking. He invented an NFS-like remote disk protocol that
worked directly on Ethernet. He also designed a version of X.25
that ran over Ethernet.

This modified X.25 was implemented by removing HDLC, since
802.3 already provided a satisfactory data link. Any transmission
necessary was moved up into the link layer of X.25. The software
even did call setup using an ARP-like mechanism. The calling sta-
tion broadcast a call setup and the target station would respond.

The result was an X.25 running at over 500 kbps, matching the
performance of TCP on the test platform, a Motorola 68010. Not
bad, to say the least. Christian, backed by the French PTT, brought
this idea to the international standards table. This was his first, but
certainly not his last, encounter with what he sardonically calls the
“classical” standards process.

The opposition camp, led by the English and the Germans, felt
that if you did X.25, you should do so by the book. This meant that
sending datagrams into the ether was out. Instead, they wanted the

120

Nice

use of Logical Link Control class II, a connection-oriented version of
Ethernet which has been widely standardized but rarely used.

The whole thing degenerated into a religious stalemate. As with
many talented people, this encounter with the standards world
moved Christian more firmly into the TCP/IP camp.

In 1986, Christian was brought on as an INRIA staff member. In
an unusual move, INRIA immediately named him a project leader.
One of the first projects in Christian’s group was an X.400 over X.25
implementation. Christian’s work on X.400 led to other projects
such as an X.400/SMTP gateway. That gateway is still up and run-
ning at INRIA and is used by several groups. SWITCH, the Swiss
Research Network, used a leased line to INRIA to link their X.400
services to the Internet SMTP base for several months. Over time,
that mail link migrated into a 64 kbps IP line from SWITCH to IN-
RIA.

Related to the X.400 project was an X.500 project called The Ob-
viously Required Name Server (THORN), run under the auspices of
the European Community Esprit project. THORN led to a DSA im-
plementation called Pizarro. Naming it after the Spanish conqueror
was a takeoff on the UCL’s Quipu, an Incan name.

Working on all these OSI-based systems gave Christian’s group
considerable experience in coding the OSI middle layers. One out-
growth of that experience was an ASN.1 compiler.

ASN.1 is a very powerful, very general specification for the
presentation layer in OSI. The problem is that most people write
programs in C, not ASN.1.

When an application receives data from an OSI network, as in
the case of an incoming X.400 message, that message is encoded in
ASN.1. An ASN.1 compiler starts with the ASN.1 specification for a
message and generates the encoding and decoding routines that al-
low an application to interact with the network.

The INRIA compiler, known as MAVROS, was a fully general
system, accepting any ASN.1 specification. It also handled variants
of ASN.1, such as the X.509 standard used for digital signatures. It
supported a lightweight version of the presentation layer, quite use-
ful when similar machines send lots of integers and floating point
numbers back and forth.

121

Exploring the Internet

An outgrowth of the compiler effort was an ASN.1 benchmark.
To test the compiler, and to compare ASN.1 to other presentation
paradigms such as the Sun XDR specification, the benchmark pro-
vides a basis for measuring performance.

Christian’s group started by examining a typical X.400 message.
Based on that message, a tree was developed with a depth of 8 and
a total of 400 ASN.1 elements. As per the typical X.400 message, 80
percent of the elements were octet strings, 10 percent strings, and 10
percent integers.

Next, they started comparing the code generated by the ASN.1
compiler with basic lightweight methods. They looked at the
amount of time to encode and decode data as well as the amount of
data transferred on the network.

Needless to say, the initial performance was slow. The results
from the benchmark were used to start optimizing the compiler. For
example, decoding is slow in ASN.1 because of the tag, length, and
value encoding for each element. In a technique similar to Van Ja-
cobson’s TCP work, the ASN.1 routines began using header predic-
tion techniques. If the next tag can be predicted from the current
context, header prediction can greatly speed up decoding.

Running the ISO layers over TCP/IP and FDDI on two DEC
5000s, the INRIA group was able to achieve presentation layer
throughput of 8 Mbps. By contrast, TCP throughput on the same
configuration was 17 Mbps, UDP 25 Mbps. The bottleneck in all
three cases appeared to be the CPU.

In addition to his research efforts, Christian has been active in
helping to link the French infrastructure to the Internet. As with
many international connections, this one can be traced back to the
ubiquitous Larry Landweber. Christian was invited to attend one of
Landweber’s networking conferences, this one hosted by Dennis
Jennings in Dublin in December 1986.

Larry and Christian started talking and decided that INRIA
should be linked to the Internet. Larry went to NSF, Christian to
INRIA and each got half the link funded. Before the link was up,
UUCP mail to the U.S. could take anywhere from hours to days.
With an initial link speed of 56 kbps, mail started taking just min-
utes, or less. The INRIA link to the. U.S. gave Christian more in-

122

Nice

volvement with the Internet community and led to his appointment
to the IAB.

I had a wonderful meal of foie gras and veal, hoping there were
no animal rights activists lurking outside. To drink, we had a nice
Bordeaux followed by an even better Beaujolais. One thing I’ve no-
ticed is that one of the qualifications for membership on the IAB
seems to be a exquisite taste in wines (many of which come with
exquisite prices). IAB members such as Stephen Kent, Vinton Cerf,
Lyman Chapin, Dan Lynch, and Christian Huitema are all renowned
in the Internet community for their technical abilities in this area.

Saturday morning, I retrieved my car and drove along the coast
of the Mediterranean back to the Nice airport. Arriving back at
Hertz 26 hours after I had picked up my keys, I was handed an
invoice for U.S. $300.

“Thank you very much,” I said, picking up my bags and head-
ing towards the terminal. You know you've been travelling too long
when a U.S. $300 bill for one day in a Renault “Junior” seems ac-
ceptable, even if only momentarily.

By the time I got to the terminal, my currency conversion proc-
ess got swapped into memory and I promptly turned around and
walked back to Hertz. Using a few choice words remembered from
my high school days in Switzerland, along with some universally
understood hand gestures, I convinced the station chief that I felt
that a fee of $300 was not totally appropriate. Through some hard
negotiating, I got the rate down to U.S. $125. What a bargain.

123

Geneva

Sunday night, I took the TGV back to Geneva. Feeling refreshed
after a day in the south of France and a weekend in Paris, I was.
ready to sink into the bureaucratic abyss of Geneva.

Sink I did. I spent the next three days battling the ITU bureauc-
racy, trying to stop a rear guard action that was threatening to kill
the Bruno project.

In four weeks, the Bruno server had been a remarkable success.
Twenty-one servers on four continents had cloned the file system
and were distributing the Blue Book. Bruno was getting as many as
35 packets per second. Over 500 hosts in 27 countries had retrieved
over 65,000 files. We had no statistics from the other servers, but it
was not unreasonable to think that several hundred thousand files
of the Blue Book had made their way out to people who were actu-
ally reading them.

How did this compare with paper copies? This was hard to say,
as profits from documents had served as a sort of discretionary fund
for the previous Secretary-General of the ITU. Knowledge of publi-
cations was highly dispersed; only finance seemed to have sales
data, and they kept this information closely guarded.

Nonetheless, it appeared that the Bruno experiment had in-
creased the distribution of the Blue Book by at least one order of
magnitude, and probably two or more. Tony had documented all
this in his “Friends of Bruno” newsletter and had papered the
[TU—paper being the only medium that appeared to work there.

Yet, despite all this, the high-level Information Systems Steering
Group had met the previous Friday to decide the future of Bruno.
Rumor had it that the outcome of the policy group was that the
experiment had not been successful and was over.

Stopping the experiment was, of course, not an option. The
server was in Colorado in a locked room and I had no intention of
stopping operation. Besides, twenty-one other servers had the data.
Tony and I had carefully structured this project so there would be
no turning back.

124

Geneva

There were, however, some important factors at stake. I had
hoped to begin putting other ITU documents on the servers. A pol-
icy decision that the experiment was over would mean that we
might have to bypass the ITU and start scanning paper copies.

More importantly was the role of the ITU in dissemination of
standards over the network. The logical outcome of the Bruno ex-
periment was to have the ITU put itself on the Internet and take
over this function.

Tony had set me up for three days packed with meetings. Many
of these people were division directors or other Very Important Bu-
reaucrats (VIBs) who sat on the Information Systems Steering
Group.

One meeting, in particular, stood out over all the rest. I was
scheduled to meet with Walter Richter, director of something impor-
tant. He was 45 minutes late, so I spent the time looking at the
stacks of file folders on his secretary’s wall-to-wall bookcase.

They had wonderful labels such as “The Preparatory Committee
on Restructuring of Subsidiary Machinery” or “The Administrative
Committee on Coordination.” One was simply labeled “High Level
Committee” and took up several folders high up on the top shelf.
The committee that seemed to take the most wall space was the
“Consultative Committee on Substantive Questions.”

Finally, Richter strode in. Speaking with a heavy Austrian ac-
cent, he preceded to tell me how my experiment “was not a success
and has been terminated.”

He seemed very certain that the experiment had not been a suc-
cess, so I asked why. It appeared that this Internet of mine (the ITU
considered the Internet to be some private project run by Tony
Rutkowski and myself) just didn’t reach the right sort of people. By
the right people, he seemed to mean those who were on the Admin-
istrative Council of the ITU or those that worked on the consultative
committees like the CCITT or the CCIR.

The conclusion that the Internet had the wrong sort of people
was Odd, since I had not analyzed the data on who was accessing
my server. In fact, anecdotal evidence was pointing to just the op-
posite conclusion. I had received personal messages from places
like AT&T, Bell Labs, and Telecom Finland.

125

Exploring the Internet

I found out later how Richter came to this view. Richter had a
buddy on the radio side of the ITU, the CCIR. His buddy had a pal
in Canada to whom he had spoken.

“Ever hear of this Bruno thing?”

“Nope.”

“Ever hear of this Internet business?”

“Yeah, but we checked it out a few years ago and it was too
expensive.”

Well, there you go. Can’t argue with a few personal anecdotes
when making a high-level policy decision. I tried trotting out a few
of my own anecdotes, but Richter had already assembled the data
he needed.

Richter had something in common with most of the other VIBs
that I met in three days at the ITU. He was very, very sure of him-
self. For example, he was absolutely convinced that the entire ITU
network architecture was fatally flawed.

I must confess, this was certainly my working assumption when
starting to deal with the ITU computer group, but the reality turned
out to be that they had a fairly decent network architecture in place.
Not what I would have chosen, but adequate for the job.

I asked Richter to tell me what was wrong.

“The Ethernet,” he replied. When his PC had first been in-
stalled, it was a diskless machine. A mistake, of course, but it had
been fixed. He was convinced that all ITU network problems had at
their root Ethernet saturation, because it had once taken several
hours after pressing a key to see the character appear on his screen.

Based on this anecdote, he was ready to completely microman-
age some fairly talented engineers that worked at the ITU computer
department. Rather than set broad policy (an area that had been
sorely lacking), he was convinced that the answer was to roll up his
sleeves and dig into the bits and bytes.

Another curious aspect of the three days of meetings was this
idea of the Internet as some academic toy that real people didn’t
use. I met with one staff member who expressed this view and
waved a piece of paper at me that had the names of delegates he
was working with as proof.

126

Geneva

I picked up the piece of paper and started going through the list.
Many of the places, such as the Centre National des Etudes de Tele-
communications in Paris, were clearly on the Internet. In fact, the
vast majority of institutions on the list appeared to be connected in
one form or another.

There was another more fundamental issue that started to focus
and helped explain this reluctance among the VIBs. Printing docu-
ments was a big empire at the ITU, and building empires was the
name of game. My project was not a good way to build big empires
(efficiency never is).

The printing department at the ITU was truly an impressive
place. I walked past the “keep out” signs and gave myself a private
tour. There were seven offset presses, four state-of-the-art, top-of-
the-line Xerox 5090 copiers, and a dozen or so other large copiers.
The ITU’s own facility generated only a fraction of the total output.
Swiss printers had a long and cozy relationship with the ITU bu-
reaucracy.

Things would be printed with no relationship to demand under
the assumption that larger print runs meant a lower per-unit cost.
True, of course, but if you throw away most of the units, your aver-
age costs can be considerable.

One of my underground sources gave me an example. For sev-
eral years, the ITU had produced a beautiful four-color “charts in
profile” document. Each time, 10,000 copies would be printed at a
cost of several hundred thousand dollars.

Of this print run, 2,000 copies would be given away and roughly
100 sold. Yet, every few years, a new edition would be put together
and 10,000 new copies would be printed. Strolling around base-
ments and subbasements, I saw enough paper to start a firestorm.
Pallet after pallet was loaded with boxes and boxes of documents
that nobody would read.

This being the decade of the environment (or was it the chil-
dren?), I naively asked about the ITU recycling program. Needless
to say, one didn’t exist.

The Bruno experiment directly threatened this paper empire.
The bureaucracy had framed its argument very cleverly. Every year,
the ITU had received several million Swiss Francs in revenue from

127

Exploring the Internet

selling documents. The official “profit” from the Blue Book had
gone to fund programs in the developing world.

In other words, Bruno was depriving the ITU of revenues that
would fund vital infrastructure. My selfish little project meant that
people who needed to call a doctor wouldn’t be able to. Project
Bruno, baby killer.

Nobody actually accused me explicitly of killing babies, but I
certainly felt that undercurrent. After donating several months and
several thousand dollars to putting ITU standards online, I had
somehow not expected this type of reaction.

The donation was the single most difficult concept for the VIBs
to understand. Why was I doing this? What was my motive?
What was in it for me?

Of course, donations to the common infrastructure are how the
Internet was built. Even formal standards bodies like the ITU run
on donations. Corporations work in the standards process as a vol-
unteer effort.

In the Internet community, volunteer efforts are the norm. The
IETF has many people who attend as private citizens, paying for the
privilege three times per year out of their own pockets. Paying for
the privilege of getting the Blue Book online was not remarkable,
but VIBs didn’t know what to make of it.

The “Bruno, baby killer” aspect was a difficult one. Profits from
document sales were virtual at best, and the simplest solution
would be to redo the accounting system to look at the total costs of
the inefficient document production process, but that proved to be
dangerous.

I thus attacked the widespread unease with giving away copies
on the Internet. Tony was advancing the novel theory that by giv-
ing copies away, you increased the market and thus increased sales.
Such an argument, although bearing a few logical flaws, seemed to
stop the VIBs, at least for a few minutes.

While the ITU was criticizing the Bruno experiment, they were
attempting to move forward on their own electronic document han-
dling system. Evidently, the Bruno situation had impressed the Sec-
retary-General enough that he had presented himself at the meeting

128

Geneva

of the Information Systems Steering Group and suggested that they
should do something.

The committee had thus spawned a task force. The task force
had formed a small working group. Their initial inclination was to
start using X.400 as a way to send out working documents, but only
to members of the committees.

Tony and I tried valiantly to switch the focus to getting the ITU
on the Internet and sending the documents out to as wide an audi-
ence as possible. This was not meeting with much success. The
computer department was worried about the resource implications
of such a move and wanted two additional staff members (in addi-
tion to their current network staff) to support the effort.

Basically, the bureaucracy desperately wanted to get back to a
world they could control. In order to control documents, however,
you need to own them. Nobody at the ITU wanted to admit that
there was a possibility that the ITU didn’t own its own documents.

Tony Rutkowski had made an analysis of the issue of copyright
and had come to the conclusion that the ITU didn’t have a sustain-
able basis for asserting copyright protection. Many of the other
VIBs, however, felt that the issue was cut and dried.

There are no apparent legal cases in which somebody has chal-
lenged copyright on a standards document. There are many factors
that must be weighed before a court will uphold a copyright claim,
and it was naive to think that the issues are so simple that the ITU
could confidently claim they would win in a court of law.

In order for a document to have a copyright applied to it, it
must, among other criteria, be original and not previously publish-
ed. Since almost all standards start out as public domain working
documents, even this fundamental requirement is not often met.

Many jurisdictions do not allow protection to be granted on offi-
cial or governmental works. Even a private standards body might
be considered by the courts to be quasi-governmental. Many places,
such as the U.S., make standards a procurement requirement, mak-
ing copyright enforcement questionable at best.

Even if standards are copyrightable, only the representation of
the standard, not the contents, can be protected. Tony’s conclusion
was that in almost any jurisdiction, running the paper through a

129

Exploring the Internet

scanner and OCR software and posting ASCII text would be defen-
sible.

Many standards do have graphics, of course. The graphics have
a stronger basis for copyright, since the representation is everything.
As we had seen with the Blue Book, though, in many cases the
graphics were not absolutely essential, at least for getting a rudi-
mentary understanding of the standard.

With these factors in mind, Tony and I walked across the street
to meet Larry Eicher, Secretary-General of ISO. My feeling was that
even if there was no copyright on standards, it was certainly easier
to work with ISO than against them.

The fact that Tony accompanied me was meant to send the mes-
sage that my efforts enjoyed at least some support from the ITU. I
brought along a copy of STACKS; Tony brought the slides from his
presentation to INTEROP in which he concluded that it was un-
likely that any standards organization could assert copyright on
documents.

“Do you think that’s diplomatic?” I asked.

“Nothing wrong with pushing forward the state of the art,” he
said with a smile.

We met Eicher and Mike Smith, one of the leaders of the task
force which supports the OSI effort. Both turned out to be very rea-
sonable people.

I gave a little speech about the moral necessity of disseminating
standards. I advanced the view that the reason that OSI had taken
so long to come to market was simply because it cost so much to
find out about it.

We then started talking about applying Bruno to the ISO world.
Eicher was quite frank: 25 percent of ISO revenues came from the
sales of standards documents. How did I propose to replace that
revenue? Even more importantly, ISO was controlled by its member
organizations, which also made much money from standards sales.
How did I propose to convince groups like ANSI that posting stand-
ards for free would help them?

Simply put, it was a question of financial survival. Interestingly
enough, Eicher was clearly unwilling to argue his case on copyright
grounds. When I ventured the theory that copyright protection for

130

Geneva

ISO documents was legally weak and that some radical might just
go ahead and post the standards, Eicher said “it’s not a question of
copyright protection, it’s a question of fair business practices.”

We began searching for a potential solution. I proposed my
high resolution/low resolution compromise. The plan would post
low resolution versions of documents for free on the network and
allow ISO and ANSI to continue to sell the high resolution versions,
either on paper or electronically.

Low resolution might mean ASCII text and 200 DPI bitmaps of
graphics, formulas, and other elements not well suited to repre-
sentation as ASCII text. Some document format such as ODA could
be used to tie the pieces together. Using ODA would help ISO by
spurring the development of the standard by giving people a sub-
stantial base of documents worth reading.

The crucial assumption was that people with the free version
would then pay for documents. I argued that free distribution of
standards would increase the base of people who read documents
by at least a factor of 10, maybe even more. Many of these would
want the paper documents. Giving away standards would lead to
increased revenues.

I then offered to test this theory on Bruno at no cost to ISO.
Eicher agreed to at least consider a formal proposal, so we went
back to the ITU and dashed off a formal letter. Kind of a long shot,
I figured, but certainly a first step. (I never received a response to
my letter, but that was no suprise. I did, however, publish the offer
in Communications Week just in case ISO had misplaced my letter
and needed a reminder.)

Tony and I had one more item of business to attend to. A semi-
nar had been scheduled for Monday morning to give people at the
ITU a briefing on Bruno. Tony had sent electronic mail on Friday,
but by Monday, none of the mail had arrived and therefore nobody
showed up at my lecture.

Turned out that the entire ITU mail system was running off a
VAXmate with a very limited amount of memory. If you sent mail
to everybody at the ITU (only a few hundred people), the system
crashed.

131

Exploring the Internet

Rather than remove the offending mailing list or even move the
message handling system up to an appropriate host, the issue of the
“mail server situation” had entered the bureaucracy and a heated
debate had begun, focusing on whether or not to expand memory
on the VAXmate. I was a bit incredulous. Running a mail system
for 900 professional bureaucrats off a VAXmate is kind of like using
a Volkswagen Beetle to haul timber out of the Amazon jungle.

We rescheduled the seminar, booking Tarjanne’s personal confer-
ence room for the occasion. The secretaries were worried that the
room would be too small (it could only hold 50 people or so), but
Tony and I insisted that the venue had the appropriate symbolism
and that having to turn away people wouldn't be all bad.

Since electronic mail wasn’t going to do the trick, we had only
one alternative: ElevatorNET. Posting notices in the elevators was
about the only effective means of communication at the ITU, the
organization that invented X.400.

The ITU has some of the strangest elevator manners in the
world. When you enter a lift, custom requires you to greet every-
one. Everybody then choruses back a hearty “bon jour.”

When you leave, you say goodbye and everyone responds with
their own “au revoir.” Nice custom, but what it means is that dur-
ing a busy period, it can cost you a dozen hellos and goodbyes to go
up just a few floors.

Tony and I pushed the button for the elevator and caught the
first one, respecting the elevator protocol. Every time we caught an
elevator, we rode a few floors, long enough to post the notice. We
then got off, pushed the button again, and hoped that a different
elevator would start. Finally, dozens of bon jours later, we had
caught the last elevator and posted the last notice.

Wednesday morning, Tony had prepared all sorts of handouts
for the eager crowds. I nervously sat in the corner and prepared my
talk.

Two people came. One was our ally in the computer depart-
ment, another a gentleman I had already briefed. Nobody else both-
ered to show up.

We all chatted for a few minutes, had a cup of coffee, then went
back to Tony’s office. I bid Tony goodbye.

132

Geneva

Thursday, I took a series of four flights from Paris to London to
New York to Ithaca, home of Cornell University. Sitting on the
planes, I had plenty of time to reflect in wonder at the ITU. Many
people were grateful that Tony Rutkowski had put the time into the
bureaucracy, but you had to wonder how somebody that talented
could survive in such a labyrinth. (He didn’t for very long. Tony is
now an employee of Sprint International and Vice President of the
Internet Society.)

133

Ithaca

Thursday night, I arrived at New York’s Kennedy Airport and took
a commuter flight to Ithaca. USAir decided that 15 cities with no
luggage lost was tempting the fates and obliged me by sending my
bags to Syracuse. My request for extra frequent flier miles, based on
the route my bags took, was greeted with a blank stare.

While negotiating with the lost luggage functionary, H. David
Lambert, my host in Ithaca, arrived to pick me up. He smiled
knowingly when he saw that I was looking for my bags. Evidently,
USAir uses some stochastic routing algorithm for luggage in upper
New York state.

Dave brought me to his house, with a stop en route for massive
quantities of Chinese take-out. There, David, his wife Carol, and I
caught up on old times.

Dave Lambert operates on the front lines of computing: he man-
ages academic computing centers. I got to know him when I was a
graduate student at Indiana University and Dave was Assistant Di-
rector of the Academic Computing Services.

In those days, Academic Computing meant a CDC 6000, punch
cards, and two terminals for the entire campus. My first word proc-
essor was on that machine. We would use punch cards and a home-
grown text formatter written in FORTRAN to input our formatting
directives, with the results sent to a line printer.

Dave brought me into the student consulting pools, rescuing me
from yet another semester teaching a required class in economics to
drooling undergraduates. My job was to sit in a cubicle and answer
any and all questions. In the summer, I continued to do my consult-
ing, but we had only a few students and most of them spent their
time at the abandoned quarries swimming.

We did, however, have 6 VAX 11/780s that had just been deliv-
ered. My consulting cubicle was a designated receptacle for one of
the sets of documentation, which filled up most of a wall. Having
nothing else to do, I started at one end of the wall and read manu-

134

Ithaca

als. It was then that I learned the key to a successful consulting
practice was to Read The FMS Manual (RTFM).

Under Dave's leadership, the Indiana University network grew
from a few hundred to several tens of thousands of active users. He
introduced a wide variety of programs that dramatically changed
how computers were used on campus. An example was the so-
called MBA Experiment.

The word “experiment” is a way of labelling some project to get
over strenuous objections from people opposed to anything new
(hence the use of the word “experiment” at the ITU). The experi-
ment gave permanent accounts to all MBA students, instead of dol-
ing out resources as part of some specific class, as had been the case
previously.

A single, over-worked VAX was loaded up with 1,200 students,
along with tools to do e-mail, financial modelling, and even access
external data sources such as the Dow Jones News Service. Non-
credit seminars were held to familiarize students with the tools, and
a consulting cubicle and terminal room were put into the business
school.

The project met strong resistance from two places. First, the sys-
tems group in the computing center were aghast at this wild, uncon-
trolled use of resources. What if everybody started using the system
at once? One system manager actually went so far as to say putting
users onto the system would make it run inefficiently.

The second source of opposition was the business school faculty,
who felt all this new-fangled nonsense would take away from a seri-
ous academic atmosphere. Behind the scenes, of course, was the
unspoken fact that if the students learned the tools, the faculty
would have to also.

Despite these objections, the experiment started up. The results
were dramatic. Students started using sophisticated modelling tools
to solve case studies, forcing the professors to learn the same tools
in order to grade the assignments. Electronic mail became such a
part of the culture that one professor even gave his exams by mail.

Eventually, the entire nature of the MBA Program changed. The
obligatory Fortran class was dropped in favor of a computer skills
seminar emphasizing tools like e-mail and Lotus. Many other

135

Exploring the Internet

classes in the curriculum were radically changed. Students left the
program computer literate, with a noticeable change in placement
SUCCESS.

The MBA Experiment was one of dozens of projects that Dave
Lambert started, usually over the strenuous objections of one or
more factions. He helped wire dormitories, develop educational
software in fields like English, obtain cheap PCs for students, and
hooked the campus to the Internet.

What Dave started, Carol Lambert had to help finish. Carol spe-
cializes in setting up support structures for computing services. At
Indiana University, this originally meant writing the reports on how
to use the CDC mainframe. This matured over time to a highly
trained, highly service-oriented operation, providing everything
from a help desk (with people who had answers) to a problem
tracking and escalation system, extensive user documentation and
classes, and standardized procedures for obtaining resources.

The computing service that Dave and Carol helped build was so
successful that it won the first INTEROP Achievement Award for
excellence in education. By then, Dave and Carol had moved on to
Ithaca, where they were part of a team trying to expand and de-
velop the Cornell campus network.

Cornell is the largest Ivy League school, with a population of
17,000 undergraduates. It is also one of the key Internet hubs. The
Cornell Theory Center was one of the original NSF supercomputer
centers. Cornell is also the site of the Network Operations Center,
supporting Sprint's role as international connection manager for the
NSFNET. Cornell monitors international links to NORDUnet and
INRIA in France and is part of both the T1 and T3 backbones.

Cornell is also one of the key sites in the Internet for work on
routing protocols. In fact, the campus network is run over routers
developed at Cornell and running on a PC/AT. These multiprotocol
routers support, along with many other networks, over 284 separate
AppleTalk networks. The Cornell routers are managed by a set of
bootservers. When new software is developed, Trivial FIP is used
to send the new boot images down to over 100 routers.

Some of the more interesting routing work at Cornell was done
by the team of Scott Brim and Jeff Honig. Scott, quiet by nature, is a

136

Ithaca

bit of an anomaly in the raucous world of routing. Meetings about
routing protocols at places like the IETF tend to be full of histrionics
and hysterics, routing being the original religious issue in the In-
ternet, but Scott talks in a whisper.

Scott and Jeff are the developers of the Gateway Daemon (gated),
a widely available piece of routing software for UNIX and other op-
erating systems. The gated software has its origins in the period
when regional networks started coming into being and hooking
themselves up to the NSFNET.

TCP/IP has two kinds of routing protocols. Interior protocols
are used to communicate reachability information within a routing
domain. On UNIX systems, the Routing Information Protocol (RIP)
was the quintessential interior protocol and was implemented in a
software module called the Route Daemon (routed).

Reachability between routing domains, as in the case of a re-
gional announcing its networks to the backbone NSFNET, uses a
protocol such as the Exterior Gateway Protocol (EGP), or its more
modern successor, the Border Gateway Protocol (BGP).

In the early days, both routed and EGP received reachability in-
formation. Each would take that information and update a master
routing table in the UNIX kernel. As long as each routing protocol
handles a separate set of target networks, having two modules up-
date a single routing table sort of works.

What started happening, however, was that a target network
might be reachable through two different routes. Information about
one might arrive via RIP, the other via EGP. The update by one
protocol would be walked all over by the other.

The gated software took the two separate modules and put them
into a single daemon, allowing coordinated updates to the routing
table in the kernel. Over time, gated was revised to make it very
modular, allowing new routing protocols to be quickly added.

The gated software has become one of the standard platforms for
prototyping new routing protocols. Current gated modules include
the OSI IS-IS and its TCP/IP cousin OSPF, the Border Gateway Pro-
tocol (BGP), the old RIP-like Hello protocol, and the policy routing
twins, IDPR and IDRP.

137

Exploring the Internet

An interesting aspect of gated is the way it is distributed. Cor-
nell maintains the reference implementation, which people can
download from anyplace in the Internet. This makes it perfect for
small organizations that have a UNIX box and want to connect to
the Internet.

Vendors that incorporate gated in their products sign a redistri-
bution agreement. This agreement requires the vendor to feed any
enhancements to gated back into the reference implementation, mak-
ing any improvements available to the entire community.

Next, I met Steve Worona, who escorted me through the gorges of
Ithaca to find some lunch while he told me about CUINFO. In the
world of Cornell, replete with supercomputers, optical disk juke
boxes, and other paraphernalia of high technology, CUINFO is cer-
tainly one of more impressive services on campus.

CUINFO is a low-tech information service. It started nine years
ago when Steve was an IBM systems programmer working on spe-
cial projects for the academic computing group. He noticed that
there were exam schedules and class rosters available online to ad-
ministrators.

Why not make this information available to students? He wrote
a program and put a terminal in the lobby of the administration
building and CUINFO was born.

Pretty soon, somebody suggested they put up the daily head-
lines from the campus paper. Sounded good, so that was added.
Pretty soon, the information on the system started growing expo-
nentially.

Perhaps the best known service is “Dear Uncle Ezra,” started by
Worona and Jerry Feist, the head of counseling services at Cornell.
Uncle Ezra is a Dear Abby-like service named after Ezra Cornell, the
founder of the school. It provides a way for students to ask ques-
tions, anonymously if they wish, of counselors. :

Answers to questions are usually posted in a bulletin board for-
mat, allowing everyone to benefit. Individual responses are often
sent directly to the student for more serious inquiries.

138

Ithaca

Questions range from the very serious, such as suicide counsel-
ing, to the inane. My favorite question, of course, is on the inane
side of the spectrum: ‘Dear Uncle Ezra, I always wonder why Carl
Sagan is so famous. I mean he doesn’t have any personality and he
speaks funny.”

Contrary to the well-written, precise, to-the-point answers usu-
ally provided by counseling services, this one got a fairly evasive,
noncommittal response.

Uncle Ezra has spawned a host of other answer-line services.
The Career Services group started Auntie Em, Mr. Chips handles
issues related to instruction, and NutriQuest provides information
on nutrition and food.

All of these services run on an IBM using the VM operating sys-
tem. TCP/IP support on the IBM extends the services to the cam-
pus network and out to the rest of the Internet. Users can access the
service by using Telnet to CUINFO.cornell.edu, specifying that port
300 should be used instead of the default Telnet port.

CUINFO has grown to the point where it receives over 100,000
accesses per month. It contains information on local restaurants and
a wide variety of the other types of information you might expect
from a high-tech videotex service. It proves that if you give people
very simple tools to publish information and a bit of support, a
videotex system can be quite successful at very low cost.

That night, Dave and Carol Lambert and I started swapping user
services horror stories. We started with the simple ones, like the
IBM staffer who couldn’t figure out how to get the command
prompt on his Macintosh. Carol brought out the old war horse of
the PC user who transformed a 5.25-inch disk in to a 3.5-inch disk
using scissors so it would fit into the slot.

Finally, Dave trotted out the ultimate user story, attributed to
Ken King, former Vice Provost at Cornell and then President of
EDUCOM.

Ken King was sitting in his office at Cornell one day when he
got a call from the machine room. There was a crazy man down

139

Exploring the Internet

there waving a knife around and threatening operators. Some For-
tran bug or something.

King promptly called security then went sprinting down the hall
towards the machine room. Halfway down the hall, his secretary
shouted after him that he had an important call. Thinking it might
be security, he raced back to take the call.

It wasn’t security. It was an irate user who proceeded to yam-
mer on about some problem with the service he was getting. Ken
tried to cut in and finally succeeded.

“I'm sorry, but I have to run. There’s a crazy man in the ma-
chine room waving a knife around,” he explained.

“Oh,” the user replied, “but this will just take a minute.”

140

Monday morning, my publisher Paul Becker got up at an ungodly
hour to pick me up at 7 AM. and bring me to breakfast. One of the
secrets to supporting yourself as a writer is to get your publisher to
buy you as many meals as possible, although with Prentice Hall this
is certainly no guarantee of fine dining.

Paul drove me up towards Harlem, near my first appointment at
Columbia University. We found a cheap diner and I had my first
American breakfast in weeks, greasy eggs and limp bacon.

I grabbed my bags and carried them across the street to the
Seeley W. Mudd building, home of Columbia’s Center for Telecom-
munications Research (CTR). CTR is one of six NSF-sponsored En-
gineering Research Centers and specializes in optical networks.

The main CTR project is called Acorn or Terabit, depending on
whom you ask. The Terabit name aptly summarizes the goal of the
project, to develop a scalable optical network providing gigabits to
the user and terabits of aggregate bandwidth by exploiting the full
capacity of optical fiber.

I rode up to the 12th floor of the Mudd building, thinking what
a silly ring the name had, and reminding myself never to allow a
building to be named after me. Not that the subject has come up
frequently.

Since I was early, I went into the office of W. Clayton Andrews
to wait for him, resisting the temptation to walk over to the work-
station and Telnet to Boulder to read my mail. A few minutes later,
Clayton Andrews came in.

A distinguished-looking professor, Andrews was formerly direc-
tor of an important IBM laboratory in Switzerland. He had joined
CTR as an associate director, working alongside such optical lumi-
naries as founder Mischa Schwartz and the current director, An-
thony Acampora.

Acorn had been successfully deployed in the laboratory when I
visited, and Andrews and others were making plans for further de-

141

Exploring the Internet

ployment into the real world. They were pursuing two different
sites to test the hardware and software.

One would use the dark fiber that Teleport had deployed in
lower Manhattan. The other use was on the Columbia campus net-
work, which was getting ready to lay multimode fiber around cam-
pus. CTR was trying to get some single-mode fiber pulled at the
same time, because of the minuscule marginal cost.

Acorn is based on the linear lightwave network (LLN) concept
developed by Thomas Stern, another Columbia professor. The LLN
uses different wavelengths of the fiber for different signals. By com-
bining wavelength multiplexing with a star-like topology, circuits
can be set up between different nodes on the LLN.

The architecture is scalable, allowing stars of stars to be built,
potentially supporting a network of a million or more nodes. Cir-
cuit establishment is provided by electro-optical filters located
throughout the LLN which establish a path from a source to the
destination.

Networks that actually do something are certainly more useful
than those that run test patterns, so Acorn is pursuing a half-dozen
test applications to help shake out the network. At the edge of the
LLN network is a Network Interface Unit (NIU). The NIU links up
to the user workstation using local interfaces such as HIPPI.

Andrews brought me into the laboratory to see the prototype
network. The entire network was on a Single laboratory table, with
a few racks of electronics located nearby to test and drive the net-
work. The system had been tested with the transmission of High
Definition TV over the network, using ATM cells to move data at
800 Mbps. The network has also been tested at the full speed of 1
Gbps.

The Acorn project would deliver this gigabit capacity to the door
of several organizations on campus, who would then work with
CTR to try and exploit the bandwidth to do something useful.

The applications included classical gigabit projects such as medi-
cal imaging and quantum chemistry. It would also bring bandwidth
to places like the law school, which was trying to move large
amounts of data around for full text search and retrieval. Of course,
the bottleneck in the law school application would probably be the

142

New York

complicated keyword searching of large optical jukeboxes, not the
network.

Another application, possibly the most useful, was a proposal by
the Teacher’s College at Columbia University to use computers
more in K-12 programs. Teachers College hoped to link two schools
in New York, one in the inner city and the other an exclusive private
school, to multimedia resources.

In 1990-91, a prototype of the system was put in at the private
school, allowing sixth-graders to communicate with a CD-ROM con-
taining a data base on ancient Greek culture called the Perseus Pro-
ject. The software allows groups of students to perform a group
excavation of a hypothetical site in Greece.

The Acorn project would set up a library of resources at Colum-
bia and make them available to students at the schools. The project
would buy as many off-the-shelf resources as possible and store
them in the library. When a user wanted to access a resource, the
data would be retrieved, compressed, and sent over the network to
the user’s workstation.

Throughout the twelfth floor, we saw laboratories filled with
esoteric-looking hardware. This was definitely a hands-on research
center. In one room was a molecular beam epitaxy system, in an-
other a testing laboratory for HDTV, and in a third an experiment to
provide different guaranteed classes of service on complex, high-
speed networks.

The real, operational nature of CTR was a sharp contrast to many
computer laboratories, where hands-on usually meant typing on a
keyboard. It was an even sharper contrast to my afternoon appoint-
ment at the Columbia Institute for Tele-Information (CITI), located
in the Business School.

CITI, as you might guess from a place with Tele-Information in
the name, is not the kind of place with lots of computers. Instead,
the offices are stuffed with paper, books, and loads and loads of
reprints of CITI publications.

143

Exploring the Internet

CITI is an economist think tank about networks. The founder is
Eli Noam, a lawyer and an economist. A former member of the
New York State Public Service Commission, his is one of those ubiq-
uitous figures in the discussions that hover around the FCC and set
much of our policies for public networks.

CITI had been trying to drag me into one of their policy studies,
a grandiose project called Private Networks and Public Objectives.
My contact was an ebullient young Irish Marxist, Aine M. Ni
Shuilleabhain, an able policy analyst putting in time at CITI to pay
the rent.

Public Networks and Private Objectives is characteristic of the
studies conducted at think tanks like CITI. Lots of people attend a
few conferences and seminars and some papers are prepared. The
papers are then sold as part of the huge Working Papers series.
Then, after a little massaging, some of the papers are collected to-
gether and published as a book.

This particular study was going to look at a taxonomy of net-
works, develop a general research framework, turn that into a con-
ceptual model of networks, and finally come up with a policy
agenda for the future.

The problem CITI was having was that they were getting lots of
speakers, but very few of them knew much about networks.
Granted, many were specialists in networks, but most were econo-
mists specializing in issues like public utility pricing models. Occa-
sional terms like “star topology” or “coaxial-fiber hybrid system”
would enter the conversation, but the general discussion was a bit
removed from reality.

Aine was certainly aware that in discussing the policy fate of
technical networks, it helped to know a bit about the technical na-
ture of those networks. She had been trying hard to recruit some
more technical people to participate and help lend a note of reality
to the process.

I ended up waffling out of the study. Having been trained as an
economist and then escaped, I was not anxious to get back into the
world of simplistic theoretical models based on unrealistic assump-
tions. Besides, I had to finish my trip and then get ready for two
more circuits around the world.

144

Washington, D.C.

Tuesday morning, I got up early for a D.C. power breakfast of the
Washington Post and black coffee. The black coffee on an empty
stomach helps counteract the nausea you get reading the latest go-
ings-on on Capital Hill, a subject the Post covers in excruciating de-
tail.

I then strolled down 16th Street, past the White House and the
Old Executive Office Building to the nondescript, generic federal-is-
sue offices of the National Science Foundation. There, I went up to
the 4th floor to find Steve Wolff, the Director of the Division of Net-
working and Communications Research and Infrastructure.

I was shown into a dark office. Steve strode over out of the
shadows and introduced himself.

“T hate fluorescent lights. If you want, I can turn them on.”

I demurred and groped around for my notepad. There was just
enough light coming in from the hall to take notes.

Steve Wolff took over the position of program manager from
Dennis Jennings, the Irishman. Wolff helped formalize the NSFNET,
supervising the shift from an interim 56 kbps backbone to a profes-
sionally managed T1, and later T3, network.

The solicitation for the T1 network had been won by a group
that included the Michigan state network (Merit), IBM, and MCI.
Later, a non-profit company was founded by MCI and IBM. This
company, Advanced Network and Services (ANS), built a T3 nation-
wide network. ANS was inserted into the organizational chain with
NSF and Merit.

When Steve Wolff and I talked in November, these organiza-
tional details had been ironed out and a nationwide T3 network,
run in parallel with the older T1 network, was slowly becoming op-
erational. Wolff was turning his attention to the next phase.

After numerous public hearings, Wolff was about to recommend
to the National Science Board what the next phase of the network
should look like. Needless to say, Wolff was not going to give me
his formal recommendations before he told his bosses. Since I

145

Exploring the Internet

wasn’t trying to scoop Network World, we discussed more general
issues.

Wolff explained the theory of network infrastructure as a pyra-
mid. You could grow it by adding to the top or by broadening the
base. Adding to the top was being done in areas like the national
gigabit testbeds, intended to make the core run faster. Gigabit net-
works were still being tested, but it was evident that in some form
they would soon allow researchers to start using higher bandwidth.

Adding to the base of the pyramid was equally important.
Every year, NSF helped another 100 schools get on the Internet.
This connections program was really quite simple. The school re-
ceived U.S. $20,000, enough money to buy a router, pay for some
telephone calls, and enslave part of a graduate student. After the
initial U.S. $20,000, the school was on its own. To Wolff’s knowl-
edge, not a single school had dropped off the network after the in-
itial money ran out.

Seeding activities and leveraging investment are the fundamen-
tal strategies the NSF uses. They view their job as one of getting
things started, not providing ongoing support. Most of the regional
networks, for example, started off with exclusive or extensive NSF
funding. By 1991, many of the regionals were self-sufficient. Even
the core, NSFNET, received only a portion of its funds from the fed-
eral government.

My next stop that day was the Office of Technology Assessment
Where I raided the document room and even found a copy of a
study I had contributed to several years ago, but had never seen the
final results. I then headed out to Dulles Airport for my last stop of
the trip, the Internet Engineering Task Force meeting in Sante Fe,
New Mexico.

146

Santa Fe

Late Tuesday night, I arrived in Santa Fe. I checked in and
walked down to the terminal room that had been set up for the
IETF. After reading several dozen mail messages, and with over 100
still left to wade through, I didn’t have the heart to continue and
called it quits.

Standing in the hall was Paul Mockapetris, the inventor of the
Domain Name System. Paul and I left the hotel to go find some
dinner, choosing a restaurant stochastically. The waitress came over,
and gave us the introduction obligatory in such chi-chi places.

“Hi, I’m Foo and I'll be your waitress this evening.”

She handed us each a flimsy little card. On the card were the
words “fish,” “chicken,” “beef,” and “lamb.” A little icon was next
to each word. There were a couple of other icons signifying potato
and onion ring. No prices. No prose.

“What is this,” Paul asked, “a RISC menu?”

I decided to see what the wine list would look like in a place
like this. Foo came back a few minutes later carrying four bottles of
Chardonnay and plopped them on the table. We pointed to our de-
sired icons, Paul pointed at a bottle of wine, and our point-and-click
supper was on its way.

The next morning, suitably fortified with a breakfast of a burrito
stuffed with Chipolte cactus and eggs, I followed a stream of what
were obviously computer engineers to the morning plenary.

The IETF differs from most other industry meetings in that it is
not a show or a conference. It is a place where people come to
work. The plenary was very short and consisted of a technical pres-
entation on ATM. Then, people went off to their working groups.
Working groups are the main business of the IETF. It is in these
meetings that people get together to develop the standards that
make the Internet work.

147

Exploring the Internet

The standards that have come out of the IETF have been impres-
sive. They are produced quickly, and they work. In the area of net-
work management, for example, the IETF produced an object
identification hierarchy, management information bases for a wide
variety of modules, and the Simple Network Management Protocol.

The IETF has been remarkably successful, but as it grows in size,
its ability to accomplish work appears to diminish. Veterans of the
IETF, like Marshall T. Rose, are vocal about the negative impact of
success on the ability to accomplish real work. Marshall likes to
distinguish between two types of attendees: the “goers” and the
“doers.” The goers are those who like to go to conferences, forget-
ting that the IETF is a working group, not a conference.

The basic difference is one of self-definition. Somebody like
Marshall T. Rose defines himself in terms of accomplishments: “the
father of ISODE” or “a leader in the development of SNMP.” Oth-
ers, however, define themselves by position: “head of a working
group’ or “member of the ACM.”

This conflict has a fundamental impact on groups like the IETF.
The goers try to get themselves named as working group chairs,
and all of a sudden committees for the sake of committees start to
flourish. The number of working groups topped 60 by the time of
the Atlanta meeting in Summer, 1991. Stev Knowles, the most vocal
minority of the IETF, suggested a working group be formed with
the purpose of reducing working groups. Somebody else suggested
that a working group be formed to study the question.

Meanwhile, it keeps getting harder to get real work done. Peo-
ple attend meetings who have not done the preliminary reading.
Working group sessions get stalled with naive questions. Even post-
ing required reading lists to the net doesn’t seem to get people to do
their homework (or to attend and keep quiet if they hadn’t done the
reading).

Marshall T. Rose grew tired enough of the “goers” that, over a
drink, he suggested a radical system of Certified Protocol Engineers.
By Marshall's scheme, in order to attend a working group meeting
and participate, you must be certified by a board of your peers in
that area. Network management, routing, and mail systems, are all
examples of possible areas.

148

Santa Fe

The board in an area would be bootstrapped by two people of
unassailable quality, who would draft the oral and written exams.
Once certified in the area, you would be part of the governing
group. The system would be quite similar to that used in the medi-
cal profession.

Working group participation would thus be limited to certified
protocol engineers. Anybody could attend a meeting, but to partici-
pate you must either be certified or be invited by the chair. This
fairly radical proposal would probably never get passed—and was
certainly proposed with tongue planted firmly in cheek—but it
makes one think about how such a technocratic priesthood might
function.

Aside from the usual work of readying router requirements and
massaging management information bases (MIBs), the IETF has a
ritual bloodletting plenary session on the penultimate day. The ple-
nary always starts very slowly, with IETF chair Phill Gross uttering
the obligatory platitudes of introduction.

While he is speaking however, you can look around the room
and see people positioning themselves at the microphones, waiting
for the meeting to open to the floor. Usually, if the controversy will
be especially loud, you can hear rumors in the halls in the days
before the plenary.

The Sante Fe meeting, however, had seemed fairly quiet and no-
body was expecting a major controversy. Lack of real issues, how-
ever, will not always stop people from speaking.

As soon as Phill Gross stopped speaking, one engineer stepped
to the mike and started raving about the Internet Activities Board.
His proposal was to abolish the board as superfluous.

I was sitting in the back of the room, looking over the shoulder
of Jon Postel, editor of the RFCs and a long-time member of the
IAB. Jon has a habit of scrunching down in his seat as he gets more
and more disgusted. As the engineer continued to rave, Jon was
almost sitting on the floor.

149

Exploring the Internet

Finally, people started to chime in with the monologue, inter-
rupting the mad ravings of the lunatic engineer and reminding him
that the IAB may have problems but it certainly had played a valu-
able role in the development of the Internet. Jon Postel slowly
started sitting up in his seat.

The IAB doesn’t always get accolades at the IETF meetings. The
role of the IETF is technical advisor to the IAB, which goes ahead
and makes policy decisions. In the previous IETF meeting, held in
Atlanta in July, tempers had flared as high as the humidity.

The story of that firestorm helps illustrate some of the inherent
conflicts in the process. The IETF had developed an MIB for SNMP-
based management of Ethernet modules. The working group that
developed the MIB included many of the standard IEEE-developed
management variables, but also included a few optional variables
that the group felt were necessary.

After several meetings and much e-mail correspondence, the
working group forwarded the MIB to the general IETF. At a ple-
nary meeting, the MIB was put on the table and no objections were
voiced. It then went to the IETF Steering Group (IESG), composed
of leaders from each of the main areas the IETF works in.

The IESG examined the MIB, saw no objections, and forwarded
the document up to the IAB with a recommendation for approval.
This is the normal process that any recommended standard origi-
nated by the IETF takes.

Once the document hit the IAB however, it sat. People were
busy and had lots of things to do. Finally, it came time to act on the
document. Tony Lauck, the chief network architect for Digital
Equipment Corporation and a member of the IAB, looked at the
MIB and thought that some existing Ethernet vendors would not be
able to easily implement the standard. He felt that it was different
from the IEEE standard and that this was not necessarily a good
thing.

This is all well and good. The role of the IAB is technical arbiter
of the TCP/IP protocol suite and Tony Lauck had identified what he
saw as technical issues. He marked up the document, crossing out
some variables and adding detailed instructions for revision of the
MIB. The IAB then sent it back down to the working group.

150

Santa Fe

The working group exploded. They felt that there had been
many opportunities to provide technical input into the process and
that the MIB reflected a technical consensus consistent with the net-
work management framework that had become an Internet stand-
ard. They saw the IAB as making a crassly commercial decision,
caving in to the wishes of a few vendors instead of providing lead-
ership in the standards arena.

In Atlanta, everybody came to the microphone and started giv-
ing their views. Why hadn’t the IAB attended the working group
meeting if they had concerns? Why had the IAB made technical
edits instead of providing policy guidance?

Karl Auerbach of Sun Microsystems got up to speak. A lawyer
by training and an accomplished engineer, he is a vocal participant
in the IETF and an active SNMP developer. The rewritten MIB had
been discussed at an unscheduled working group meeting the pre-
vious evening, and Karl was concerned that this violated the open,
consensual nature of the IETF.

Others got up and complained that the IAB had issued several
different explanations of what had happened, each different. Others
were concerned that the IAB had tried not to tread on the IEEE turf,
neglecting technical leadership for politics.

Finally, members of the IAB admitted they had fumbled the
question. The issue was not technical, it was procedural. How to
run an informal standards process like the IETF, yet preserve due
process safeguards, has been a continuing problem as the body
grew in size from the original 13 participants to the several hundred
attending at Atlanta.

The issue of due process was certainly a crucial one. It appeared
to some members of the IETF that Digital Equipment had sabotaged
a standards action on commercial grounds. Karl Auerbach raised
the question of the potential legal liability of the group.

The questions of accountability and liability had not escaped the
attention of people like Vint Cerf, then chair of the IAB. He was
actively working to promote a new professional society, the Internet
Society, which would sponsor the activities of the IAB and appoint
the members.

151

Exploring the Internet

The IETF was fun, but by the time an angry mob got around to
putting Jordan Becker, vice president of ANS, on the grill about the
transition from the T1 to the T3 NSFNET backbone, I was ready to
leave. Seven weeks on the road had certainly been enough for me.

152

Boulder

Getting off the plane in Denver, I found myself in the middle of a
minor blizzard. I collected my bags and fought my way through
the snow to find the shuttle bus back to Boulder.

On the bus, I struck up a conversation with the clean-cut young
man sitting next to me.

“What do you do?” I asked with some apprehension.

“I install custom industrial equipment,” he replied.

Well, I thought to myself, that’s not too bad. In Boulder, I hesi-
tate to ask that question because more often than not the person is a
professional aromatherapist, an astrologer, or a trance-channeler for
mythical beings.

“What kind of equipment?” I asked.

“Large-scale bean sprout growing machines,” he replied.

The Boulder version of heavy industry. At least he didn’t work
for Wholebrain Technology, the prototypical Boulder boutique, a
business establishment specializing in “light and sound technology
for mind expansion and brain development.”

The next four weeks were spent reading mail, paying bills, and
sorting out hotel reservations in 43 cities. About half-way through
my stay at home, I came home from the travel agent to find a fax
from Pekka Tarjanne of the ITU.

The letter congratulated me for the wonderful work I had done
and terminated the Bruno experiment as of December 31, 1991, a
mere 90 days after it went operational.

The letter disingenously said that although the experiment was
terminated, “measures are in progress for a similar service to be
made available under ITU auspices.”

I had talked in Geneva to Robert Shaw, the technical staff mem-
ber working on an Electronic Document Handling System. His vi-
sion, still very much in the conceptual stage, was to put a PC with
X.400 software on the ITU network and offer only working docu-
ments to a tightly controlled group of people.

153

Exploring the Internet

The letter from Tarjanne also insisted that I somehow convey to
“all those who are operating info-servers with copies of the ITU
standards that IT’s authorization for distribution of this material
ceases after December 1991.”

Kind of like publishing a newspaper and then asking people to
return the copies. Even worse, many of the 21 sites around the
world had invested money in upgrading their equipment to handle
the ITU standards. A 90 day experiment was a farce. Pekka Tar-
janne, despite his good intentions, had finally been beaten down by
the bureaucracy.

I tried to find out what had happened from Tony Rutkowski.
He had been called in (but only after the letter was dispatched to
me), sat down at a table, and read the letter. He had protested
vainly that once the cat is out of the bag, there is not much you can
do, but his arguments fell on the unwilling ears of a bureaucracy
threatened by any sign of innovation.

Meanwhile, I sent out notes to the Internet community, inform-
ing them of the ITU’s decision. Even if the ITU had not kept its
commitment, we didn’t want to expose any of the other sites to an
unwitting legal liability.

Of course, chances were highly unlikely that the ITU would sue
anybody. A strong legal argument could be made that by allowing
uncontrolled dissemination of its standards, the ITU had given up
any claim to a copyright it might have once had.

One problem with undoing distribution on the Internet was that
I had no idea where all the copies were. Just to make sure that I
reached all the servers, I dashed off a piece for Communications Week,
letting Tarjanne’s office know that I was performing this service for
him.

154

Round 2

From Christmas to Chicago

a5) mone pao ate
standuxds that IT's coronene he 4 fisirily
Cases after Decembeg 1992.” =
Kind of ike poblishiog! « newspaper aud
Qwerty ropies.: Emer wormd: tamny/ot' tee Oh it
we rid had inwersted pinney in apgrtding viele ex ‘
the ITU standards, A: Sd day’ experiment “as. a: farce

wert iw nie ant Siow, by aes a qt oy
Oger ..ah sé meds MOM ee

Pe 8 :

’ Vis:
rote as
. | - emis

i tried { e. qul Wht ha id { happens di fro Sony, ats tke
Me had been walled in That ¢ mty attnr the lettes WS) CAN patch F
‘tt | i, and read the leties: fie had, . tes
iniy that once ye cat ont oF tne} ag, there in. not Gy Rit
at his argamnents fell an the unwilling pers of a F reg
' y any Sen of winds atic fi. i a Poe
a wre, 1 2 mt ou moins to the intervie community, i R

ine treat cf ee (Tle cledisiorn@ Byven if thu: TTY nad nov x
+ get! be yp OF i wliessatte ite

witiaslon
2pies Watt,
é off a piece fork:

Berkeley

I didn’t realize that traveling on Christmas day would be such a
novelty, but it was in the Wonderbread world of Boulder. Being
Jewish, I long ago learned that travel was optimal on Christian holi-
days. While the silent majority practiced their rituals, I was able to
travel in peace. The flight attendants feel sorry for you and feed
you champagne, which you can enjoy with plenty of elbow room.

In Boulder, though, the news that I would be traveling on
Christmas was inevitably met with great remonstrations of pity.

“It’s OK, I’m Jewish,” I would try to explain.

“Oh, but still,” they would sigh.

Boulder is a cowboy version of California. It sports an uneasy
mix of starry-eyed new-agers and redneck wannabees, with an odd
combination of the intensely mellow and the highly midwestern.
This is the kind of place where, when you tell a joke, people thank
you for sharing it with them. It’s the kind of place where every-
body eats organic sprouted bran breads, but they go to mini-malls
to get it. Its the kind of place that thinks it’s the center of the world,
but in reality is simply too close to Kansas and not nearly far
enough away from Marin County.

You know you've come from an island when Berkeley seems real
and down-to-earth. I picked up my car from the airport and drove
up the Bay, past the hole where the Embarcadero Freeway had been
torn down, and over the Bay Bridge to Berkeley.

Thursday morning, with a day to kill, 1 walked down to Tele-
graph Avenue. Although increasingly littered with junkies and wi-
nos, the street still has some charm left from its heyday in the 60s.

First stop was the Café Mediterraneum, a cavernous coffee shop
where you can sit for hours nursing a café latte and watching the
passing parade of residents sporting “proud to be weird” buttons
and teenagers dressed like Hippies. On the wall of the Café Med is

157

Exploring the Internet

a “no soliciting or dealing” sign. It was on the second-floor terrace
of this café that I wrote two of my books, typing every day until my
café latte and laptop batteries ran out.

The only problem with typing in Bay Area cafés is the questions
you must periodically field. At the Café Med, every day that I
typed I would receive several inquiries about the feasibility of calcu-
lating astrology charts on my device.

The astrologically-inclined would-be cyberpunks were an inter-
esting diversion from trying to decipher vendor manuals. Some of
the queries were even more basic, such as the one I received while
working on another book in the Café Picarro, located in the heart of
San Francisco’s Mission district. I was busily typing away, trying to
wax poetic on master-slave replication when I realized that some-
body had been standing staring at me for a good ten minutes. |
looked up and saw one of the local bums, a benign psychotic who
wrote random ravings on a piece of paper, xeroxed a dozen copies,
and sold them as a newsletter to buy a U.S. $1 bottle of wine. Feel-
ing that we had something in common, I always tried to be nice to
him.

He had a very puzzled look on his face, staring at my notebook
computer. Seeing me look up, he seized his chance.

“Uh, excuse me, but, like, how do you print on this thing?”

“You don’t,” I replied in a matter-of-fact tone of voice, and went
back to typing.

He stood there for another ten minutes watching me and
scratching his head, trying to figure out what the point was to this
write-only memory device.

I wasn’t sitting in the Café Med, however, to answer any existen-
tial questions. I had been gone from the Bay Area for several
months and I was stoking up on caffeine to get myself ready for a
tour of the bookstores.

My standard tour has three stops and takes all day. It starts
with Cody’s in Berkeley, one of the best general purpose bookstores
in the country and one of the places bombed for selling Salman
Rushdie’s Satanic Verses.

After stocking up on obscure travel literature (and making sure
my own books were in stock), I headed out of the Berkeley sun,

158

Berkeley

over the Bay Bridge, and into the mist of San Francisco. The city
hadn’t yet fully recovered from the earthquake yet and instead of
being dumped in the middle of North Beach, I took a round-about
route through the warehouses and the financial district.

In previous years, I had taken this route every day, but on a
scooter instead of in a car. Then, I had blasted through traffic, doing
the messenger slalom across the lanes. This time, I had a rental car
and had to resist the temptation to treat the gas pedal as a digital
device.

Finally, I found myself on Columbus avenue and darted into a
parking space across from the City Lights bookstore. On the top
floor is the publishing house that, in 1953, became become famous
for being the only house willing to publish beat authors like Jack
Kerouac. The first floor and the basement are stuffed with a won-
derful collection of literature, communist propaganda, and the like.

After spending my allowance, I walked up into North Beach.
First stop was Molinari’s deli. In a few weeks, I would be in Bang-
kok visiting the restaurant critic of The Bangkok Post and it was un-
thinkable to arrive without an appropriate sausage. I debated the
merits of diverse salamis with the clerks, purchased a few of the
better specimens, popped a few prosciutto-and-mozzarella stuffed
peppers into my mouth, then headed back to the car for the trip to
the South Bay. On the way, though, I received a shock. The Con-
dor, San Francisco’s first topless bar and a sleaze magnet for the
entire state, had become the Condor Bistro, a fern-and-cappucino
establishment.

On the way to the South Bay, I made a diversion to the Mission
for lunch. La Tacqueria Menudo is named after their specialty, tripe
tacos. No tripe tacos for me, though. I prefer lengua, beef tongue.

La Menudo is no ersatz chimichanga joint. Tacos are the real
thing here, served on soft tortillas with lots of salsa and cilantro,
washed down with a can of Tecate beer. Rubbermaid containers on
the tables hold fresh radishes, pickled hot peppers, and more salsas.
Mexican MTV booms out over the speakers, unless a soccer match is
on.

Suitably refurbished, I took the I-280 freeway down to the South
Bay to Computer Literacy, the largest and definitely the best com-

159

Exploring the Internet

puter bookstore in the world. With 20,000 different titles, this store
is a mecca for the computer literate. For the first two years | wrote,
I calculated that I spent as much money buying books here as I
earned from writing them.

With a new collection of reading matter, I continued playing the
digital tourist and headed up to Fry’s. Fry’s is to Silicon Valley
What the Louvre is to Paris. This is a supermarket for hackers.

The doors to this store are labeled Enter and Escape. Over 20
full-duplex aisles, each 100 feet long, are stacked with every con-
ceivable piece of software, peripheral, chip, and cable. People wield
shopping carts up and down the aisles and then report to 35 check-
out stands that can easily process several thousand people per day.
On your way out with your new peripherals, you can pick up
twinkies and Jolt to fuel the creative process.

After picking up a new fax modem, I headed next door to
browse in the Weird Stuff Warehouse, a store that specializes in
used and surplus equipment. This is the kind of place where you
can pick up a VAX 11/750 for $50 or a Symbolics 3600 for $125,
always useful if you are remodelling your recreation room and want
to put in a bar.

After a stop in Los Altos Hills for the best wine I’ve ever tasted,
courtesy of Dan Lynch of Interop, I headed back to the hotel to sort
out my modems, books, and salamis.

160

Mountain View

Saturday afternoon, I headed down to Mountain View, the home of
Sun Microsystems. I was going to visit Steve Roberts, better known
as “the guy with the bike.”

In 1983, Steve decided that a life of consulting and writing books
on the subject of microprocessors was not for him. He tallied up the
things he liked to do, and decided on a short list of writing, meeting
people, and riding bikes. He sold his home in Columbus, Ohio, and
invested everything he had in a recumbent bicycle, an early laptop,
solar panels to power his ham radio and other devices, and then hit
the road.

For 10,000 miles, he kept on pedaling, going through Florida, the
Rockies, the deserts of Utah and California, and the bleak desolation
of San Clemente. On the way, he used CompuServe as his elec-
tronic mailbox and wrote articles for any publication that would
send him money. The bike was certainly a natural conversation
opener, helping to fulfill his most important ambition of meeting
lots of people.

The original “Winnebiko” was only a couple of hundred pounds
and had 18 gears. Over time, that original frame has grown to sup-
port 105 gears, 580 pounds, and an incredibly sophisticated collec-
tion of on-board computers.

Sun Labs lent space to Steve to let him work on the latest incar-
nation of the Bike, called the Big Electronic Human-Energized Ma-
chine... Only Too Heavy (Behemoth). I reported to Sun’s Building
4, and was met by Steve, a tall bearded man in his late thirties,
wearing a “Peace” T-shirt. He led me through dark empty corridors
to a locked door bearing a neatly labeled sign which read “Nomadic
Research Labs.”

“Labs” was an understatement in this case. Under one of the
desks was a futon and scattered throughout were other signs that
the lab doubled as a home. Walking in the first time, though, I
didn’t immediately notice the futon, the flute, or the cereal bowl.
My eye naturally turned to the huge bike in the middle of the floor.

161

Exploring the Internet

At first glance, I saw a recumbent bicycle with a storage unit
mounted behind the seat and an aerodynamic hood up front. Be-
hind the bike is a trailer, covered with 72 watts of solar panels used
to charge 45 amp-hours worth of batteries.

Sticking up from the rear of the trailer is a big yellow pole con-
taining ham radio antennas. The pole is six feet tall and can be ex-
tended to 12 feet for even better reception. Another antenna on the
back is a little Qualcomm satellite dish, similar to what you would
use on a delivery truck. Roberts keeps a Sun 4/260 at Qualcomm
headquarters and runs 165 bps UUCP-based transfers into the bike.

The satellite dish provides coverage anywhere, but 165 bps isn’t
really optimal for true connectivity. A high-speed modem and a cel-
lular phone allows high-speed IP access to a server located at Sun.

The bike has a variety of different computers. A Sparcstation, a
Macintosh portable, and a DOS portable had all been ripped apart
and mounted onto the bike. The screens were mounted on the bike
console, with one screen flipping up to expose another.

Another laptop is in a removable carrying case, allowing Steve
to compute in diners and other places where the bike is not wel-
come. The carrying case is, of course, in constant communication

with the bike.

Input to computers while riding is provided by a handlebar key-
board and a head mouse. The handlebar device is based on an In-
fogrip BAT chord keyboard which has 7 keys. Force-sensing
resistors are built into the handlebar grips and are linked to the BAT
keyboard controller.

The chord keyboard allows Steve to type at roughly 35 words
per minute, not an optimal rate when you make your living being
paid by the word. Once the keyboard controller generates charac-
ters, the data is sent to a macro package on a DOS machine that
expands the data. Through careful use of macros, the data hitting
the target computer appears to be a typist working at 100 words per
minute.

The head mouse is the other major form of input. Three
transducers are mounted on Steve’s helmet and they are used to
sense an ultrasonic beam generated from the console. The resulting

162

Mountain View

motion detected is fed into the controller of a Macintosh ADB
mouse.

The helmet has a few other features to help the digital nomad.
A 720 x 280 pixel red image floats in front of him on a little display
mounted on the helmet and acts as a console. A heat exchange sys-
tem built into an inner liner for the helmet acts as an air conditioner,
using the refrigeration unit on the bike to cool fluids which travel
up and remove heat from the helmet.

The helmet also has a little microphone on it which connects to a
serial cross-connect bus. The bus can set up connections to cellular
phones, the ham radio, or even the voice recognition unit on the
bike. The cross-connect bus also links modems to serial ports and
any other serial links that need to be established. The audio bus can
handle up to 8 simultaneous events and the serial bus can handle
up to 4 simultaneous events.

Just as the serial bus goes throughout the bike, so does a power
bus. Power is distributed in a series of batteries. The main power
source is three 15 amp-hour batteries, plus there are various other
special purpose batteries scattered about.

On sunny days, photovoltaic cells recharge the system. Addi-
tional power is provided by a regenerative braking controller, which
transforms the heat generated by braking into power. If there are
no hills and no sun, a cable connected to a car’s cigarette lighter will
do the trick.

A Motorola 68HC11, programmed in Forth, acts as the power
controller, sending power where needed and monitoring usage. The
controller signals the main bicycle control unit when power is low,
which in turn tells Steve so he can take corrective action. Two other
68HC11 controllers handle the serial and audio buses.

The audio bus allow sound output from many sources to be
mixed. Two 4 inch Blaupunkt speakers are mounted behind the
driver and connected to an automotive stereo amp. Data from the
CD player, the cellular phone, the ham radio, or the speech synthe-
sizer can all be mixed. Steve carries over 100 CDs with him on the
road, removed from their original packaging to save space.

163

Exploring the Internet

This entire contraption is worth over $300,000 in parts alone. If
you added in time, it is easily worth over $1 million. Security natu-
rally becomes an issue with an asset of this magnitude.

Security for the Behemoth is provided on several levels. At the
most basic level, a microwave doppler motion detector reports any
motion within 8 to 10 feet. Other detectors signal when a person
touches the bike or sits on the seat. A Global Positioning System
(GPS) receiver on the bike hood reports bike movement and speed.

Different responses can be set to security events depending on
the circumstances. The voice synthesizer can use the speakers to
utter an appropriate phrase if somebody touches the bike, such as
“do not touch, or you will be vaporized by a laser beam!” A siren
can be activated, or Steve can be paged.

If the bike starts moving without a password, more drastic ac-
tions are possible. The wheel can lock, a call can be placed to 911
and the speech synthesizer can request help, the siren can be acti-
vated, and, most effective of all, a wild-eyed digital nomad is pro-
grammed to come automatically bursting out of his tent in a mad
frenzy.

The Behemoth and its previous incarnations attracted the media
from the word go. CBS, USA Today, and even the trade press
quickly realized that this was a story. Being “the guy with the bike”
has made Steve instantly recognizable throughout the U.S.

For the first 17,000 miles of journeys, Steve simply wandered.
An office manager at home base took care of things in return for a
cut of the revenue. Over time, though, Nomadic Research Labs has
grown into quite a business. Steve publishes a somewhat quarterly
journal called Nomadness and is in frequent demand on the lecture
circuit, providing an interesting alternative for groups that can’t af-
ford Oliver North.

To meet the demands of speaking engagements and trade
shows, the bike had just added a mother ship, a large trailer pulled
by a nice van. The bike went inside the mother ship and hooked up
to the antennas in the trailer and a console in the van. The bike
continued to operate as the communications hub, even though it
was inside a trailer.

164

Mountain View

I thought that my three round-the-world trips for this book were
quite a journey, but I felt like a digital dilettante compared to Steve's
wholehearted commitment. His lifestyle was different from that of
your usual commute-to-cubicle engineer, but he has certainly
proved that you can have fun and put together a technical tour de
force at the same time.

I left Steve to drive up the peninsula to meet Ole Jacobsen, my
editor at ConneXions. Over a dinner of pasta with garlic, olive oil,
and anchovies, we plotted how we could get Steve Roberts to pedal
around the INTEROP show floor, hooking, of course, into the
ShowNet. How about remote SNMP bike management? A Behe-
moth MIB? The possibilities were endless.

165

San Francisco

Monday morning, I headed into the Mission district. With an hour
to kill, I walked past El Pollo Loco and the Carlos Club (“commut-
ers welcome”), to the Café La Bohéme, one of my old haunts. I
threaded my way through the couches and tables with derelicts
playing chess to get a cup of coffee.

In the corner of the café, I spotted something new. There was a
computer terminal with a sign saying “SF Net” and a message on
the screen indicating I should inquire at the counter. I paid a dollar
and received a ticket with a code on the back that entitled me to
fifteen minutes.

SF Net had the usual PC-based access, but decided to add bul-
let-proofed terminals in places like La Bohéme and the Brain Wash,
a combination laundromat, diner, and bar. The terminals had a
screen built under a glass table top and the keyboard was covered
with plastic.

This was your basic BBS, but with a California twist. After en-
tering the code for time, you were asked for a handle. I chose
“Carl.” You could then chat with online users, post messages to
bulletin boards, send mail, or even buy and sell used CDs.

Aside from the obligatory Love Chat, a digital singles bar, SF
Net boasted a large Metaphysics and Astrology area. Messages
from people like Spud Muffin and Venus Anemone examined the
relative merits of Jupiter and Mars to our daily lives.

Finishing my coffee, I walked through the Mission, past a man
piling whole butchered hogs into the back seat of his station wagon,
and cut over to 23rd street to Brewster Kahle’s gingerbread Victo-
rian house. Brewster was the architect of the CPU of the Thinking
Machines’ CM2, one of the most innovative computers on the mar-
ket. He was also a former member of the Artificial Intelligence
Laboratory at MIT, home of Marvin Minsky.

After delving into the arcana of message-passing protocols for
massively parallel processors, Brewster turned his attention to the
much more difficult problem of finding and using information on

166

San Francisco

networks. The result was the Wide Area Information Servers
(WAIS).

The living room of Brewster’s house was packed full of books
on subjects ranging from the works of Goya to those of Marvin Min-
sky. This was not surprising, since Brewster viewed himself as a
publisher and WAIS as the ultimate publishing platform.

WAIS has three pieces. A server has documents and makes
them available to the world, perhaps for free and perhaps charging
for access. A client gets these documents from one or many servers.
Clients and servers communicate using an enhanced version of the
239.50 protocol, originally developed to search online card catalogs.

Brewster ushered me into his office, where he sat down on a
beat-up old easy chair and balanced a keyboard on his lap. The
screen and rollerball mouse were conveniently nearby, making this a
highly comfortable work or play station.

There was no need to start up his WAIS client since it was al-
ready up and running. Deployed for only a few months on the In-
ternet, WAIS was quickly becoming a part of people’s routines, and
had certainly been integrated into Brewster’s daily work.

Brewster typed in a query: “Is there any information about Biol-
ogy?’ The query was sent, in its entirety, to the server of servers
that Brewster maintained, quake.think.com. Servers of servers were
no different than document servers, they simply kept a list of other
servers and a description of the information they maintained.

We got back a list of servers throughout the world that had in-
formation on biology, such as a database of 981 metabolic intermedi-
ate compounds maintained in the Netherlands. At this point, we
refined our query and sent it out to many servers simply by point-
ing to them on the screen. Servers returned lists of document de-
scriptions; pointing to those documents retrieved the full text.

WAIS makes very few assumptions about servers or clients. In
our query, we used natural English. The server has indexed docu-
ments by keyword and parsed our query to form the lookup. There
is no requirement that a WAIS server handle natural English, it can
simply take keywords, SQL queries, and even do fuzzy matches to
correct for poor spelling.

167

Exploring the Internet

The documents stored in the server can also vary. They can be
simple ASCII text, as in the case of things like the RFCs. Or, the
documents can be in the Macintosh PICT format, PostScript, or
might even be some virtual document, dynamically generated from
a database at query time.

Likewise, the clients can range from simple to complex. At the
most basic, the client used Telnet and VT100 emulation, an interface
known as simple WAIS (SWAIS). Clients can be much more intelli-
gent. On a Macintosh or NeXT, complex queries on multiple servers
can be monitored in a simple intuitive manner. Queries can be saved
and periodically reexecuted, allowing updates to information to be
periodically generated.

Most users of WAIS use a Macintosh interface called WAISta-
tion, not surprising since Apple worked with Thinking Machines on
the initial prototype of the system that was tested at Peat Marwick.

WAIStation uses two metaphors to structure the user interface:
storage in folders and searching as a conversation. You started a
search by asking a question. You get back document descriptions,
and those results are used to refine the query. Eventually, the query
returns the information the user wants. The query can then be
stored in a folder where it can be pulled up later and reexecuted. |

Queries in WAIStation can also be passed by looking for docu-
ments “similar to” those already found. The conversation thus con-
sists of asking for documents by a query or by example and
pointing to one or more servers. Servers can be remote, or can be
local document stores such as the entire hard disk or a folder con-
taining all e-mail ever sent or received by the client.

WAIS, with over 10,000 users in 24 countries by the end of 1991,
requires, for decent performance, at least a 9,600 bps connection into
the network. A local WAIS environment can easily be put together
using Ethernet, either as an alternative or a supplement to the In-
ternet-based servers, useful in places where Internet connectivity
had not yet achieved the necessary bandwidth.

When I talked to Brewster, he was working with Apple’s Ad-
vanced Technology Group on an even fancier interface called Rose-
bud, named after the sled in Citizen Kane. Rosebud would use a
newsroom metaphor for the user interface.

168

San Francisco

Users would give assignments to reporters, a process like the
WA\IStation query dispatch. Different reporters can specialize in dif-
ferent assignments. The user would then look at the results from a
“newspaper” window, which would show the documents recovered
by different reporters. Reporters can be instructed to reexecute their
tasks on a periodic basis, allowing the newspaper to be dynamic.

Several other client interfaces also exists. NeXT would be ship-
ping one with every workstation. XWAIS uses the X Window Sys-
tem and was similar to the WAIStation. Other interfaces are
available for GNU Emacs and DOS platforms.

The number of WAIS servers that had emerged in a few months
on the Internet was truly incredible. It started with a couple of da-
tabases from Thinking Machines, holding things like manuals and
bug reports. Thinking Machines also uses WAIS as a corporate in-
formation system and even saved all messages from all mailing lists
for future queries.

Other databases that had emerged ranged from CERT advisories
to weather maps to archives of USENET bulletin boards to the en-
tire text of the Koran, the Bible, and the Book of Mormon. Three
cookbooks, a huge poetry database, Supreme Court opinions, and
even the CIA’s World Factbook were online.

Between June 1991 and the end of the year, WAIS spread to in-
clude over 160 databases in 8 countries, including (of course) Fin-
land, Norway, and Australia. Brewster described WAIS as
“uncontrolled and uncontrollable,” as any good service should be.
For example, Finland had started keeping a server of servers and
soon it might be necessary to dispatch queries looking for servers to
several different places.

Brewster’s goal was to enable anybody with a computer, even a
lowly PC, to become a publisher. The first PC-based WAIS server
had recently gone online, running in somebody’s basement, and
Brewster was quite excited by the prospect.

Brewster’s interest in publishing was personal as well as profes-
sional. His fianceé ran a printing museum and in the basement was
an old printing press. For entertainment, friends went downstairs
and made calling cards. The only requirement was that they leave

169

Exploring the Internet

one on the wall. Cards from people like John Quarterman and Mi-
chael Schwartz adorned the walls of the basement.

After a Salvadorean lunch of pupusas and empanadas, I bid
Brewster goodbye and went down to the South Bay to thank Sun
Labs for the use of their machines and inform them of the ITU’s
imminent assassination of Bruno.

170

Moffet Field

Milo Medin is best known in the Internet community as one of the
developers of Open Shortest Path First (OSPF), the new and im-
proved dynamic routing protocol. He also runs the NASA Science
Internet (NSI), one of the largest operational networks and one of
the core backbones for the Internet.

NASA Ames Research Center, where Milo worked, is located at
the Moffet Naval Air Station near Mountain View, the site of an im-
pressive collection of wind tunnels that includes one big enough to
fit a full-size Boeing 727. Another tunnel can achieve hypersonic
speeds of Mach 9. The big tunnel is powered by six 28,000 horse-
power engines and takes so much power that it can only be run at
night (after first warning the power company).

NSI links all the big NASA laboratories, such as Ames, the Jet
Propulsion Laboratory, Goddard in Maryland, and the Kennedy
Space Center in Florida. It also links scientists in universities and at
research facilities in locations as diverse as the Antarctic, Greenland,
and Easter Island.

NSI is both a general-purpose Internet and a mission-oriented
operational network. As a mission-oriented network, NSI has some
fascinating applications with demanding resource requirements.

For example, a telescope at Owens Valley is connected to NSI,
allowing astronomers to reposition the telescope remotely. Another
set of applications distribute data from the Upper Atmospheric Re-
search Satellite, moving telemetry data from the satellite to White
Sands. From there, the data are moved to Goddard’s Consolidated
Data Handling Facility, where they are distributed over NSI to re-
mote analysis centers.

Perhaps the most unusual application I saw was a C141 airplane
used for infra-red data collection. The plane regularly flies between
Ames and Christchurch, New Zealand. The plane has its own Class
C network address and is filled with Sparcstations on an Ethernet.

The plane flies its route and collects data along the way. When
it lands, a thinwire BNC connection drops out of the back and con-

171

Exploring the Internet

nects to a cable in the hanger, making the plane a part, once again,
of NSI. The movement of the plane from one facility to another is
transparent to users. Routing protocols, such as OSPF and EGP, are
used to announce the availability of the plane network to the rest of
the Internet.

Keeping 70 routers, plus underwater fiber, satellite, and dozens
of high-bandwidth terrestrial links up and running is quite a chal-
lenge. In many cases, the links not only have to be up but also need
to provide adequate performance for the transfer of large images or
low latency delivery of operational commands.

The network is managed from within a Network Control Center
at Ames. The current status of the network is shown on a wall dis-
play, with bad links in red and good ones in green. Operators at
workstations monitor these links 24 hours per day.

Given Milo’s active IETF involvement, it is not surprising that
SNMP is used extensively to monitor network status. SNMP, how-
ever, provides in-band network management and does not help very
much when the network was down. To supplement SNMP, each of
the routers in the NSI network has a modem on the console port,
allowing operators at the Network Control Center to dial in and di-
agnose problems, often while the users on-site sleep.

While TCP/IP is the main protocol suite in NSI, there is also a
heavy DECnet contingent. Unlike TCP/IP, DECnet Phase IV has no
globally administered address space. The architecture assumes a
single network which has up to 63 areas, each area under some
management entity.

When a global community starts to interconnect, as in the case
of space sciences or physics, you quickly get many more than 63
areas. DEC had a hard time believing that physicists would need
more than 63 areas. “Just get organized,” they would say.

“Just get expletized,” the physicists would respond.

To solve the problem, groups like NSI and HEPnet have been
forced to resort to hidden areas. NSI and HEPnet agreed to make
areas 1 to 46 globally administered areas, available to the entire
physics and space sciences DECnet communities. These areas mean
the same thing to all routers in NSI. CERN, for example, used area
22 under DECnet Phase IV and area 23 for Phase V migration.

172

Moffet Field

Areas between 47 and 63 are reusable. A particular DECnet
router would know about one of these areas, but another router in
the network might use that same area number for a totally different
network.

On many links, DECnet and TCP/IP had equal status, being
sent across by multiprotocol routers. Occasionally, for political or
technical reasons, the underlying link could only transfer TCP/IP
traffic. A process known as “tunnelling” encapsulates DECnet traf-
fic inside of IP packets, making two DECnet routers appear to be
next to each other.

NSI is both a special-purpose network and a general-purpose
Internet. Owning the links means that NSI is able to control them,
helping to tune the network for the bandwidth and latency require-
ments of specific end-to-end requirements. This doesn’t mean that
the network is isolated, only that NSI planners have a fairly good
idea of what is happening on their part of the Internet.

NSI also acts as one of the Internet backbones. Links to New
Zealand, Hong Kong, Australia, Hawaii, London, and many other
sites come in at NASA Ames, Goddard, or other NSI sites. The data
from these sites is routed on through to its final destination, which
might well be outside of NSI.

One of the key links into the rest of the Internet, indeed one of
the two key hubs in the entire internetwork, is at Ames. This is the
Federal Internet Exchange for the West coast, known as FIX West.
The other FIX is in College Park, Maryland, and is called Fix East.

FIX West is located in a small squat building in the middle of a
sweet potato patch on the edge of Ames. The building is shared by
NSI and the NASA Telephone Company, which maintains the un-
derlying bandwidth on which NSI runs.

Inside the windowless building, Pacific Bell has the termination
for 12 fiber optic cables. Two other sets of 12 fibers terminate at
other locations on site to provide redundancy. The fiber provides 30
to 40 T1 links, but could easily provide as much as a dozen OC-3
links running at 155 Mbps each.

The room is packed with termination equipment, patch panels,
and other telesoterica. In a locked wire cage are all the routers for
NSI. One 19 inch rack holds routers for the Bay Area Research Net-

173

Exploring the Internet

work (BARRnet). Another has NSS routers for the NSFnet. Other
racks hold equipment for local networks, such as the link to the
huge National Aerodynamic Simulation center at Ames. Yet more
racks have routers for the Energy Sciences network (ESnet), NSI,
and many others.

In a corner is a huge, antique piece of equipment, known as a
Butterfly. The Butterfly is made by BBN and is still used by MIL-
NET as a “Mailbridge” to the outside world. Prying off the top of
this ancient-looking device, Milo showed me the Macintosh hidden
inside which is used as the console.

Also located in this building are the electronics for two complete
C band satellite earth stations, not to mention videoconferencing
equipment, racks for the experimental Terrestrial Wideband net-
work, time servers, and a host of other equipment needed to pro-
vide the gateway for the west coast, the Pacific Rim, and much of
the rest of the Internet.

Finishing what Milo called the “cheap tour,” we drove by the
huge hangers of Moffet Field, originally built to house dirigibles and
so immense that clouds form inside when the doors are closed.
Milo stopped at another hanger, where we saw Harriers, Hornets,
U2s, and many other NASA aircraft being worked on. The U2s
were used to pinpoint firestorm centers during the Yellowstone
fires, helping the firefighters to identify targets.

After a nice lunch of generic Chinese food, I borrowed a termi-
nal to read my mail. Being in the center of the world, I figured my
response time ought to be good, and it certainly was.

Heading out, I stopped first at the visitors center. Here you can
get all sorts of educational books on atmospheric sciences and scale
models of historic aircraft. I headed straight for the astronaut food,
Where I bought five pouches of freeze-dried, vacuum-packed ice
cream, a chalky, marshmallow-like bar that sort of resembles ice
cream if you freeze it and squint your eyes.

174

Wellin gton

Tuesday night, New Year’s Eve, I boarded the plane for New Zea-
land. Travelling west on New Year’s Eve gives you at least two
excuses to celebrate, provided you cross the international date line
after midnight. If you cross before, you get no chances unless, of
course, you don’t tie your celebrations to arbitrary chronometric in-
dicators.

Twenty hours later, I landed in Auckland to await my flight to
Wellington. Customs impounded the salami I had bought as an of-
fering for the restaurant critic of The Bangkok Post and told me that if
I did not reclaim my links within 28 days they would be destroyed.
Not wishing to be responsible for the wanton destruction of four
innocent salamis, I paid my NZ $5 fee for a sausage hotel at the
Ministry of Agriculture and Finance, and assured the officer that I
would return as soon as I could.

I must admit that in the course of travelling, I’ve gotten a bit
jaded about waiting lounges, but the Air New Zealand Koru Club in
Auckland was a refreshing change. After a breakfast of fresh kiwi
salad and coffee, I was offered the use of a shower, razors, and the
daily newspaper. Computers, Xerox and fax machines, and other
office equipment were warmed up and waiting.

Arriving in Wellington, I was met by Richard Naylor and his
two sons, Jeremy and Chris. Richard runs the computer department
for the city of Wellington; I had come to learn about City Net.

Even though it was the middle of summer vacation, a “south-
erly” had blown in and it was cool and windy. Richard dropped
me off at the St. George hotel for a nap. Before going to sleep,
though, I had to flush the toilet a few times just to watch the water
swirl down in the opposite direction from what I was accustomed
to. Having empirically narrowed my geographical coordinates to
the Southern Hemisphere, I went to sleep.

Two hours later, Richard and the kids picked me up for a tour of
New Zealand’s capital. We went first to the new NZ $150-million
city complex just being finished on the edge of Lambton harbor.

175

Exploring the Internet

The complex includes not only the city offices, but a hands-on sci-
ence museum for children, an art gallery, a huge library, and many
other facilities for the public.

The Council in Wellington used to own everything from a milk
bottling company to the bus service and the power company. Al-
though many of these functions have been privatized, Richard’s
group still provides computing services to many of them. The col-
lection of computers is known as City Net.

Most of the computing power comes from a large DEC installa-
tion, including nine big VAXen and a dozen pVAXen. Over 60 giga-
bytes of disk space are available, most of it in a VAX Cluster. The
cluster provides disk striping, volume shadowing, and dual porting
of disks, all facilities that help ensure good performance and reli-
ability. In addition to the DEC equipment, there is an AS400 used to
automate the library, as well as 50 PCs, a Honeywell, and an IBM RS
6000.

Since the city used to own the tram service, Richard was able to
pull fiber throughout the downtown area, stringing it above the
tram lines. Inside the buildings, unshielded twisted pair was used
to wire the individual offices. Both voice and data went over this
common infrastructure.

With a staff of 55, the computer dearth is able to deploy an
impressive number of applications. A layered mapping database
has the entire city of Wellington in it, including everything from
sewers to phone lines to all roads and buildings. The mapping sys-
tem is extensively used by city employees ranging from planners to
mechanical engineers, and is even available to the general public
through terminals in the library.

What is most interesting about City Net is its open philosophy.
An Ultrix machine on the edge of the network is on the Internet and
city employees have access to the full range of Internet services.
City Net even provides accounts to the general public on the gate-
way system. Over 500 citizens of Wellington use the machine at no
charge.

Wellington has a very active bulletin board community with
over 35 boards in a city of only 150,000 people. Many of these
boards use the Waffle BBS software, which includes support for the

176

Wellington

UUCP protocols, making integration of the bulletin board commu-
nity into the Internet as simple as dialing up the City Net hub. That
night, I watched 9-year old Jeremy composing messages for a pen
pal in Pittsburgh on a little Toshiba laptop running the BBS soft-
ware. The minute Jeremy was done, his older brother grabbed the
machine. Periodically, the laptop would be plugged into an RJ11
jack and mail exchanged with the central system.

Everything from a World Wildlife Fund training center to the
children’s museum to the milk company accessed City Net. Richard
even wanted to put a hospitality database on the Internet to allow
tourists to Telnet in and plan their vacations.

City Net is one of the best examples of a network as a truly
public infrastructure I have seen. Extensive use of e-mail by chil-
dren as part of Kids Net, the Discovery Museum, or just individual
correspondence with pen pals helps to ensure that there will be a
new generation of computer-literate children.

Providing a public utility to organizations like the milk company
or the general public is done at very little marginal cost to the city.
Many of the resources, such as the Internet gateway, are viewed as
necessary for the city to do its job. Extending that service to the
public or to commercial enterprises helps the city as a whole, as
well as the direct users.

Invited for dinner to Richard’s house, I brought two packets of
Astronaut Ice Cream, which turned out to be the perfect gift. Chris
was heavily involved in the local chapter of the Astronautics Asso-
ciation of New Zealand and had heard of this astronomical delicacy
only through rumors. Chris wanted more information for his club,
so his dad suggested that he Telnet into the NASA public server or
send e-mail to some of the online PR hacks. Chris wanted my e-
mail address so he could ask me questions later.

After a delicious dinner with a truly computer-literate family,
Richard dropped me off back at my hotel. I had a Steinlager beer
and watched steer wrestling from Montana in the bar before going
up to pack for the next day’s trip.

177

Dunedin

Saturday morning, I gulped down some fresh strawberries and a
savory and took the plane down to Dunedin, on the southern tip of
the South Island. The plane flew by the Southern Alps, then headed
into the white-speckled hills of sheep country. New Zealand has 65
million sheep and 3 million people, and the South Island is the site
of huge farms.

In Dunedin, I met Ian Forrester, Chief Scientist of the Ministry of
Research, Science, and Technology. A former professor at Otago
University in Dunedin, he keeps a house there even though he lives
most of the time in Wellington.

The house was an old farm house, honeycombed with gardens
and pastures for the horses and sheep. Ian’s son, a medical student
at Otago, lives there year round and Ian had come down for the
summer holidays. We sat in a rose garden and drank Speight’s Beer
(“Pride of the South” as the label reads, though in much the same
sense as Stroh’s would be considered the pride of Detroit) while
Ian’s son cooked up some artichokes from the garden.

Ian started at Otago as a biochemistry professor. He was mar-
ginally interested in computers, using tools like Medline to search
databases and using an early Apple computer. His main interests,
though, were medicine and agriculture. Agriculture is the key in-
dustry in New Zealand and Ian was studying problems related to
livestock.

Although bats were the only native mammal to the country,
deer, elk (brought in by Teddy Roosevelt), and lots and lots of sheep
had taken root. Ian’s interest was in helping find ways to keep the
population of animals genetically healthy. On a Ford Foundation
fellowship in Laramie, Wyoming, he looked at topics like how to
freeze elk semen to increase the diversity of the genetic pool.
Throughout the 1980’s, Ian continued this line of research. He spent
a few more years at Otago, got interested in biotechnology, and
went to Madison, Wisconsin for a few years.

178

Dunedin

Meanwhile, New Zealand was undergoing a revolution. The
1987 stock market crash had reduced the stock index by one-third
and recovery was only marginal. GDP was growing by only a frac-
tion of a percent.

In this dismal economic climate, and with a bloated government
bureaucracy, New Zealand initiated a drastic deregulation cam-
paign. Many large government departments became state-owned
enterprises and many were then sold off. The telephone company,
for example, was sold to Ameritech and Bell Atlantic.

Even traditional government departments began to be run like
businesses. Heads of government departments were called chief ex-
ecutives. These executives negotiated contracts with specific output
goals with the cabinet ministers for the area. Government employ-
ees were put under contracts.

Change was equally drastic in the area of government-spon-
sored research. New Zealand had long emphasized government
laboratories over university efforts for research. The government re-
searchers were spread out in different ministries, such as agriculture
and forestry.

A small ministry of 30 people, to which Ian belongs, was formed
to advise the cabinet on national research policies. Once the cabinet
decided on national policy, money was funneled into a Public Good
Science Fund, administered by a foundation. Researchers who
wanted money submitted bids and, if they were awarded, signed
contracts. Researchers were no longer guaranteed money if their
work did not meet the strategic directions set out by the cabinet.

The next step was to remove the research efforts from the opera-
tional arms of the government. Ian and others worked as part of a
committee which recommended that 10 independent research insti-
tutes be established, each headed by a chief executive. This plan
was approved in June 1991 and had been implemented so quickly
that the institutes were opening their doors a mere one year later.

As Chief Scientist, Ian was both scientific advisor to the govern-
ment and an advocate for researchers. In going around and talking
to people, he quickly saw that getting computers and communica-
tions onto the desks of researchers was a way to help the entire
scientific community.

179

Exploring the Internet

Although there was a university network in place, linked to
PACCOM, each of the government research groups also had their
own networks. The technical level of networking was uneven
throughout the country and interconnection was sporadic.

With independent research institutes being formed, it was evi-
dent that some form of unified research network was needed. As
part of the proposal to form the Crown Research Institutes, a recom-
mendation was inserted to make a national research network a stra-
tegic direction. When the cabinet approved the divestiture plan for
the institutes, a side benefit was approval for the national net.

Working out the mechanics for institutes and networks kept Ian
busy most of the time, but as the first occupant of the post of Chief
Scientist, he also had some freedom to help define his job. Looking
at a national research network exposed Ian to topics like broadband
networks. If this was the wave of the future, was there some way to
speed up the future?

This was the genesis for the idea of making New Zealand into a
world communications laboratory. Ian started to push the idea that
the country could become a testbed for new technologies. Massive
deregulation, small size, and an educated workforce would all make
New Zealand an attractive place for large corporations to deploy
and tune new technologies.

With unemployment at 10 percent and GDP flat, this also might
be a way out of New Zealand’s economic doldrums. Applications
could be deployed to make industry more competitive and the
workforce could be trained in high technology. He prepared a pa-
per on the idea for the cabinet which authorized an in-depth study.
When the second briefing paper was submitted, the cabinet asked
Ian to start working with the private sector to develop a business
plan.

In New Zealand’s climate of deregulation, the World Communi-
cations Laboratory was certainly not going to be a large government
project. Instead, lan was trying to form a coalition, led by govern-
ment but working with industry.

How could New Zealand get a state-of-the-art telecommunica-
tions system, the prerequisite to becoming a world testbed? Ian was
hoping to follow the example set by the State of Connecticut. By

180

Dunedin

pooling the purchasing power for telecommunications services and
promising a long-term commitment, the state was able to convince
Southern New England Bell that it was a sound business decision to
invest in a large fiber network.

When I spoke to Ian, many of the details of the World Commu-
nications Laboratory were vague; the project was long on vision and
short on substance. However, an active campaign was being waged
to convince large corporations inside and outside of New Zealand
as well as the general population that the project was viable.

The project was publicly unveiled to people in New Zealand in
November 1991. To help build a constituency, Ian emphasized the
benefits of high technology to traditional industries such as agricul-
ture. He painted a picture of food stores in England using interac-
tive video to decide how they wanted lamb chops cut. The
specifications would go straight to the meat packing plant, which
would scan the carcasses and make the optimal cuts. Just-in-time
inventory management for the meat-packing industry.

At the same time, he was trying to drum up support from large
multinational corporations for the project. He held a telephone con-
ference with participants from groups like American Airlines and
Citicorp to brief them on his vision.

Ian had adopted the “highways of the future” metaphor of Sena-
tor Gore with a vengeance. He saw the ports of the information age
being established, and he wanted New Zealand to be one of those
ports. The challenge would be building a political consensus and
then performing the hard technical work to turn the vision into real-
ity.

Ian Forrestor’s World Communications Laboratory might or
might not get off the ground, but this high-level policy focus on
information technology was an important first step. Even if the
only result turned out to be a good national research network, Ian's
efforts would have paid off handsomely.

181

Auckland

Saturday morning, I flew up to Auckland. With the whole weekend
to kill, the largest city in New Zealand seemed to be a likely hunting
range. Unfortunately, Auckland in January is kind of like Paris in
August. A one-hour walk through town uncovered various flavors
of tour groups but no Kiwis.

It was time for guerilla tourism. I reported to my local Hertz
dealer and negotiated a rental. Lacking any knowledge of New
Zealand, I flipped a coin and headed North. One hour later, I had
successfully managed to stay on the left side of the road and hadn’t
run over any of the locals, so I left the motorway and got onto a
sometimes-paved road down the Western shore.

I ended up at Muriwai beach, one of the few nesting sites for the
gannet. After watching mama gannets regurgitating fish into their
offspring, it felt like dinner time so I headed back into Auckland.

The pretty attendant at the Quality Inn suggested three restau-
rants for me. At the first, located right on the water, the maitre d’
sniffed when I admitted to not having reservations and I was
grudgingly shown to a table with a view of the wall next to several
screaming infants. I declined.

At the second restaurant, I was summarily dismissed for not be-
ing a member of a tour group. At the third, I was told I could have
a table, but would have to vacate it within the hour for the next
booking.

With visions of being tossed onto the street in mid-potato, I hit
the streets to go hunting. Ten minutes later, I spied a hip-looking
couple dashing into a nondescript door. I followed them into a
funky little place that called itself Four Steps to Heaven.

Four Steps to Heaven had a half-dozen tables and menus writ-
ten on chalk boards. The grizzled cook in a white chef’s hat peered
out from behind his counter to see if customers were enjoying them-
selves. The waiter had a long spiel about the menu prepared and
insisted on going through the talk in its entirety (despite the fact
that I had heard three other tables receive the identical lecture).

182

Auckland

The cuisine was absolutely first rate. I started with smoked
salmon from the South Islands, served with a mustard sauce dusted
with paprika, nestled in a sea shell. My main course was the most
impressive rack of lamb I’ve ever seen served with a piquant mint
sauce. My table was so small I kept kicking my laptop in the case at
my feet, hoping not to damage my rack of RAM.

Accompanying the lamb was a large platter of vegetables. The
cauliflower was steamed and covered with a delicate cheese sauce.
The potatoes were gratineed with onions. The piece de resistance,
though, was zucchini, doused with ouzo, tomato juice, and hot pep-
pers, and cooked with sesame oil.

The next day, I drove south to the harbors and bays on the east-
ern coast, cutting through farm land and country homes. Refreshed
and ready to work, I dropped my car off and went home to pack for
the next day’s trip to Australia.

183

Melbourne

Sitting in the Auckland airport lounge, the power went out. The
two ladies next to me immediately blamed the outage on computers
and spent the next ten minutes yammering on about how comput-
ers had ruined everything and had given rise to a younger genera-
tion devoid of any useful skills.

“They can’t even add without a calculator,” one sighed. I re-
strained myself from commenting that she probably couldn't pro-
gram a bubble sort to save her life. The large umbrella she kept in
readiness at her side contributed to my reticence.

Arriving in Melbourne amidst a spate of “g’days,” I presented
my gift salamis at customs for inspection. Despite my pleas, the
Sausages were impounded and destroyed, leaving me without a
food offering for my upcoming visit to Bangkok. Without a food
offering, I had no bribe to motivate my friend the restauraunt critic
and I had visions of him exacting his revenge by taking me out for
Swiss steak and succotash.

I deposited my bags at my hotel and set off to try and read my
e-mail. I took a tram towards the city, crossed over the Yarra River
and Batman Avenue and headed up the hill to the University of
Melbourne. There, I found the location of the Information Technol-
ogy Center on the map and sat myself down in a large room filled
with PCs and Macs. I picked a likely looking 386 and typed “‘tel-
net,” and sure enough, a few seconds later I was happily deleting
messages.

Since I wasn’t expected until the next day, I walked back down
the hill to find the world famous Royal Botanical Gardens. Both a
formal garden and an arboretum, this is the kind of place you find
written up in coffee table books and issues of the National Geo-
graphic.

The gardens certainly deserve their reputation. I walked past a
cluster of Hoop pines, past a huge Monterey cypress planted by
Prince Leopold in 1910, and through a grove of small but pungent
Bhutan cypress. Next to all these evergreens was the Canary Island

184

Melbourne

Bed, with a rare arrow-wood tree surrounded by cactus-like succu-
lents and asteraceae.

Further down the path was a row of aloe plants ranging from a
few inches tall to huge bushes. Spiny agave and Mexican Yucca
were mixed in, giving the desert-like setting a sharp contrast to the
California air of the evergreens.

Below the beds was an intricate ornamental lake filled with
ducks and black swans, edged with English Elms, Irish Strawberry
trees, and groves of bamboo and Brazilian Pampas grass. To my
disappointment, that evening’s showing of A Midsummer Night's
Dream was sold out, so I set out to find some dinner.

The closest I could find to local food was an Aussie Burger, a
nice hamburger topped with the traditional bacon, egg, and beets. |
headed instead up to Toorak Road for some Italian food.

Most of the restaurants were BYOs, so I went into the bottle
shop and picked up a half bottle of local Cabernet Sauvignon, then
went into Molina’s Bistro, truly a fortuitous find. The appetizer was
a Cervello alla Grenoblaise, lamb brains coated in beer batter and
served with a delicate sauce of chives, capers, and lemon wedges,
garnished with sprigs of fresh basil and fennel. It was accompanied
by fresh, warm bread, dusted with basil and oregano from the herb
garden out back.

My main dish was Saltimbocca a la Romano, baby veal in a
sauce of sage and white wine, and served with a timbale of potato,
onions, mushrooms, and cheese. Dessert was a home-made liqueur
ice cream. Stuffed, I walked back to the hotel to get ready for my
meetings the next day.

Tuesday morning, I went back up the hill to find Chris Chaundy,
Networks Manager for the University of Melbourne. The university
network connected 74 buildings with a mix of fiber, unshielded
twisted pair, microwave, and leased lines. Originally a massively
bridged network, the university was rapidly completing the transi-
tion to a network based on routers.

185

Exploring the Internet

The university is split into Mac and PC camps. The Macs used
AppleTalk and 53 Banyan Vines servers take care of the PCs. The
backbone network is mainly TCP/IP, although large DECnet and
CDCNET applications are still running.

AppleTalk clients access the TCP/IP world using a Webster mul-
tiport gateway, a device similar to the Kinetics Fastpath. The mul-
tiport gateway was developed on campus. before being
commercialized by Webster.

All told, the network is your basic professionally-run, produc-
tion environment. Large machines such as a Cyber 990, an IBM
3090, and a Maspar are all on the net. The library card catalogue,
WAIS, X.500, and all the modern information services you would
expect are all available.

This is one of those operations that has everything you would
want. The information technology center has a regular, frequent
schedule of training classes, extensive documentation, a custom pro-
gramming group, a proactive help desk, spacious labs, and a trouble
ticket system.

This is great for the university community, but certainly posed a
challenge for the digital tourist. After all, what can you say about a
place that does everything right? |

After my briefing with Chris, I went across the street to meet the
elusive Robert Elz, a long-time inhabitant of the Internet. Robert
rarely works, or, for that matter even wakes, during the day, but I
was in luck. Cricket matches were on this week and Robert had
inverted his chronology to watch the games and post the results into
netnews.

Robert’s desk was stacked a few feet deep with paper, leaving
only enough room for a keyboard and a half-dozen empty Coke
cans. Robert had just returned from giving seminars in Thailand
and the top of his stack included Thai-language keyboard templates.
His wall had a map of all undersea cables in the Pacific region.

After some chitchat with Robert about mutual friends in Bang-
kok, I had worked a full two hours and then it was tea time. Back
at the hotel, sitting in the lobby, I listened to the drivel of a lounge
lizard playing a baby grand piano.

186

Melbourne

After the second drink, I realized that the piano was unmanned.
The thing had broken into a jaunty Elton John tune, so I walked
over to investigate. Under the piano was a Yamaha Disklavier, a
digital version of the old player piano, hooked up with solenoids to
activate the keys and provide a very exact replica of the original
performance. The Elton John tune was from the album “Don’t
Shoot Me, I’m Only the Piano [Player].”

187

Sydney

Wednesday evening, I flew in over the Sydney opera house and,
after checking in, darted across the street for a quick bite of barbe-
cued octopus. With a little time to kill, I strolled through neighbor-
ing King’s Cross, the home of Sydney’s red light district. Sipping a
Foster’s, I watched the strange melange of prostitutes looking for
their first action of the night, world-weary backpackers trying to en-
tertain themselves on a dollar, and a hip-looking hippie driving an
antique Jaguar. Several Hare Krishnas trooped by, reminding me
that it was time to go to work.

I was soon met by Bob Kummerfeld, chairman of the University
of Sydney’s Computer Science department. We dodged raindrops
from my hotel over to the Hernandez Café, where we sat and talked
over a café latte and a cappucino. Bob Kummerfeld was part of the
first wave of networking, sparked by the Landweber symposiums.
Bob attended his first one in Dublin, and ended up hosting the 1989
symposium in Sydney, the last one before the informal seminars
metamorphisized into the much larger INET conferences.

In the late 70s, armed with VAXen and PDPs, Australia, like eve-
ryplace else, wanted to connect their computers together. Running
an early version of UNIX, the obvious choice would have been to
use the UUCP protocols. UUCP had some features that not every-
one liked. Its designers had intended UUCP as an interim solution,
but, as often happens with prototypes, it rapidly became en-
trenched.

Kummerfeld and his partner, Piers Lauder, decided that they
could do better and set about designing an alternative set of proto-
cols. The result was the Sydney University Network, known as the
SUN protocols. The SUN protocol suite started out being used at
Sydney, the software was quickly picked up by Robert Elz in Mel-
bourne, and, by 1988, 1,000 sites were on the network. Since the
acronym SUN was identified with the Stanford University Network,
the collection of computers that used Kummerfeld and Lauder’s

188

Sydney

protocols eventually became known as the Australian Computer Sci-
ence network (ACSnet).

SUN is a protocol suite optimized for message handling. Unlike
TCP/IP, it was not designed for interactive services such as Telnet.
Message handling can be much more than e-mail and the SUN pro-
tocols supported services such as file transfer and remote printing.

The unit of transfer is the message, which can be infinitely long
(although in practice, very long messages are broken up into man-
ageable pieces). While UUCP performs error correction as it trans-
fers the data, SUN uses a streaming protocol to send the entire
message, identified as a series of packets. When the message is
transmitted, the receiving end sends back a bitmap of corrupted or
missing packets and those packets are then retransmitted.

When a connection is established between two hosts, UUCP
only supports transfer of a single message in one direction. SUN,
by contrast, is multichannel and full duplex. Full duplex means that
messages can go in both directions. Multichannel means each direc-
tion is divided into four logical channels of different priorities.

Channel 1 is the highest priority and is used for network man-
agement messages. An example would be a message to decrease
the packet size from the default of 1 kbyte down to the minimum of
32 bytes. On a dirty line, small packets reduces the amount of data
to be resent later since an error would affect a smaller chunk of the
message. It does so, of course, at the expense of throughput.

The other three channels are used for different kinds of message
transfer. E-mail messages would typically travel on channel 2, al-
lowing mail to get through even though a large file might be occu-
pying channel 4. While it might take a longer time to get a large file
through a link, it does mean that other applications are not frozen
totally off the line.

With multiple channels sending data, a cut in the line can inter-
rupt several different messages. Rather than start sending the mes-
sage again, SUN allows the session to start up where it left off,
something that UUCP does not do.

Use of multiple channels also tends to improve line utilization.
When a message is received, the destination node must typically
perform some processing. In a single channel application, the line

189

Exploring the Internet

would be idle during that time. With four channels, data transfer
can continue.

Riding on top of the message transfer facility are applications.
The suite has standard facilities such as file transfer and electronic
mail and users can design their own protocols.

The file transfer protocol allows both senders and receivers to
initiate transfers. To send a file, for example, the recipient’s user-
name and host are specified. The file would be spooled onto the
target machine and e-mail sent to the recipient. The recipient could
then chose whether or not to accept the file, moving it into personal
disk space.

E-mail supported standard Internet-style mail based on RFC 822,
making a gateway to the TCP/IP mail system almost a trivial task.
E-mail with the SUN protocol suite allows the transfer of arbitrary
binary images, a feature that has been only recently added into
SMTP and RFC 822.

The architecture of messages and packets used in SUN allows a
wide variety of different substrates to be used. Direct lines, mo-
dems, and X.25 are all used and the protocols can run on top of IP.
Bob even claimed that the protocol could run successfully on tele-
graph lines, although nobody had yet had a pressing desire to run
applications at 50 bps.

While Bob went on and described features ranging from dy-
namic routing protocols to broadcast and multicast support, I began
to wonder why I was using UUCP on my home systems. The an-
swer was fairly simple: UUCP was free and SUN wasn't.

Kummerfeld and his partner had formed a company called Mes-
sage Handling Systems to market the software. The University of
Sydney had a one-third stake in the company and the software was
being pedaled as MHSnet. Formed in 1988, the company was still
in its early stages, but Bob told me about a few recent successes.
One was a large insurance network that links brokers together. The
other was the Australian diplomatic corps, which would use
MHSnet to link their worldwide operations.

Both groups were attracted to MHSnet by its ability to run over
a wide variety of different transports, ranging from telegraph to IP.
The message-based paradigm is quite appropriate for environments

190

Sydney

with only sporadic connectivity. The software is small and efficient,
taking less than 100,000 lines of code.

Selling MHSnet to the government brought up the spectre of
GOSIP.

“Didn't you have problems with OSI bigots?” I asked.

OSI was certainly an issue, he explained, but eventually it came
down to practical issues such as whether software existed that could
do the job. OSI as a general architecture certainly had appeal to
many, but a lean and mean OSI that could run on low-quality, low-
speed lines was not readily available. Indeed, one could argue that
searching for such a beast was akin to looking for a hippopotamus
capable of doing the limbo.

To Bob and many others in the research community, OSI has ac-
tually had a negative effect. If he proposed to do research on mes-
sage handling, for example, somebody would invariably suggest
that X.400 had already solved the issue and further research was
superfluous.

Bob’s theory was that the OSI wave had finally crested and peo-
ple were settling on an environment characterized by IP over Foo
and Foo over IP. OSI had promise, but while large groups had large
meetings in a large number of locations, the IP community had
mounted an equally ambitious but much more successful standards
effort.

The IP process was organized to suit the needs of the technical
people, not the desire to reach a political consensus among market-
ing representatives. The crucial difference between the two groups
was that implementation and testing was the cardinal rule for the
engineers. Bob cited the recent extensions to SMTP and RFC 822 to
handle different kinds of data as an example. By the time the im-
plementations reached standards status, a half-dozen implementa-
tions would already exist and be deployed.

The success of the Internet community didn’t necessarily mean
that there was no role for the more formal (i.e., slower) procedures
like those used at the ITU. Bob pointed to a low-level substrate,
such as the definition of a digital hierarchy of line speeds, as an
issue that needed to be codified, ratified, and digested by a hierar-
chy of committees.

191

Exploring the Internet

Of course, if low-level issues are the province of the public
standards cartel in Geneva, that kind of leaves OSI out in the cold.
After all, OSI starts work at about the same level as TCP/IP does.
Since the OSI standards potatoes insist on defining theory first and
leaving the implementations to “the market,” they have little to con-
tribute to the upper layers of the network, where implementation
and testing are crucial.

I bid Bob Kummerfeld goodbye and went to my room to watch
Dallas. It struck me that the politics of the Geneva standards cartel
had an overwhelming resemblance to the goings-on at the Ewing
Ranch. Perhaps I could write a nighttime soap called “ISO,” com-
plete with a season finale of “Who Killed FTAM?”

192

Canberra

Canberra became the capital of Australia for the same reason Wash-
ington, D.C. became the capital of the U.S. In Australia’s case, Can-
berra was chosen because it wasn’t Melbourne and it wasn’t Sydney
and because it was somewhere between the two.

Canberra is also the capital of the Australian Academic Research
Network (AARNet), a network managed by Geoff Huston. The net-
work came into being after ACSnet began to feel the strain of, as
Geoff puts it, “networking on two bob.” ACSnet was strictly a vol-
unteer effort between the computer science departments, with no
management and no funding.

DEC’s educational group had been providing international ac-
cess, but after somebody took over the pipe for two days for a mas-
sive file transfer, they pulled the plug. Robert Elz, who had been
handling most of the international liaison work, arranged an X.25
connection to UUnet in Virginia.

X.25 costs money, so Elz wrote software to provide accounting
and chargeback to the users. Then, the first leased line, operating at
2,400 bps, was installed between Melbourne and Sydney. Network-
ing was starting to cost money and it was evident that a more coor-
dinated effort was needed.

About this same time, the central computer departments de-
cided that networking shouldn't be left to computer science. The
computer departments put together a proposal for a national re-
search network based on X.25 and the U.K. Coloured Book stand-
ards.

This group went to the federal government, but the feds had no
interest in funding a national network. Next, they approached the
committee of Vice-Chancellors, a group of the heads of all 38 uni-
versities formed to lobby for the interests of higher education
against other groups competing for a slice of the federal pie.

The Vice-Chancellors were interested, but weren’t about to just
hand the computer centers carte blanche. They initiated their own
feasibility study. With the feasibility study, they decided the idea

193

Exploring the Internet

was worth pursuing and, in January 1989, formed two groups to
develop the idea.

The first, the technical group, met for the first and last time that
same month. “It was bloody obvious what needed to be done,”
Geoff said. The network would simply be multiprotocol routers
and leased lines. No code to write, no research issues to investigate.
You simply took out the price books and tariff charts and figured
out what you could afford.

The policy and steering group only met three times. The Vice-
Chancellor who chaired that group made it clear that they were
there to build a network, “so let’s get on with it.” Geoff Huston is
credited by many in Australia for helping to make sure that the
process moved smoothly in both the policy and technical groups.

By November 1989, everything was done. Geoff placed orders
for leased lines and Cisco routers, telling both vendors he wanted a
delivery date of April. He told Cisco how he wanted the routers
pre-configured. With that much lead time, there was no excuse for
delivery problems and none occurred.

In the last week of April, all the boxes were in their locations
and the 48 kbps lines were up and running. For the next three
weeks, Geoff and his one other staff member toured the country
turning on the boxes and connecting them to the campus Ethernet
and the leased lines. In a total of three weeks, the entire network
was up and running.

The network continued to be run with only two people, even
though it had expanded to 2 Mbps lines on key routes and had a
huge growth in traffic. By January 1991, over 40 Gbytes per week
were being transferred on the network. One year later, 77 Gbytes
per week went through, and peaks of 120 Gbytes per week were not
uncommon.

The lack of religious battles over the network and the clockwork
installation are widely credited to Geoff Huston’s deft leadership.
He defused any attempts to make the network overly complicated
and insisted on a multiprotocol backbone so that any traffic the cus-
tomer wanted to transfer could be transferred. He showed that in-
stalling a national network can truly be no fuss, no bother.

194

Canberra

While hooking up the network internally was straightforward,
there still remained issues in the Internet as a whole that were hav-
ing a real effect on the operation of AARNet. Geoff gave me a dra-
matic example of one of these issues.

He started a traceroute to Barry Shein of Software Tool and Die,
a well-known member of the Internet community. The traceroute
went out from Canberra to the AARNet routers, reached NASA
Ames and then failed. Barry is connected to the Internet via one of
the Commercial Internet Exchange (CIX) members and AARNet
connects to the Fix West. Since the NSFNET had refused to allow
its routers to announce the presence of networks that had not signed
the Appropriate Use Policy (AUP), the result was a fragmented In-
ternet.

Software Tool and Die was a classic example of commercializa-
tion of the network. Barry Shein started the operation as a commer-
cial IP service provider, giving over 1,000 people access to the
Internet for an hourly fee, exactly the kind of thing the AUP warned
against. Software Tool and Die could have waffled the issue, sign-
ing the policy and issuing severe warnings to its users, but that
seemed kind of silly.

Since Software Tool and Die was an AlterNet hub, Barry Shein
estimated that they had access to over 80 percent of the Internet.
AlterNet was on the CIX, and most regionals had set up back door
paths to one or the other of the commercial networks. Several com-
mercial links to Europe existed, making international access to the
EBONE and its regionals straightforward. In the U.S., only a few
regionals in the Midwest had a single connection to NSFNET and
were unreachable by Software Tool and Die customers. Most of the
Pacific Rim, however, squeezed through the single Fix West link,
and was cut off.

Geoff Huston was quite adamant that the Internet was facing
many serious issues, fragmentation being one of the more visible.
The address space was quickly filling up, there were so many net-
works that routing protocols were thrashing, and a host of other
problems were surfacing.

These problems were not inevitable in Geoff’s eyes. The Internet
began as a single global community with ad hoc regulation of the

195

Exploring the Internet

network for the common good. Over time, under the pressures of
massive growth and commercialization, the common view splin-
tered into many camps. Global connectivity is a key issue for the
Australian community. At the end of a long, skinny pipe, much of
their traffic is communication with other parts of the world.

There is no group that worries about global connectivity, or even
provides much in the way of a global regulatory framework and
places like Australia, on the frontier of the Internet, feel the impact
of this lawlessness first. The IAB had long been charged with look-
ing out for the Internet, but Geoff pointed out that the heavy Ameri-
can bias of the IAB doomed it, particularly in Europe. Many
Europeans continue to refer to the TCP/IP protocol suite as “DoD
IP’ to emphasize the American military origins of the Internet.

While the IAB had failed to provide a global framework for con-
nectivity, the engineering arm, the IETF, had not always provided
solutions to the problems that needed solving. Geoff emphasized
that the IETF was a “tremendous asset’ but also cited Marshall T.
Rose’s comment that “adult supervision” was needed.

The problem he saw was the lack of direction given to the IETF.
The IAB and the IETF Steering Group (IESG) failed at times to pro-
vide a solid agenda and timetables of engineering problems that
needed solutions. Problems such as the address space exhaustion
were studied in a fairly ad hoc fashion, even as the problems be-
came more and more pressing.

The IETF also suffered from an influx of goers and an increasing
tendency by many to treat the occasion as a social forum rather than
a working environment.

So what forum should guide the evolution of the Internet?
Geoff placed some hope in bodies like the Internet Society, but also
wanted to see an operational solution to the problem of global con-
nectivity.

To participate in the global telephone network, you have to
agree to certain sets of regulations in order to plug in. Geoff wanted
to see some baseline regulations defining how one cooperates in the
global network and, most importantly, a place to plug in to a global,
neutral backbone.

196

INTEROP Before Midnight

The Akihabara ‘With Me’ Girls

An ISDN Pay Phone

Motherboard Sale

Juha Heinanen in His Castle

Milo Medin and the Class C Airplane

Canberra

The NSFNET is an American backbone that also doubles as a
global backbone, provided you play by NSF rules. Geoff wanted to
see some higher tier, say some FDDI-like point of interconnection.
This ring would have no policies prohibiting certain classes of traf-
fic. If you pay your share of the ring infrastructure, and abide by
certain technical rules, you can join.

Once on the ring, it is up to each individual member to decide
whether or not to accept traffic from other ring members. This bi-
lateral decision would allow the NSFNET to decline to accept com-
mercial traffic without at the same time fragmenting the
connectivity of the entire Pacific region.

That night, over dinner and drinks at Geoff’s house, I brought
up the question of international bodies again. He had high hopes
for the Internet Society and I was curious whether he shared the
opinion of Bob Kummerfeld in Sydney that the OSI process was a
negative influence.

Geoff felt that OSI was a good thing. It had kept all the goers
busy, preventing them from turning their attention to forums where
real work was being done. Had OSI not existed, all these people
would have looked around for something to do and found things
like IETF meetings.

Under the Huston theory of standards development (and like
any good theory, it helped predict and explain many phenomena of
the standards world), one of the best things that could be done for
the Internet community would be to initiate a movement for OSI++,
keeping the standards potatoes fully occupied for another ten years.

197

Adelaide

Friday morning, I was up at an obscenely early hour to catch my
flight to Adelaide. It was the end of a hectic week in Australia and I
was off to meet Simon Hackett.

Simon took me to Internode Systems, his startup company of
three people with a suite of offices located behind an Adelaide re-
alty company. While Simon handled the usual run of emergencies
caused by being away for an hour, I examined a disemboweled Sun
2 in the corner, an ancient machine based on a Motorola 68010. The
huge 100 amp power supply had died, which was just as well be-
cause, as Simon pointed out, the thing drew enough power to drive
a welder. Instead, a power supply from a little PC had been put
into place and the Sun was up and running.

Simon is best known for putting funny things on the Internet.
At INTEROP 90, with the help of TGV, he showed off his Internet
CD player. The Internet CD player was a stereo system controllable
over the network. Connectivity to the net was provided by a
“magic box,” consisting of a Motorola 68000 CPU, some memory,
and a couple of RS-232 ports.

The box was connected to the INTEROP show network using a
SLIP link off one of the RS-232 ports. Portable TCP/IP software
from TGV and portable SNMP code from Epilogue Technology were
put into the magic box.

The stereo system was a Pioneer tuner/amplifier, a Pioneer PD-
M910 six-disc CD player, and two Klipsch speakers. The CD player
has a “remote control in” jack, allowing control of both the tuner
and the CD player. The output signal from the CD player has a mix
of digital audio samples and status information.

An SNMP MIB was defined and some X Window System soft-
ware was written to interact with the SNMP module on the magic
box. Of course, any SNMP network management station could have
been used, but the X software had a nice visual representation of the
stereo system. When the network management station changed a

198

Adelaide

MIB variable, that information was used by Simon’s code on the
magic box to control the CD player.

The problem with doing something cute like the Internet CD
player is that everybody expects you to top yourself the next year.
At INTEROP 91 Fall, Simon separated the speakers from the CD
player, using the Internet to distribute sound. The result was the
NetPhone.

The NetPhone is basically just a protocol for encoding sound (or
any other data) on the network. This protocol, the MultiMedia Data
Switch (MMDS), is built on top of UDP. MMDS takes a sample of
data, and puts it into a UDP packet along with information such as
the sampling rate and encoding format. The protocol can handle
most isochronous data such as video or audio.

The user input to the NetPhone is simply an audio board, such
as those that are commonly found on Sun, NeXT, PC, or Macintosh
computers. A telephone, speakers, or a microphone can all be con-
nected to these boards.

To provide connections between two NetPhones, a switchboard
was implemented on a VAX. A phone makes its presence known by
registering itself with this central switchboard. To place a call, a
user issues a request to the switchboard, which then contacts the
destination. If the two devices are available and have compatible
sampling rates, a connection is established.

If only two nodes want to talk, it doesn’t make sense to have the
switchboard in the middle. In that case, the switchboard issues re-
direct commands allowing the two nodes to send data packets di-
rectly to each other. In the case of multiport calls, data goes to the
switchboard which replicates the packets and resends them.

MMDS is thus more than just a phone protocol, it is a way of
moving sound and video around a wide-area network. At IN-
TEROP 91 Fall, there were NetPhones on various computers, allow-
ing people to talk to each other. People could place calls to the CD
player in the TGV booth and listen to music. There were also two
tuners, one in Melbourne and the other in Santa Cruz, allowing peo-
ple to listen to radio broadcasts from either of the two cities.

NetPhones are one paradigm of moving sound over the net-
work, and similar efforts are underway in several places, including

199

Exploring the Internet

Xerox and ISI, to develop ways to use this technology. Another ef-
fort to incorporate sound into the network is multimedia electronic
mail.

Were the two paradigms complementary or would one win? Si-
mon felt that not only did we need both message-based and real-
time voice transfer, the two technologies should be aligned, so a
voice message could be redirected to a NetPhone, or so that a Net-
Phone might act as an answering machine, recording messages and
then mailing them to the user.

Messaging is an important way to interact with people and cer-
tainly works on a one-to-one basis. However, the NetPhone also
supports broadcast technologies, as in the case of allowing people
all over the world to tune in to Melbourne radio and catch the
cricket matches.

Simon was continuing to work on the question of moving voice
around the network. He was looking at ways that the switchboard
could be taken out of the loop and various other enhancements to
the MMDS protocol.

His real interest, however, was the ability of the magic boxes to
put arbitrary devices on a network. In that vein, he was working
with Epilogue Technology to develop portable networking code that
could be put on small, cheap boxes. That code includes TCP/IP,
SNMP, and even higher level services such as NFS or the Ipr proto-
col for line printing. The code runs on a box as stand-alone soft-
ware, alleviating the overhead of a full operating system.

The possibilities for such a box are very intriguing. Think about
being able to put an NFS server on the network by packaging a few
disk drives into a small box. Or, put a tape drive in with a box and
use that as a remote backup device on the network. The number of
devices one could put on the network with a cheap interface are
endless. How about a machine room temperature monitor? Or a
toaster?

After we were done playing with the NetPhone and I had read
my mail, we went to Simon’s house where I met his partner in Inter-
node Systems, Robyn Hill. Robyn is also a naturalist and conducts
tours of the Great Barrier Reef, the outback, and other scenic attrac-
tions in Australia.

200

Adelaide

The next day, Robyn, Simon, and I went for a tour of the Barossa
Valley, one of Australia’s premier wine making regions. As we did
the Barossa crawl, I was able to sample wines ranging from the typi-
cal mixes of Cabernet Sauvignon and Shiraz to more unusual varie-
ties such as a fruity Alicante.

I was also able to replace the salami that was destroyed at cus-
toms with an appropriate offering of food for my friend in Bangkok.
Each vinyard sold a variety of condiments, and I stocked up on
pickled prunes and kiwi jams.

That night, we went to dinner at the Magic Flute, a posh restau-
rant in trendy North Adelaide. I started with a truffled boned quail,
stuffed with a rich pheasant pate and glazed with aspic. My main
course was an exquisitely prepared pigeon served over wild rice
and finished with a Madeira glaze. Simon had the peppered kanga-
roo filet in a port gravy, garnished with the omnipresent beets, and
Robyn chose rare mignons of venison.

Over dinner, we discussed the bizarre bribery trial of Joh Bjelke-
Petersen, for decades the Premier of Queensland. It was alleged
that Bjelke-Petersen wanted money from Alan Bond, one of the
more prominent rat-pack Aussie businessmen. Instead of just hand-
ing Over money in a brown paper bag, however, it is alleged that the
bribe had an ingen ious twist. Allegedly as a prearranged pretext,
Bond's national television network publicly insulted the prime min-
ister. Bjelke-Petersen turned around and sued for libel and the case
was settled out of court for cash.

While our dinner was good, dessert was absolutely spectacular.
Robyn insisted we all get different things so she could taste them
all. She ordered a gratin of raspberries and blueberries, cooked in a
lemon custard and blowtorched on top to caramelize the sugar, then
finished with an orange butter sauce. Simon had the hazelnut cake
on a coulis of fresh coffee beans, and I had an ice cream confection
of raspberry, mango, and passionfruit, separated by thin layers of
chocolate licorice ice cream. Many fine lunches and dinners, indeed.

We headed back and I slept in their spare room, along with
three very fat tree frogs that Robyn keeps, and a mountain of old
disk drives that Simon keeps.

201

Singapore

Monday morning, I headed down Singapore’s spotless streets and,
for Asia, an eerie lack of chaos, up to the National Computer Board
to meet with Lew Yii Der, an engineer with Singapore’s Public
Works Department. On my last visit, I was told that a contract had
been awarded for an electronic road pricing system, and I was eager
to learn more details.

Since 1975, Singapore has attempted to reduce congestion in the
central business district by charging a fee to enter the area during
morning and evening rush hours. There are 26 entry points to the
central area, known as the Restricted Zone.

Before entering the Restricted Zone, a motorist pays SG $3 (U.S.
$1.76) for a license and, when passing the entry point, ensures that
the large pass is prominently visible on the left side of the wind-
shield. Failure to prominently display results in your license num-
ber being noted by the police patiently waiting by the side of the
road and a ticket being mailed to your home.

On a typical day, 300,000 vehicles enter the Restricted Zone, of
Which 74,000 enter during the controlled periods. While the system
has certainly reduced congestion by keeping those without the will
or the resources away during rush hour, it has some drawbacks.

The biggest problem is the manual nature of the system. Cars
must stop to buy a license and then slow down to wave it around.
Fifty-two officers must be present at the 26 entry points to enforce
the system.

Many cities throughout the world are investigating ways to re-
duce congestion in central business districts. Lew Yii Der explained
the two classes of solutions: passive and active. In a passive system,
people maintain accounts with the government. When you enter a
restricted area, a charge is made to a person’s account. Passive sys-
tems thus require central billing, are difficult to maintain, and, even
for Singapore, are considered too intrusive. An active system uses a
stored value card. Buying a card for a subway and then running it

202

Singapore

through a reader to activate the turnstile is an example of an active
system.

While I had understood, perhaps wrongly, on my previous visit
that a contract for an active system had already been awarded, a bit
of further discussion turned up the information that not only had
the contract not been awarded, but the RFP was not yet out. An
RFP had been issued previously, but it was worded so vaguely that
vendors were not sure what to bid.

Lew Yii Der gave me an outline of what the new RFP would
specify. Smart cards would be mounted on all car windshields.
Value for the cards would be added at sales outlets, where the mo-
torist could also review the transaction log stored inside the card.

When a motorist entered the Restricted Zone, the car would pass
under a microwave beam, which would debit the appropriate
amount from the card. If the card didn’t have enough money on it,
a camera would be activated to take a picture of the rear license
plate.

Since the system didn’t exist, it was a bit hard to get more tech-
nical details, so I spent the rest of my allotted hour trying to learn
more about other traffic control systems.

Like most big cities, Singapore has computerized its traffic lights
in the busy parts of the city. The original system, installed in 1983,
was a fixed system, meaning that the lights changed on some pre-
arranged timetable. Each light had a phone line to a PDP computer
which controlled the lights and attempted to form “green waves” so
motorists could drive without stopping.

In 1987, Singapore installed an Australian dynamic traffic man-
agement system called SCATS. In addition to the traffic light, each
intersection in this system has magnetic detector loops under the
pavement, used to measure the flow of cars. A local controller at an
intersection can use the absence or presence of cars to change lights.
The data from the detector goes to the regional PDP computers,
which use the traffic data to adjust light patterns while trying to
establish green waves. Data from regional computers goes to a VAX
11/750, which is used by headquarters staff to monitor traffic status.

The traffic subsystem is also tied into the emergency fire system.
Every major building in Singapore has a fire detection system tied

203

Exploring the Internet

into the fire stations. When there is a fire, trucks are automatically
dispatched from fire stations to the fire location and the traffic sys-
tem is used to provide a green wave for the trucks.

Between the parking control, the Restricted Zone, the traffic light
management, the incredibly clean streets, and the gum-free subway
system, it was certainly clear that Singapore was able to control the
flow of people and vehicles in the city. I had to ask myself, how-
ever, if there wasn’t some tradeoff between efficiency on the one
hand and, on the other, an atmosphere that promotes a vibrant,
creative, and (yes) fun place to live.

On my previous visit, I had learned about Tradenet, an EDI applica-
tion for clearing customs paperwork. I had heard that not only was
the customs work computerized, but that the Port of Singapore
Authority (PSA) also had extensive systems. As with all things
computer related, the National Computer Board was the place to
start.

Singapore is one of the world’s busiest ports. Since the opening
of the Suez Canal in 1869, the island has been a major entrepdt port,
serving as a clearinghouse between destinations in Europe and Asia.

That afternoon, Chew Keng Wah of NCB brought me down to
the Tanjong Pagar terminal where container ships are loaded and
unloaded. We met with Ang Chong Hoat, an MIS manager for the
port.

Running a terminal as large as Tanjong Pagar is a highly com-
plex scheduling problem. When a ship arrives, containers have to
be unloaded and other containers loaded. Cranes mounted on rails
are moved up and down the length of the wharf and are used to
move containers on and off ships.

Incoming containers are then placed on trucks and run over to
an adjacent storage yard. There, containers are picked up by an
even bigger crane, a transtainer, a device that constructs huge piles.
When the ship arrives that is to carry the container to its destina-
tion, the container makes the reverse journey back onto a ship.

204

Singapore

The trick is to pile the containers so they are readily available
when needed. You hope that the container you need at a given time
is not buried under a pile of other containers that are not needed
until later. It also makes sense to keep containers for similar desti-
nations in the same area so the transtainer does not have to be
moved around.

The goal of all this is to minimize the amount of time a ship
stays berthed at the terminal, thus making more effective use of
scarce resources. This minimization is done in an environment of
scarcity, where cranes, ships, transtainers, and berths are all finite
resources.

Planning the loading and unloading of ships and the subsequent
placement of containers in the yard was once the province of a few
highly skilled experts churning out plans by hand. A plan for a
single ship could take many hours to work out. To automate the
process, the Port Authority and the NCB developed an expert sys-
tem to assist in the process, cutting the amount of time to develop a
plan for a ship in half. The expert system was developed on Sun
Workstations in Objective C and Lisp and took over 20 person-
years.

The ship planning system is the first of six that the PSA will
eventually deploy. Other systems will try to place ships in berths
adjacent to the correct yards, or try to take ships in a yard and place
them in a correct berth, or make available the proper number of
resources for the unloading and loading of a ship.

The planning system thus tries to solve a complex optimization
problem a level at a time. The results from one planning exercise,
such as putting ships in the correct yards, is moved into a Sybase
database. The next planning process takes that data out of Sybase
and uses it as input for the next level of decision making.

Since a yard has many simultaneous ships, it is still possible for
one ship-unloading plan to interfere with another ship-unloading
plan. The last planning module is a terminal operations simulator,
used to take all the separately developed plans and see if they con-
flict. If so, adjustments are made back up the chain.

The planning network at Tanjong Pagar terminal consists of two
mirrored Sun servers with approximately 5 Gbytes of disk space.

205

Exploring the Internet

Planners use 13 workstations and various plotters and printers.
Some of the modules, such as the ship planning system, were al-
ready deployed. Others were just going online.

Planning is one of three major clusters of computer applications
in use at the terminal. A large mainframe is used for documentation
and a third cluster of applications runs on a Stratus computer for
fault-tolerant control of yard operations.

The documentation system contains the raw data, such as when
a ship is expected to arrive and the contents of that ship. This infor-
mation is used to feed the planning process and also is used as in-
put for back-office functions such as billing.

Data for the documentation system is fed in electronically by
shipping agents. Small companies emulate a 3270 terminal and dial
in to fill out forms. Larger organizations use an APPC-based pro-
gram to transfer data. Documentation data resides in an Ad-
abase/ Natural application and an APPC-based program running on
a SunLink gateway transfers data over to the Sybase server for plan-
ning.

The output from the planning process is used to feed the yard
operations system. This system, which was preparing to go online
when I visited, run on a Stratus platform. The Stratus is linked to
the IBM using SNA APPC programs and to the Sun workstations
using TCP/IP.

The output of the planning process is a series of loading and
unloading sequences, specifying container ID numbers, yard loca-
tions, ship berth locations, and other identifying data. This informa-
tion is translated into a series of micro instructions, such as a
command to a transtainer to go to a certain location and expect to
load a container with a certain label. Another command might give
a truck instructions to move to a certain berth.

All this is computerized, with custom PC-based systems on
cranes and transtainers. PSA trucks get transponders to track their
location and simple display terminals give truckers instructions.
Trucks entering the terminal from local shipping companies are
given a window during which they are allowed to enter the yard,
and at the gate the truck is fitted with a transponder.

206

Singapore

As we were driving back from the port, Chew Keng Wah told
me that yard operations were only half of the port automation pro-
gram. The other half was a marine-based control system that would
automate the deployment of tugs, pilot allocation, and utilization of
deep sea channels in and out of the terminals.

Having finished my whirlwind tour of Singapore, I went back to
my hotel to meet a friend for a drink. Bob Cook (not his real name)
is a salesman for a large computer company. Salesmen for this com-
pany are assigned territories by country. One will work in Thai-
land, another in Singapore. Bob gets whatever is left. His beat
includes places like Brunei and Bangladesh.

I had brought Bob a copy of Stacker for his laptop PC. In many
places, including Singapore, the only way to obtain legitimate soft-
ware locally is to pay three times (or more) the U.S. list price. Even
then, many U.S. companies refuse to provide technical support or
updates.

Lack of legitimate software drives many otherwise honest users
into the arms of the software pirates. Bob always gave me a list of
software he needed to purchase and I brought it over for him on my
trips. Updates and other luxuries he had to get for himself from the
pirates.

After taking his software, Bob looked around the bar furtively.
Everybody was watching the lounge lizards doing an Elton John
tune. The audience was captivated, although the fact that the per-
formers were clearly out of the Barry Manilow school just made me
thirsty.

The coast clear, Bob told me he had something for me, giving
me a conspiratorial wink and pulling out a package of Big Red
chewing gum. Singapore had just banned all chewing gum on the
island, a reaction to the fact that somebody had discarded their gum
in a couple of subway doors, jamming them. The fine for importing
this vile substance with intent to deal had been set at SG $10,000
and a year in jail. The government had just arrested four gum deal-
ers and was threatening swift retaliatory action to any others fool
enough to chew and talk about it. A government official was quick
to point out, however, that only the importation and sale of the sub-
stance was banned, not its chewing.

207

Exploring the Internet

Bob’s frequent travels to free trade meccas like Bangladesh had
given him plenty of opportunities, and he had become a key gum
trafficker, supplying his teenage daughter and even occasionally of-
fering a stick to friends.

208

Kuala Lumpur

After an interminable wait for my luggage in front of a sign reading
“baggage claim,” which had spent the time circling a belt in an adja-
cent unlabeled room, I cleared customs. Nobody was around as I
emerged from customs at some obscure side door. Seeing nobody
around an exit door in an Asian airport gives one an extremely eerie
feeling, as these places usually pack several hundred expectant
families into a mob around the exit, making the walk to the taxi
stand challenging.

I looked around and saw that the mob was clustered around the
other exit. I worked my way through the crowd into the middle,
then turned around and walked back out. Sure enough, I soon saw
a sign with my name on it and introduced myself to Rafee Yusoff
from the Malaysian Institute of Microelectronic Systems (MIMOS).

Rafee spoke perfect English, not surprising since he went to
school in Iowa. We drove past palms and the heavy vegetation that
grows in this steaming jungle climate. The traffic was a far cry from
Singapore. Cars here easily turned three lanes into 4 or 5 as they
darted in and out in increasingly complex geometric progressions.

Malaysia is the world’s top exporter of integrated circuits, the
result of an aggressive government policy of free trade zones, cheap
labor, and a strong work force. Malaysia had hoped that the huge
manufacturing operations of companies like Motorola, Texas Instru-
ments, and Sony would lead to technology transfer and increased
value added by the Malaysian work force.

This didn’t happen. The Malaysian computer industry contin-
ues to be devoted primarily to chip manufacture and assembly of
components. MIMOS was created to help increase the Malaysian
R&D effort and thus increase the technology transfer into the local
community.

Rafee deposited me in a waiting area and soon ushered me into
a large conference room with a very, very large conference table,
capable of seating groups of 40 or 50 people at a time.

209

Exploring the Internet

Two women in chadors came in, followed by Chandron Elam-
vazuthi, a bearded UNIX guru. Finally, the head of the MIMOS
computer systems division, Dr. Mohamed B. Awang-Lah came in
with a very large pile of transparencies.

The purpose of my visit was a bit unclear, an understandable
confusion since I was researching the Internet and Malaysia was not
on said Internet. Dr. Mohamed gave a nice little speech, welcoming
me to Malaysia and MIMOS for whatever reason it was that I had
come. I then gave a nice little speech thanking MIMOS for their
hospitality and thanking them in advance for whatever program
they might have set up.

After the introductory formalities, Dr. Mohamed started pulling
out transparencies and sliding them across the table so I could see
them. Soon a large stack had been fanned out in front of me, ex-
plaining the purposes behind MIMOS and the various projects of
the computer systems division.

One project that MIMOS had worked on was to help the Minis-
try of Education design the specifications for a PC to use in the
school. Why not just buy some system that already existed? Chan-
dran excused himself and ran pao to get a prototype of the
Atom-1 computer for me to see.

The entire PC had been built into a wedge which fit under the
keyboard, and included a 3.5-inch floppy drive, serial and printer
ports, and even a network interface. When you set the computer on
the table, the keyboard tilted up a few inches and used much less
real estate than any normal desktop. You still needed room, though,
for a display on the desk (or hanging someplace nearby).

After developing the prototypes, the Ministry of Education had
then issued a Request for Proposal (RFP) and a private company
took over the manufacture and delivered units for a fairly low cost.
Sixty schools were in the process of putting in 20 PCs each and the
Ministry had plans to computerize all 1,300 secondary schools.

At this point, we paused while cups of Nescafé were poured for
all of us, accompanied by sticky green sweets and some spicy meat
pastries. While sipping my coffee, I looked around the room and
noticed that everybody had notebooks and everyone seemed to be
taking lots of notes.

210

Kuala Lumpur

Between bites of pastries, we examined some more transparen-
cies describing JARING, the proposed national research network for
Malaysia. JARING means net in Malaysian, but I also learned that
in English it had a much longer meaning, spelled out in a little
poem at the beginning of a brochure I was given:

Joint Development IT Infra-structure Built Upon

Advanced Technology; Supports Multidisciplinary Collaborative
Research, Development, and Educational Activities; Reflects
Integrated Development Strategy Through Computer
Networking; and has Worldwide Connectivity for

Global Communication and Information Exchange.

Wow! I wasn’t quite able to discern from the poem, however,
what the network might look like. There were plans to link all the
major universities and cities, and JARING was waiting until
Telekom Malaysia would be able to deliver a 64 kbps line, a process
that would take some time.

In Kuala Lumpur, several institutions were connected together
using X.25 with UUCP and X.28/X.29 protocols. An international
X.25 connection provided the link to UUnet. TCP/IP links within
Kuala Lumpur were running, but I was unsure of exactly where and
what they were used for.

MIMOS was planning on making the network some mix of X.25
and TCP/IP. Apparently, they needed a solution that would allow
some fairly ancient IBM computers in government ministries to par-
ticipate in the network without requiring those ministries to spend
any money, and straight X.25 solutions were old enough to work on
all of these machines.

We broke for a fine lunch of curried chicken, shrimp sambal
with chilies, and a delicious omelette stuffed with sweet and spicy
vegetables. With some time to kill before the afternoon meeting MI-
MOS had set up, I was left in the MIMOS library. Perusing the
shelves, I spotted a copy of the Internet Managers Guide right next
to a copy of the Concise Encyclopedia of Islam.

211

Exploring the Internet

My MIMOS experience was a bit different from my usual site visit.
Normally, I asked people for help in a new country and got back a
list of e-mail addresses, which I used to set up a series of appoint-
ments. Some appointments were time definite, others were vague
promises of intent.

Rather than giving me pointers, MIMOS had set up my entire
itinerary. In fact, they would be accompanying me to all meetings.
Dr. Mohamed made it clear that the MIMOS name had been ad-
vanced on my behalf and it was understood that I should behave
accordingly. Rafee continually probed to find out what the nature
of my writing would be, perhaps hoping to avert a potentially em-
barrassing piece.

That afternoon, Chandran and one of the women in chadors, ac-
companied by a MIMOS driver, were my escorts to go visit the
headquarters of Plus, a private toll road that would span the Malay
peninsula from the northern border with Thailand down to Singa-
pore.

I had heard that Malaysia was laying fiber optic cable when they
built roads and it turned out that the rumors I had been hearing
about for two years had to do with the Plus project. While the toll
road was being installed, the corporation had planted a bundle of 36
fiber cores, each capable of operating at anywhere from 34 Mbps up
to 140 Mbps. The business case for the fiber had been made for
internal operation of the toll road, but it was evident that quite a bit

of spare capacity had been installed.

I met with a team of five Plus MIS employees, headed by Rosli
Md Tan. They joined our MIMOS delegation to form a fairly large
party sitting around a large conference table.

The MIS staff estimated that they would take roughly 8 Mbps of
the fiber capacity for their internal network. They would also use
fiber for the toll plaza voice network. The 8 Mbps data bandwidth
was being set up as a large extended Ethernet spanning the entire
Malay peninsula.

The lowest unit of operation in a toll road is the lane in a toll
plaza. Each lane in Plus would get a lane controller unit, a custom-
ized PC that connected to the cash register, the light on top of the

212

Kuala Lumpur

booth to signal that the lane is open, and various other toll collec-
tion accoutrements.

Up to 17 lanes in a plaza would then be connected by 9,600 bps
serial lines into a PC that would act as a data collection unit. An
18th virtual lane controller would be available inside the toll plaza
headquarters to be used for collection of cash for things like
monthly passes.

The PC concentrator would in turn connect to a pVAX, which
would act as the central computing resource for the toll plaza. The
uVAX would also support management terminals as well as a point-
of-sale terminal.

The toll plaza pVAX would use the fiber infrastructure to peri-
odically upload data to regional headquarters, which are equipped
with a VAX Cluster.

Inbetween the region and the toll plaza is an intermediate unit
of management, the section. Section headquarters don’t have
VAXen and must therefore go to regional headquarters for their
data. Each section would have a set of terminals connected into the
fiber patch panel, just as a voice phone would connect. At regional,
a terminal server would take the incoming lines, thus connecting the
terminal to the VAX cluster using the DEC LAT protocols.

One interesting aspect of the toll road is that it must break in
Kuala Lumpur, since Plus does not have permission to cross the city.
To connect the northern and southern segments, Plus had leased an
8 Mbps line from Telekom Malaysia. One more 8 Mbps leased seg-
ment linked the northern segment to a VAX 6100 located at Plus
headquarters, a building located, for some reason I could not under-
stand, away from the toll road.

Plus hoped to have the first segment, going 100 kilometers
northward from Kuala Lumpur, up by August 1992. The entire
road, and thus the management network, would be completed by
1994.

How would the spare bandwidth be allocated? Plus was owned
by a holding company called the Renang Group, which also owned
an electromechanical consulting organization known as Time Engi-
neering.

213

Exploring the Internet

Time Engineering had applied for a license and was about to
become Malaysia’s second telephone company, providing at least
long-haul data communications and possibly other services such as
voice. The fact that Time Engineering had applied for a license was
somewhat of a formality as the Renang Group was owned by Ma-
laysia’s ruling party, a group that had been in power since 1957. As
one embassy official sardonically put it, approval seemed “likely.”

The next morning, I passed up the opportunity provided by my ho-
tel for a “western breakfast buffet’ featuring such delicacies as beef
bacon, stale toasted white bread, and a tray of what was labeled
“exotic tropical fruits” but contained instead grapes and honeydew
melon.

Maybe honeydew was exotic in Malaysia, but I opted instead for
a dish one-third the price of the buffet (but still an order of magni-
tude more than I would have paid on the street) of teocheow, a rice
porridge accompanied by small portions of chicken sambal, dried
anchovies mixed with peanuts, a pungent fish with black beans, and
half a salted egg.

Later, Rafee picked me up to bring me to Telekom Malaysia, the
recently privatized government PTT. I was interested to see the
status of their fiber infrastructure given the bold moves by Plus and
Time Engineering.

I noticed that everywhere we went, people always spoke to
Rafee in English, even though I stood in the background. Even
parking lot attendants would break into English, asking Rafee how
long he intended to stay, as if he had turned into a foreigner by
being with me.

We reported to the 21st floor of the Wisma building only to find
that the head of the Network Technical Services Division, whom we
were supposed to meet, was not in that day. Suddenly, the conver-
sation switched into rapid Malay. After a few minutes, a deputy
division head was produced and we were ushered into the division
director’s office to meet with Ahmad Tarmidi.

214

Kuala Lumpur

Rafee made a nice speech thanking Mr. Tarmidi for seeing us
and introducing MIMOS. I introduced myself as a reporter for Com-
munications Week and explained that I was interested in the status of
the fiber infrastructure.

Ahmad Tarmidi then made a short, indefinite speech explaining
that fiber played an important role in Malaysia’s future, as did tech-
nologies like ISDN. The PIT was pulling fiber down the major
north-south roads (with the obvious exception of Plus) and hoped
to have coverage of the Peninsula by 1995.

As to the specific services or the specific infrastructure, the con-
versation was more than a bit vague. Yet, on the way outside, Mr.
Tarmadi called Rafee back for a hurried conversation. Rafee ex-
plained that the fact that I was a reporter had hit home and Tarmadi
was worried I would give away sensitive business plans. I assured
everybody that no strategic business advantage would be lost as a
result of our little chat.

After a nice lunch in the MIMOS cafeteria, where I decided to skip
the fish head curry in favor of less exotic fare, we headed off to visit
Tenaga Nasional Berhad, the national electric company and another
group that was laying a fiber infrastructure.

At Tenaga, a group of only four people met us, still outnumber-
ing our own delegation of three. The assistant general manager for
information systems, Abdul Rahman Bin Shafi, welcomed us and
gave us all, including his own staff, a document describing the Dis-
tributed Source Data Generation Project.

I flipped through the four-page document, scanning it for infor-
mation. Abdul called the meeting to order and asked us all to turn
to page 1, where he commenced to read the entire text aloud.

Tenaga was proud to have what it called the largest distributed
systems project in ASEAN. Implemented over a 36-month period at
a cost of 18 million ringgit (U.S. $7.2 million), the project had con-
nected 135 district offices and power stations of the electric utility
system.

215

Exploring the Internet

Processing oomph for the system was provided by a pair of
fault-tolerant IBM 4300s. Distributed processing was the realm of
Nixdorf computers, the prime contractor on the project. Telekom
Malaysia’s X.25 network, Maypac, was used to connect district of-
fices to headquarters at speeds ranging from 2,400 to 4,800 bps and
occasionally at 9,600 bps.

Applications consisted of various ways to input paper records,
such as allowing the fact that a bill was paid and money collected to
be transmitted to headquarters. Previously, the data had been input
at headquarters, meaning that errors in entry codes and the like had
not been caught until the data was well downstream.

After the briefing paper was read, a huge TV was wheeled in.
While it was being set up, Rafee made polite chitchat about the re-
cent renovation of the MIS office space. We all commented on how
well-lit the offices were, certainly an appropriate attribute for the
offices of a monopoly electric utility.

The video was then displayed. Ten minutes and forty-eight sec-
onds long, the in-house video explained the success of ASEAN’s
largest distributed project, accompanied throughout by a driving
disco beat. Near the end, the disco turned into an anthem (though
still accompanied by the disco drums), the hyperbole flew thick and
fast and then finally the lights went back up to monopoly bright-
ness.

I looked around the table and saw that everybody in the meet-
ing had a briefing packet, consisting of my resume, a Xerox of the
cover of STACKS, a letter from MIMOS, and an internal Tenaga
memo with a very long routing slip on it.

Next on the agenda was a description of a home-grown commu-
nications system, used to control the power substations. Substations
could be located in very remote areas, so it was not possible to
count on the telephone company to provide even voice service.

The system consisted of twisted pair lines running along with
the power lines and providing data transmission at 100 and 600 bps.
One function of this communications system was to allow central
headquarters to activate an emergency cutoff switch at a substation
to prevent damage to the larger network.

216

Kuala Lumpur

Tenaga was in the process of replacing the control system with a
fiber optic network running along the power lines. The fiber would
support cascaded, automated remote power substations and would
even allow transmission of video images for monitoring and secu-
rity. Tenaga was laying cables of 10 fibers, giving it a huge capacity
for future expansion (or sale to a telephone company).

We finished our meeting just before 4:00 PM., allowing Rafee to
drop me at a taxi stand a few minutes after four. Government em-
ployees are released at precisely 4:15, and a 20-minute ride immedi-
ately turns into a 90-minute ordeal. I caught my taxi at 4:08, hitting
the airport just ahead of the wave.

Kuala Lumpur, like many airports in developing countries, has
an acute shortage of terminal space, meaning that you cannot check
in until two hours before a flight. It was a hot, humid jungle after-
noon and I had several hours to kill before I could check in.

I found a cart with a broken wheel and piled my garment bag
and computer case into it and took off my jacket and rolled up my
sleeves. I jumped my cart over the high curb, played chicken with a
taxi (he won), and went across the street to find a bar.

The airport hotel had a choice of stairs and an elevator. The
stairs had a big picture of a durian, with the circle and the slash
around it spelling out in terms suitable for international tourists that
this hotel was a “No Durian” area. The elevator, used by western-
ers to get into the main lobby, had no need for such a sign. The
picture of the durian reminded me that I was soon going to be in
Bangkok, where I would be able to get my fill of this strange, odor-
iferous fruit.

After a snack in the restauraunt of baby octopi on a bed of
shredded jellyfish, I went to find the bar. Sitting in the small
crowded bar, I watched a TV special on the wonders of Malaysian
industry, focusing in this episode on the life of a pineapple.

At 7:23, the screen went blank and was replaced with a picture
of a big clock. How handy, I thought, a time service that randomly
appears. The clock was soon replaced by pictures of Mecca and the
sound of the call to prayer. Arabic and Malay subtitles started fill-
ing the screen and rolling off, accompanied by an upbeat, yet inspi-
rational soundtrack. Around me, the bartenders continued

217

Exploring the Internet

preparing drinks. Prayers finished, the screen did another context
switch back to the marvelous Malay miracle, now in the middle of a
scene of pineapples getting lobotomies, interspersed with shots that
featured happy foreign tourists on idyllic beaches eating pineapple.

I left to stand in line and check in, cleared immigration, and
walked past the endless row of duty-free shops, including one sell-
ing bags of dried mangoes and prunes, items that I couldn’t recall
falling under the duty-free exemption in too many countries.

With great relief, I found the doors of the CIP lounge, a cool
plush oasis for Commercially Important Persons. Making a living
as a writer is kind of the antithesis of the CIP, and I take great de-
light in making myself at home in such facilities whenever I can fool
the authorities into letting me in.

218

Bangkok

As you travel up the peninsula from Singapore, things get more
chaotic, more exotic, more Asian. For sheer density of activity,
Hong Kong represents the peak, with things getting a bit more or-
derly as you head north towards Korea and Japan. Orderly is, of
course, relative by Asian standards; the chaos level of any of these
places is several orders of magnitude over any U.S. city.

For mystery, though, you find your peak load in Bangkok. This
was my tenth visit to Bangkok and I still had only a vague inkling
of how the place worked. The chaos is there, but unlike Hong
Kong, you can’t get a handle on it. Things are happening, but you
don’t know what, and nobody stops to tell you.

With the temperature a balmy 85° F and the Thais all wearing
sweaters to ward off the cold, I set out on foot down Ploenchit Road
to find the stock exchange. Dozens of large construction projects
were all packed in together, part of Bangkok’s expansion as the only
real city in a country with the fastest growing economy in the
world. For years, the growth in GNP had been averaging over 7
percent per year and from 1987 to 1990, GNP grew at an amazing 11
percent per year.

Each construction site was packed with a hundred or more la-
borers from the Isan northeast of Thailand, near the Laotian and
Cambodian borders. Teenage boys and girls worked alongside a
few older workers in their 20s or 30s. They would stay at the same
site for weeks or even months, living in the building as it shot up.
A dozen or so food stalls clustered around each site, with old ladies
and young girls making Isan food such as the fiery som tam, a pa-
paya and chili salad made with lime juice and fish sauce (and, in the
more authentic versions of the dish, served with crickets or land
crabs for that extra crunch). Other stalls had dried fish, sticky rice,
or grilled chicken, with construction workers ambling from one to
the other, poking at the food and trying to decide what to eat.

The morning rush hour was in full swing and I waited at the
intersection with Wittayu Road, watching the police in their traffic

219

Exploring the Internet

booth act as DJs, interspersing jaunty music with inspirational
speeches intended to motivate motorists and pedestrians to conduct
their business with dispatch. Of course, everybody went their own
way. This particular section of Ploenchit road had once been desig-
nated by the police chief as a “traffic law observance zone,” a title
that did more to underscore the status of laws as suggestions than it
did to convince anybody to change their driving habits.

Wittayu road is a wide, shaded boulevard, lined with embassies
and a few fancy hotels. I walked past the Imperial Hotel, site of the
annual Fourth of July festival that offers free vasectomies and hot
dogs. This event is sponsored by a Thai non-profit family planning
and rural development group I had worked with.

My appointment was at the Stock Exchange of Thailand (SET)
with Dr. Surat Palalikit, a senior vice-president in charge of the MIS
operation. I presented my business card to the security guard, an-
nouncing myself in Thai. He nodded and picked up the phone.

“A farang is here for Dr. Surat,” he said. Except for a few close
friends, Thais always refer to me as “farang,” a foreigner, rather
than by name. I occasionally bridled at being called “Mr. Foreign-
er,” but realized that my name was virtually unpronounceable in
Thai. The closest people usually got was “Mah-Lah-Moo” which
translates into Thai as “come hunt the pig.”

Dr. Surat was an energetic man, full of charm and speaking per-
fect English. Trained as a molecular chemist, he had been a profes-
sor at both Mahidon and Chulalongkorn Universities, two of the top
Thai schools.

A lack of computers made it challenging to do much research in
molecular chemistry, so Dr. Surat decided to moonlight. A friend of
his ran Chase Manhattan’s computers in Thailand and asked Dr.
Surat to supervise the creation of a general ledger system.

The system went in and worked. His friend was promoted to
Hong Kong and Dr. Surat was given the job as EDP manager. After
a couple of years doing that, he supervised the creation of an ATM
network linking seven large banks. It was evident that, by now, mo-
lecular chemistry was a hobby and he had found his calling in the

financial markets.

220

Bangkok

At that time, the Stock Exchange of Thailand was running a few
PCs on a local area network. A committee was formed to go up to
Taiwan and investigate their EDP system and Dr. Surat was invited
along. The Taiwanese exchange was working well and it was evi-
dent that Thailand needed to engage in some form of expansion to
support a rapidly growing stock market.

The choice was whether to make or buy a system. Dr. Surat
presented the case for buying as much as possible, so strongly that
they made him the EDP manager. That’s when Dr. Surat met Mary
Jo Moccia, the person who had built the Midwest Stock Exchange
system from the ground up and was porting her solution to stock
exchanges all over the world.

I had met Mary Jo on my trip to Thailand the previous year. We
had sat next to each other for six hours on a plane and spent the
time talking about everything from DECnet bugs to the differences
in CPU architectures and their implication for efficient database en-
gines to the long-term financial viability of DEC.

Gruff, chain smoking, highly technical, everybody who worked
with Mary Jo described her as one of the best MIS managers they
had ever met. We had both been staying at the Regent Hotel, and
all week I kept bumping into her. My consulting was done at the
end of the week, and we had made tentative plans to get together
Saturday morning for breakfast.

Saturday morning, I picked up the phone and asked the opera-
tor to connect me to Mary Jo’s room.

“I'm sorry, but she has passed away,” was the reply. Thinking
this was some kind of bizarre translation problem, I went down-
stairs to investigate.

While I was trying to get a manager at the desk, one of her staff
passed by and overheard my conversation. I found out that Mary
Jo had gone to bed after her weekly get-together with her staff and
died during the night.

Four very young, very shocked staff members stood around not
knowing what to do. A U.S. embassy official took care of the details
and got them booked on the next flight home, coincidentally the
same flight I took.

221

Exploring the Internet

I tried briefly to cheer them up, but there is really nothing you
can say or do that will help in such a situation. I wrote an obituary
for Computerworld, but the gesture felt empty and meaningless in the
face of such a sudden and tragic loss.

Dr. Surat and I talked about Mary Jo Moccia and it was evident
that in the year that had passed since her death the project had got-
ten back on track, but the loss still hurt.

SET had resisted the temptation to run their operation on a large
IBM or a fault-tolerant platform like Stratus or Tandem. Instead,
their system was based on distributed minicomputers and worksta-
tions, centered around a fairly small VAX Cluster.

The back-end matching system, which takes care of the mechan-
ics of buying and selling, is a cluster of two VAX 6510s. Three VAX
3800 systems provide front-end processing to the network of 40 bro-
kers. Each broker has a choice of an IBM AS 400, a Stratus, or DEC
equipment.

All the SET VAXen share an Ethernet segment. The brokers con-
nect to the front end VAXen using simple leased lines. Since no
modems are involved (and leased lines are hard to hack), security is
fairly straightforward.

In case the computer-to-computer links between the front-end
systems and the brokers fail, each broker also has a terminal con-
necting to a terminal server on the SET Ethernet. The terminals
function as a backup in case the fancier front-end system goes
down.

The day before I visited, the SET had reached an all-time high
volume of 8.585 billion baht (U.S. $343 million). The day I was
there, as the Bangkok Post screamed out the next day, “the bull
turned bionic,” reaching a total volume of 12.341 billion baht. The
business section of the Post also went bionic, raving on about the
“Super Bull” and the “Galloping Bull,” and generally carrying this
metaphor so far that I felt like reaching for a shovel.

Despite the metaphors, things were pretty calm at the exchange
offices. People drifted in and out of Dr. Surat’s office, sat and lis-
tened to our conversation for a while, and occasionally he would
step out and talk with them for a few minutes.

222

Bangkok

The system seemed to work fine. One broker switched to the
terminal server backup and a meeting was called with the MIS staff
and the brokers just to make sure things were under control. Even
the broker’s offices, in a frenzy by Thai standards, seemed fairly
calm to me.

While the bull was going bionic, Dr. Surat and I sipped our cof-
fees and chatted about the other two pieces of his MIS operation.
One was the price reporting system, designed to feed out status in-
formation to the ultra-modern, wall-sized displays on the walls of
most brokerages, as well as VT terminals and PCs all over town.

The other big component was called the SET Information Man-
agement System. Under development, this system consisted of an
IBM RISC processor running AIX and Ingres. Brokers and compa-
nies would use this system to enter and retrieve information such as
financial statements, profiles, and other fairly static information
about stocks traded on the exchange.

I asked how he planned to support generalized dial-in in an en-
vironment where security was an issue. We tossed around the idea
of a PC with modems, isolated from the net. The PC would accept
uploads from people. At the end of the day, simple virus checks
and authentication would be run. Then, the modems would be
turned off and a switch flipped to connect the PC to the AIX system.
The semi-validated data would be loaded into the Ingres database,
presumably with more stringent data checking.

Letting the public retrieve information without giving them ac-
cess to the Ingres engine could be done in a similar fashion, caching
standard reports to a PC and then disconnecting it from the net.
The database engine could still be available to the brokers over the
network, allowing them to retrieve more sophisticated reports.

Like a good consultant, I threw this scheme on the table, then
left before the dangers of reality exposed any flaws. Having
worked my two hours for the day, I headed up Sarasin road, passed
a pushcart selling crab mousse in banana leaf, and stopped in front
of the black chicken joint.

This shop specializes in chickens that have been bred to be black
all the way through to the bone. The chicken is steeped in medici-
nal herbs and is considered to be good for all sorts of things like

223

Exploring the Internet

prosperity and virility (although I must admit, it tastes quite similar
to your basic New England boiled chicken). I once brought a farang
acquaintance in there and, needing one more dish to round out the
meal, ordered “chicken leg in garlic sauce.” We were presented
with a platter of steamed chicken feet in a sticky gel of garlic and
agar.

I turned down Ratjadamri road and walked past the Lumpini
Park, one of the few large parks in a city of 8 million people, and
headed towards the Bangkok Post. I wrote myself into the visitor log
and then reached over the receptionist’s desk and into the top
drawer and grabbed a badge. If you execute this maneuver with
enough confidence, you walk right in. Show the slightest hesitation,
however, and the bored receptionist will confine you to a plastic
couch to await an escort.

In the news room, I walked past desks all jammed together, sev-
eral feet deep in old newspapers, press releases, books, and posters,
and sat down at the empty desk of my friend Bob Halliday. Bob is
officially a sub-editor, charged with turning copy for the Outlook
section into a semblance of English.

He is better known, however, for his book and music reviews.
He approaches legendary status in Bangkok for his restaurant re-
views, written under a pen name which means Sea Toad in Thai.

Bob and I have an understanding. I bring him offerings of food
and he introduces me to some of the best restaurants I’ve ever been
to. I call him periodically from the States and he briefs me in great
detail on the latest saffron chicken joint or some obscure noodle
pushcart that has a line of Mercedes a mile long waiting for food.

I dug into my dignified-looking black leather briefcase and
pulled out a jar of pickled prunes and the last remaining packet of
Astronaut Ice Cream. I also told him about the grisly fate of his
salami, destroyed by the Australian meat police, a story so sad that
it prompted Bob to wax poetic about a previous batch of sausages I
brought over.

After a brief period of contemplation over the salami (Bob was
sure that customs in Australia was at that moment gorging them-
selves on the links that were rightfully his), he began eyeing the ice
cream. I insisted that this chalky, freeze-dried substance would

224

Bangkok

have much better ice cream-like qualities if served cold, but I could
tell that my arguments were having little weight.

While Bob struggled with this dilemma, I went over to the of-
fices of United Press International to meet a friend who rents space
in the back. Peter Jansen is a long-time Asia resident who now
writes a report about doing business in Vietnam. Writing about do-
ing business in Vietnam is kind of like writing artificial intelligence
programs in Cobol, so I always make a point of looking Peter up to
see how he is getting along.

Peter and I walked over to the Dusit Thani hotel and took the
elevator up to the top floor. Instead of heading into the penthouse
bar, we took a set of stairs marked only with a sign reading “mem-
bers only” to the Foreign Correspondents Club of Thailand. Sitting
in this quiet bar, surrounded by pictures of wars and riots in Laos,
Burma, Vietnam, Cambodia, and Thailand, we drank a long string
of Kloster beers. Most expatriate residents of Thailand prefer Klos-
ter over the better-known Singha, Singha being rumored (probably
unjustly) to contain formaldehyde as a preservative against heat.

That night, Bob had a column to write, so we met up with two
other friends to do research. Peter Skilling is a noted scholar of Ti-
betan literature and Norman Bottorff spent years helping to coordi-
nate ABC News coverage for Southeast Asia, including extended
stints coordinating the coverage of hot spots like Tianamen Square
and the Iran-Iraq war.

The four of us piled into a taxi and headed out to Soi Nana.
Our taxi went skidding into a major road and we all groaned as we
saw that it was backed up for at least an hour. We piled back out
and walked over to Nasir Al-Masri, an Egyptian restaurant that, be-
fore the Kuwait war, would have been crawling with Arabs and
prostitutes eyeing each other hungrily.

The war had drastically reduced the Arab tourist trade and the
restaurant was filled instead with the world traveller set, sitting in
pairs looking sophisticated (and bored). We sat down, making lots
of noise, and Bob and Peter entered a long, complicated negotiation
in Thai with the waiter. Norman and I flagged down another waiter
and ordered beer.

225

Exploring the Internet

First to arrive on our table was a wonderful dish of hummus,
labeled on the menu as “creamy stuff of crains Conianing Sesame
Paste.” We instantly devoured the first plate and sent the waiter
back for more.

Soon, more and more plates arrived, filled with roast lamb, a
fava bean salad soaked in garlic, stuffed vegetables, and more.
Around us, the world travellers stopped eating and watched the pa-
rade of food to our table with growing awe. Mountains of bread
disappeared in a flurry as we dipped into plates, stopping only to
order yet more food or to hand empty plates to waiters as they
passed by.

After 18 plates of food, I could see that we had made the diary
of the backpacker sitting next to us. She was so impressed she for-
got totally about her own dinner, her mouth opening wider and
wider as we went for the world hummus-eating record.

We finished the last of the grilled pigeon and ordered Turkish
coffee, served in thimble sized glasses and having the consistency of
40 weight motor oil. Our research done, we paid the bill of 1,500
baht (U.S. $60), quite high by Thai standards, and left.

Friday morning, I walked a block from my hotel to Chitlom Towers
to meet Narong Intante, president of the Value Group. Like many
Thai companies, this one seemed to have a million names. The
lobby listing sent me to the Value Systems Co., Ltd. On the seventh
floor, the signs read Value Component Co., and Value Data Co, Ltd.

I sat in the lobby of Value* and sipped my cup of freshly brewed
Nescafé. The office was chilled to a level that signified a prosperous
company and all the office girls wore heavy coats to ward off the
cold.

I was ushered into Khun Narong’s office while he finished up a
deal for a whole bunch of workstations to some multinational. The
Value Group was the official distributor for Microsoft and I had
come to discuss the issue of intellectual property protection. Khun
Narong had started the company four years ago as a distributor and
it had grown to over 120 people.

226

Bangkok

I wanted to know more about how you make a killing selling
legal software in a country where piracy is not only rampant but
almost legal. For 1,000 baht (U.S. $40) a pirate shop will sell you a
complete Microsoft Windows, including disks and a photcopy of the
manual. For 5,700 baht, you can buy a legitimate copy from Khun
Narong.

Value Systems had recently introduced a version of Windows,
localized with Thai fonts and a Thai manual for 4,700 baht. To pro-
tect their investment, Value was using a hard lock, a piece of hard-
ware that must be present (usually on a serial port) for the software
to run.

Selling the simple English version, Narong estimated he had
provided only 2 percent of the copies on the market. With the Thai
version, however, he was hoping that legal copies would make up
20 percent, or maybe even 30 percent, of the market share.

With no real copyright protection, Narong had to rely on things
like support and service to lure in corporations. A few large Thai
corporations were beginning to be concerned about image and
would buy at least a couple of legal copies to put into their machine
rooms. A few even decided that legal copies reduced the amount of
internal support needed and were a sound business decision.

Increasingly though, it was a proliferation of viruses that was
driving customers into Narong’s hands. Many large corporations
had relied exclusively on pirate copies and some were paying the
price. I heard from several sources, for example, that one of Thai-
land’s largest banks had used pirated software not only for PCs but
for larger systems as well. My sources informed me that the bank
had been hard hit by a particularly virulent strain of virus and had
only just recovered.

Did the high price of software in Thailand have anything to do
with piracy? One of the more persuasive arguments in favor of
software piracy was that legal software was quite expensive, selling
for several times the U.S. list price in a country with a much lower
standard of living.

Narong saw this as a chicken-and-egg problem. As long as soft-
ware had no protection, there was no incentive to localize for the

227

Exploring the Internet

Thai market. Lack of Thai language support kept the market low
and the prices high.

This argument seemed to confirm an impression I had that most
of the piracy was destined for the Thai market. Although I had cer-
tainly seen foreigners in the Patpong red light district bring in a
laptop and say “fill ‘er up,” the vast majority of the copies made in
Thailand were made as an incentive to sell hardware.

I left Narong’s office and went out to the head of the road where
a half-dozen teenagers sat passing the time playing checkers with
bottle caps on a board marked on the sidewalk in chalk. Every
street had one of these groups, serving as a motorcycle ferry to
houses located deep within the complex maze of side streets that
spread out from each road. Every morning and evening, you could
see office girls being shuttled on these bikes, riding sidesaddle in
tight skirts and high heels, hanging on nonchalantly with one hand
as their chauffeurs went scuttling between lanes and over sidewalks.

That afternoon, I was briefed by a gentleman who insisted that I
identify him only as a “diplomat” from a “western embassy.” I re-
ported to the guard post of the Western Embassy, went past two sets
of metal detectors and incredible numbers of security guards, and
entered the inner sanctum of foreign soil.

There, I learned some of the background on the intellectual
property dispute between the U.S. and Thailand. The heart of the
problem was the patent law, which had explicitly disallowed protec-
tion for pharmaceuticals, leading to a huge market in clones. Drugs
can be a touchy trade negotiation issue, especially when a huge
American multinational is pitted against a poor, sick Thai child.

With the coup on February 23, 1991, a relatively honest govern-
ment had been put into power, surprising the hell out of the mili-
tary junta that had installed them to act as puppets. This new
government had been diligently at work passing laws, and one that
had almost gone into effect was a new patent law.

Likewise, trademark protection was finally going into place.
While trademarks had been nominally protected, penalties were so
lax that products ranging from perfume to watches to clothing were
regularly manufactured and sold both internally and to a huge tour-
ist market. Of course, the pirates don’t always get their copies ex-

228

Bangkok

act—I remember one Gucci watch knockoff with a picture of Snoopy
on the dial.

Copyright is a particularly sensitive issue in Thai politics. There
was a 1979 law on the books, but the problem was that the Thai
judiciary refused to enforce it, throwing cases out on technicalities.
Software was not mentioned in the law at all and no cases had gone
to court, so pirates had a virtual carte blanche to operate.

The government of General Prem, the first democratically
elected government in Thailand, introduced a new copyright law.
The legislation was perceived as a bow to foreign pressure in a
country that prides itself on its independence. General Prem’s
goverment fell over the issue.

Prem was followed by General Chatchai who had the honor of
presiding over one of the most corrupt Thai governments ever. I
remember sitting in Bahn Pitsanulok, the opulent former official
state guest house and the offices of the Prime Minister’s personal
staff. I was talking with one of the most senior members of the
Prime Minster’s staff in an on-the-record interview and asked him
what the current rate for bribes was on large government projects.
He promptly let me know that the usual rate was 10 percent, but
that this was, of course, negotiable. This same senior advisor cheer-
fully admitted that almost all the software in the offices was pirated.

This government was so venal that they were actually hurting
business, an unpardonable sin. The military, itself a huge business,
stepped in to protect their interests. Later, it came to light that sev-
eral government ministers had managed to save so much on their
meager salaries that in one case a minister increased his net worth
by U.S. $50 million during only a few years in office.

The government of the Anand administration was run by a half-
dozen technocrats who were busily trying to straighten out the mess
of the previous few years. (He was later replaced by Suchinda, the
junta leader.) Things were simmering on the diplomatic front and
negotiations had gone from the noise and publicity of 301 actions to
the back room workings of the diplomatic process. Progress was
being made, but no firm results had been achieved.

229

Exploring the Internet

Back at the hotel, I walked down the deserted hall of the hotel’s
shopping arcade, each shop holding one or two very bored-looking
attendants, shivering in the cold of the air conditioning. My room
was located in a remote wing, absolutely deserted except for the
elevator button pusher. This poor kid stood for twelve hours every
day in an uncomfortable uniform, his only job being to push the up
button if he saw somebody coming to save them any unnecessary
wait.

In my room, I needed to recharge my notebook and I noticed
that, while I had packed an appropriate two prong adaptor, I didn’t
have an adaptor for the three-prong, grounded plugs. I called
housekeeping and a uniformed engineer was sent up. I explained
the problem and he nodded and smiled, clearly understanding my
dilemna. He reached over and grabbed the ground plug and started
to snap it off. I recoiled in horror.

“No, not good,” I protested, grabbing back all three of my
prongs and moving to the other side of the room.

“Cut! Cut!” he repeated over and over, chasing after me. The
music wafting in from the hall speakers was playing “Somewhere
Over the Rainbow.”

Saturday morning, I went out to the Pratunam area, where textile
vendors crowd the sidewalks with bolts of cloth and silk, piled onto
tables covered with umbrellas to provide some shade from the over-
powering heat. Occasionally, an overloaded motorcycle would dart
out of a shophouse and onto the sidewalk, somehow missing the
milling crowds, and jump onto the street to go make a delivery.

At the edge of Pratunam is Pantip Plaza, a shopping center with
a large number of computer dealers. My favorite copy shop, a little
shophouse jammed with 6 PC ATs and 6 young women sitting in
front of them keeping the floppies humming, had been torn down.
These low-rent retail operations were getting harder to find in Bang-
kok and most of the businesses had moved into the big shopping
centers favored by the Thais, huge air-conditioned atriums with
three or four stories of balconies.

230

Bangkok

Pantip Plaza had its share of computer dealers, and even one or
two copy shops, but this was pretty tame stuff compared to the
Golden Shopping Center in Hong Kong. At Pantip, there were 20 or
30 shops, and a few had some halfhearted employees sitting around
playing games. A few customers drifted by, but most had the slow,
purposeless amble of the window shopper.

I snapped a few pictures of pirates in action, then gave up and
went down to the S&P Bakery for a dish of rice cooked with shrimp
roe and chilies. Roe, of course, being a Thai euphemism on English-
language menus for the orange fat in the shrimp’s head.

Dinner that night was at an unassuming little restaurant, a
brightly lit room with a few movie posters on the wall and lots of
big, round tables, around which were seated large families, each
with several generations present. Bottles of Mekong Whisky and
soda water and buckets of ice were piled up on little carts next to
each table.

We started with grilled pork, doused in a fiery salsa of lime
juice, garlic, and chilies and a big bowl of sour soup. The specialty
of the house was oysters and mussels, served half-cooked with egg
and vegetables, and we consumed a half-dozen plates of this deli-
cious seafood. Dessert was fried taro root, dusted in pistachio and
sugar, leaving even Bob so full that we omitted our traditional post-
dinner search for the ultimate durian ice cream.

Sunday morning, I left the hotel for my usual breakfast of a bowl of
noodles from a street stand. Returning to the hotel, I gathered a pile
of notebooks, business cards, and photos and walked down to the
business center. Two young attendants were seated at the desk,
waiting for business.

“T’d like to send these papers by Federal Express,” I said.

“One hour,” an attendant replied.

“Federal Express?” I repeated, with a puzzled look.

“One hour, must go to shop,” the other attendant insisted, reach-
ing for my stack of papers.

231

Exploring the Internet

After repeating this little routine a few times I realized that they
were referring to Photo Express, which, of course, didn’t really ex-
plain why they thought I wanted to have notebooks developed.

My flight wasn’t until evening, so I had time to have lunch with
Bob. We went to an old hangout, the Italian restaurant Pan Pan.
Normally, I try to stick to Asian food while in Asia, but Pan Pan is
an exception. The Italian food is good and, more importantly, they
have excellent durian ice cream.

Durian is one of those mysteries of the East. On the outside, it’s
about twice the size of a pineapple, with very sharp spikes about a
half-inch tall sticking out all over, making it advisable not to fall
asleep under a durian tree at harvest time. Inside, there are a half
dozen segments of creamy, pale flesh that looks sort of like a ba-
nana. The durian’s most famous feature, however, is its powerful,
distinctive smell. |

The taste is great, but the smell does tend to dissuade many
westerners from taking an immediate liking to the “king of fruits.”
I describe it as tasting something like a cross between a mushy ba-
nana and brie, but one Englishman I know refers to it as “a bit like
eating strawberries and cream in a public lavatory.”

Well, I took an immediate liking to it. The first time I came to
Bangkok, a mutual friend of ours, a freelancer who specialized in
travel writing and auto magazines, had arranged for Bob to meet
me at the airport.

We went straight to a night market where prostitutes, pimps,
and the like grab their midnight lunch. I was introduced to a wide
variety of tropical fruits, including mangosteens, rambutans, and
durians.

On a subsequent trip to Bangkok, I achieved a very modest de-
gree of fame by inventing the durian cheesecake. The creamy
durian mixed perfectly with the eggs and cream cheese, and a coco-
nut biscuit crust gave the cake a nice Asian twist. Of course, the
cake did still smell like a durian, but for many people that was con-
sidered a feature and not a bug. The recipe was an immediate hit at
the Post and Bob published it in his column the following week.

232

Amsterdam

Amsterdam was 0° C, certainly a change from Bangkok. I made a
mad dash to the taxi stand and got to my hotel with plenty of time
for a one-hour nap after my all-night flight.

At 1 PM., I took a tram to Muiderpoort station, then got in a
minibus that shot over a narrow little path running next to a canal
and dropped me off at the complex of buildings holding Dutch re-
search laboratories including the Centrum voor Wiskunde en Infor-
matica (CWI), the computer science and mathematics institute, and
the Nationaal Instituut voor Kernfysica en Hoge-Energiefysica (NIK-
HEF), the Dutch institute for atomic high energy physics.

NIKHEF and CWI have both been prominent centers of net-
working in Europe for many years. NIKHEF was the Dutch HEPnet
site, CWI had given an early home to EUnet, and this complex of
buildings on Kruislaan road kept popping up in one context or an-
other. It housed the RARE secretariat, it hosted a gateway to IXI,
and the president of RIPE was Rob Blokzijl, a NIKHEF employee.

It was to get to the bottom of all these projects and acronyms
that I had come to Amsterdam. I had heard more and more about
the maze of conflicting institutions, consortiums, organizations, and
associations that made up the confused politics of European net-
working.

I was just in time for the eleventh meeting of Réseaux IP
Européens (RIPE), an informal consortium of the TCP/IP research
networks of Europe. I figured that RIPE must be something like the
IETF, and thus a good place to gather information.

The meeting was about to start, so I gathered my badge and one
each of the dozens of handouts and went into a university-style lec-
ture hall. Down at the bottom of the hall sat three serious-looking
RIPE officials facing a room of 50 or so equally serious-looking engi-
neers.

Rob Blokzijl was going through a long agenda, reviewing action
items from the previous meeting and making announcements. |

233

Exploring the Internet

propped my jet-lagged eyes open with toothpicks and started leaf-
ing through the materials I had picked up.

RIPE was formed as a sort of anti-organization, a reaction to the
total ineffectiveness of other groups in setting up a pan-European
Internet. At the time RIPE was formed, there had been several years
of thrashing while people tried to figure out how to make OSI into
something real.

Meanwhile, several organizations with work to accomplish had
been putting together TCP/IP networks. An Internet was being
formed, but on an ad hoc basis. Networks were touching and acci-
dents were starting to happen.

“What kinds of accidents were happening?” I naively asked Rob
the next day while we were hiding in his office during a slack time
in the working group sessions.

“Routing loops, black holes, the usual,” Rob said. To try and
solve the problems, Rob had convened a group of six people from
the real networks in Europe. Nuclear physics was the lead player,
so in addition to Rob there were representatives from CERN in Ge-
neva and the Italian physics community. The Italians were operat-
ing the only 2 Mbps international link in Europe at the time. To
supplement the physicists, representatives from NORDUnet and
EUnet also came.

At first, the six tried to work through the Réseaux Associés pour
la Recherche Européenne (RARE), the association of research net-
working groups in Europe. RARE wouldn’t touch this issue,
though. The six researchers were operating TCP/IP networks and
RARE was exclusively and officially an OSI group.

Borrowing a phrase from Daniel Karrenberg of CWI that the
“time was ripe for networking in Europe,” an extremely lightweight
organization was founded in September 1990 with the mission of
helping to ensure cooperation among European IP networks. Rob
Blokzijl formally announced the formation of the group in RFC 1181,
a very carefully worded one-page document.

In Europe, careful wording is crucial and I’ve seen more than
one group spend several hours wordsmithing seemingly innocuous
phrases. Rob proudly pointed out the triumphs in the RFC, such as

234

Amsterdam

the sentence that “RIPE serves as a focal point for other common
activities of the participants related to IP networking.”

Apparently meaningless to the untrained eye, this sentence is
rife with subtleties meant to steer a course through the shoals of
politics. RIPE is a focal point, not an operational unit. It worries
about only its own voluntary participants and is not trying to im-
pose solutions on others. It deals only with IP networking and is
thus not advocating against OSI but merely serves as a forum for
those who have already chosen.

By the time of the eleventh meeting, RIPE had grown to nearly
30 organizations and was meeting quarterly. NIKHEF hosted many
of the meetings and acted as the organization’s secretariat. In terms
of developing new standards, RIPE was a pretty sleepy forum.
Those wanting a marathon of 50 working groups producing moun-
tains of paper were directed instead to the IETF. RIPE was much
more collegial and informal.

The week before the meeting, two days had been set aside for
informal tutorials to bring new participants in TCP/IP up to speed.
At the same time, a bake-off had been held featuring equipment
from various vendors and even including an ISDN switch. People
worked together to form different network topologies and see what
types of problems arose.

The meeting itself was a one-day plenary, followed by a day of
working groups, followed by another day of plenary meetings.
There were only a few working groups, devoted to issues like rout-
ing and European connectivity.

In addition to representatives from most of the central European
countries, there was a heavy contingent of East Europeans, includ-
ing delegations from Czechoslovakia, Poland, Hungary, and Yugo-
slavia. Most of the representatives from Eastern Europe migrated to
the working group session on connectivity. After a brief review of
who was coming on line, the conversation turned to problems of
mutual interest.

One of the biggest issues was simply getting equipment. All the
countries were on the U.S. export restrictions list and the Poles told
how it had taken them a year just to get a Cisco router. Various
suggestions were bandied about on just exactly how the USS.

235

Exploring the Internet

wanted the myriad forms filled out, but it was obvious that the
rules the U.S. imposed remained mysterious and arbitrary.

Another working group was devoted to the issue of routing co-
ordination. Routing was the single biggest technical issue at the
meeting, and it is in this area that RIPE made its biggest contribu-
tion. In the U.S., there was a single backbone which used the EGP
protocol to announce available routes to the second-tier regional
networks. The backbone was centrally managed, so all the regionals
got fairly consistent information. In Europe, there was no backbone,
so it was pretty much guaranteed that conflicting routing informa-
tion would be propagated, leading to routing loops and black holes.
RIPE took a look at the map of Europe and it became clear that four
sites acted as the main hubs.

These four sites, in Stockholm, Amsterdam, CERN, and Bologna,
served as the centers for large numbers of secondary networks.
Stockholm was the center of NORDUnet and Bologna was the cen-
ter of the Italian network. Amsterdam and CERN had huge num-
bers of links going out to different countries.

The four sites agreed to take the lines running from Stockholm
to Amsterdam to CERN to Bologna and turn them into a de facto
European backbone. The routers at the four sites were combined
into a single routing domain, thus presenting a consistent view of
the world. To make the backbone a reality, all four sites agreed that
transit traffic could traverse their links.

This was certainly not a centrally managed, fault resilient, highly
designed backbone, but it did the job of connecting the European
Internet. Since the four main sites also had links to the U.S., the
European Internet was well connected to NSFNET and the rest of
the world.

In a centrally managed backbone environment like NSFNET,
low-level networks that wish to make their availability known to the
world (and thus enable communications with said world) simply
have to inform the NIC or NOC running the backbone. If the low-
level network doesn’t fill out the appropriate forms or jump through
the appropriate hoops, the network is simply not announced and
thus is effectively isolated. It would be as if a road existed into a
country but there was no indication of the road on any map.

236

Amsterdam

In Europe, maintaining the name space was a trickier situation.
RIPE came up with a novel plan to keep information up to date. A
European WHOIS service was put into place containing network
names, domain names, IP addresses, and the like.

The incentive to keep this information up to date was quite sim-
ple. The four backbone sites used the WHOIS database to load their
routers. Sites that didn’t keep information up to date were simply
not announced over the de facto backbone.

RIPE was a simple, effective tool for coordination and instruc-
tion, all done on a shoestring. The informal nature of the group and
the advocacy of the heretical TCP/IP was too much for the network-
ing establishment and RIPE came under heavy attack at first from
officially sponsored groups like RARE.

To understand why RARE would object, one has only to remem-
ber that TCP/IP continues to be referred to by many as “DoD IP.”
In contrast, OSI, developed with very heavy participation from
European academics and PTTs, was the open way to do networking.

While RIPE was under attack, it was also clear that IP network-
ing was gaining a strong hold in Europe. In one of those bizarre
twists of European politics, RIPE became an official RARE “activ-
ity.” Not an official working group, mind you, with the connota-
tions of European Commission financial support, official voting by
one participant per country, and a chairman appointed by the RARE
Council of Administration. Simply an activity.

Under the merger and acquisition agreement, RIPE would con-
tinue as an independent group and would elect its own chairman.
RARE got to bring the renegades under its umbrella and unity once
again reigned, at least on the surface.

To tell the truth, I wasn’t exactly sure why the merger took
place, but I did get a glimpse when Rob told me about the new
RIPE NCC being formed. The informal RIPE backbone had been
working well, but members decided they needed a more profes-
sional Network Control Center. Not a mere NIC, several people
cautioned me (but certainly less than a NOC), this NCC would
serve a coordination and troubleshooting function.

The NCC would be funded by voluntary contributions. Since
RIPE was not a legal entity, the money would be funneled through

237

Exploring the Internet

RARE. RIPE/RARE had already put out an RFP for ECU 250,000
(U.S. $307,500), enough they hoped for a manager, two staff, some
equipment, a travel budget, and some office space. Obviously, it
was hoped that a group like NIKHEF or NORDUnet would respond
with an offer within the budget, and would donate resources on top
of the ECU 250,000.

A committee had been formed to evaluate the responses to the
RFP, consisting of two RIPE officers and two RARE officers. This
group, upon making its decision, would forward it to the RARE Ex-
ecutive, which in turn would make its recommendation and send it
on up to the RARE Council of Administration. RARE would then
officially hire (and fire) the manager.

Following this discussion, I must admit I was still a bit confused as
to what was going on. While I wasn’t clear on the organizational
ins and outs, it was certainly clear that RIPE was performing a valu-
able coordinating function, allowing the technical people to gather
in one room and work out solutions to problems.

Sometimes the issues on the table were highly complex, such as
how to do policy routing in an environment characterized by too
many policies. Often, though, the issues were painfully simple.

Take the issue of country codes, the two letter abbreviations by
which you send a letter or provide the root portion of a DNS name.
The list of such codes used to be fairly static, but with the disinte-
gration of the Soviet Union there were a bunch of new roots to the
name tree.

Managing that root was a sign of power, and in more than one
case, fights had broken out within a country between groups who
felt they were the rightful namespace administrators. The right to
control the top of the tree was a touchy issue in a place like Yugosla-
via. Imagine a Serb administration of the Croatian subdomain, for
example.

RIPE served as a place where these types of issues could be
ironed out. Not everything can be solved on an informal, technical

238

Amsterdam

level, of course, but it is amazing how many problems can be
worked out this way.

One exchange in the RIPE plenary that I found interesting was
the issue of where to find the current country codes. Officially,
these codes are listed in an ISO standard and, in response to a query
from the audience, a bureaucrat got up and explained that one ob-
tained ISO standards from one’s own national standards body.

Not a very useful solution to a simple problem, so somebody
else got up and suggested that the information be obtained from the
back of a Sun manual. The representative from Israel suggested that
Larry Landweber maintained this information as part of his list of
connectivity.

Finally, somebody wanted to know if this information wasn’t
online. He had heard about some standards server “someplace in
the U.S.,” but when he tried to obtain standards there he had re-
ceived an error message.

I had the dubious honor of trying to explain why the ITU had
withdrawn their permission for posting standards since the server
wasn’t helping “the right sort of people.” Many of the Eastern
Europeans were furious about the decision since it had been their
only access to the vital documents. Would ISO be posting their
standards, somebody wanted to know?

I explained the concept of a standards haven in some friendly

location like Budapest. Everybody laughed and the plenary broke
up.

That afternoon, I managed to drag Glenn Kowack away from the
center of the vortex he creates to find out about EUnet. Glenn had
just finished his first year as the CEO of EUnet. EUnet was sort of
the analog of the U.S. USENET, but had evolved into a much more
structured network.

In the U.S., the UUCP protocols had long been used to allow
people to form a very loose confederation of systems, a sort of net-
work anarchy. A new system would join USENET by setting up an
arrangement with some existing system for the transfer of mail and

239

Exploring the Internet

news, usually using dial-up links. The only real rule was the un-
written tradition that if somebody was nice enough to let you join,
you ought to do the same for somebody else. In this way, a loose
web formed across the U.S. and a de facto network came into being.
The informal nature of USENET meant no acceptable use poli-
cies or votes to allow a new member in. Lack of formal structure,
together with the low cost, meant that USENET (and the underlying
UUCP protocols) spread quickly among small startup companies.

Much of USENET ended up falling under the umbrella of
USENIX, the U.S. UNIX users group. In Europe, simple networks
were also forming in the context of UNIX user groups and the
UUCP protocols. In Europe, though, there wasn’t quite the luxury
for anarchy that the U.S. had.

For one thing, calls across national borders were expensive
enough that it made sense to concentrate traffic through a star con-
figuration of national nodes and a central switch. The central
switch was the famous mcvax (now mcsun), housed at CWI in Am-
sterdam. Since an international call to one point in Europe costs
pretty much the same as to any other point, there was no need for a
fancy mesh topology, and the star stayed in place.

Because calls cost so much, users started paying their share of
line charges from the very beginning. Administration and manage-
ment was totally volunteer and fell under the umbrella of the Euro-
pean UNIX Users Group (EUUG), and the net came to be known as
EUnet.

By the time Glenn joined EUnet in January 1990, it had grown to
be a TCP/IP and UUCP network with over 1,600 sites. Over the
next year, it was to grow to well over 2,500 UUCP sites, 100 IP sites,
and 750 news sites. All this traffic was going from local networks,
run by a local UNIX user’s group, into a national node, and then
over to Amsterdam.

EUnet, with all these users, began playing a major role in Euro-
pean networking, pooling its money with other groups like NOR-
DUnet so that bigger pipes could be leased. Many of these shared
lines formed the de facto European backbone of RIPE, and then later
of EBONE.

240

Amsterdam

One of the interesting aspects of EUnet is that it is often the very
first network in a country. By January 1992, EUnet already had 22
national members, each representing a national UNIX user’s group,
running a national node, and paying for their own line to mcsun.
The cost of international links, the Amsterdam equipment, and mis-
cellaneous overhead like Glenn were also apportioned among the |
members.

With the craze towards open systems at a peak, EEUG changed
its name to EurOpen to signify openness, truth, and beauty. Thank-
fully, EUnet did not become OpenNet.

Both parts of the EurOpen name are applied in a fairly loose
fashion. Open included things like UNIX and TCP/IP instead of
being a code-word for only OSI. European is also interpreted kind
of loosely. Iceland, Tunisia, and Moscow are members of EUnet,
stretching the word far beyond the confines of the European Com-
munity, the European Free Trade Association, and even the Euro-
pean continental boundaries. Egypt and Algeria were coming on
line soon and Glenn didn’t really seem to care who joined as long as
he could provide service and spread networking into new countries.

In fact, it was the desire to spread networking to those who
could use it that got Glenn his job with EUnet, in a very indirect
fashion. In the 70s and 80s, Glenn led a reasonably normal life, tak-
ing jobs like that of an R&D manager for Gould in Champaign-Ur-
bana, Illinois. Of course, he did take a few years out to help found
and run a community radio station, a job that many engineers don’t
sport on their resumes.

At the tail end of his stint with Gould, Glenn ended up as the
first employee of the newly formed UNIX International. He enjoyed
the job so much that, after the organization was off the ground, he
quit and took three months off to go look at computers in Asia,
visiting places like Hong Kong and Singapore.

He came back from his sabbatical and set up shop again in
Shampoo-Banana, this time hanging out his shingle as a freelance
consultant. He was doing various consulting jobs when Eastern
Europe collapsed.

Glenn bought a Zenith laptop, packed his backpack, and took
off. By the time he looked up, he found himself in Warsaw chairing

241

Exploring the Internet

the first meeting of the Polish UNIX Users Group, an event made all
the more challenging by his lack of any knowledge of Polish.

In his travels, he made enough visits through Holland that he
got to know the Amsterdam crowd well. EUnet decided to hire a
real manager, and Glenn certainly seemed not only highly qualified
but highly available. He took the job.

EUnet is proof that networks can be set up with minimal outlays
of funds. With a headquarters budget of ECU 600,000 (U.S.
$738,000) and a total pan-European outlay of well under ECU 10
million per year, EUnet has become one of the key European net-
works, acting as the lead player in bringing new countries into the
world Internet.

EUnet has also played a key role in helping to form the EBONE
consortium. EUnet’s historical role in working with other organiza-
tions on line sharing has meant that lines to the U.S. and across
Europe have been possible even when groups like RARE were un-
able to act.

One would think that a low-budget, technically astute, ex-
tremely effective network would win plaudits from all over, but
EUnet has its detractors. OSI-only groups view EUnet and its infor-
mality as a definite threat. In some cases, active efforts are under-
way to try and prevent EUnet from operating.

A recent example was brought to light at the RIPE meeting.
SWITCH, the national research network in Switzerland, had barred
EUnet traffic from its network. Not only had the traffic been barred,
it was cut off without any advance notice to EUnet, causing several
days of confusion as people all over Europe tried to figure out their
sudden loss of connectivity.

At the RIPE plenary, Michael Nowlan, the Irish chairman of
EurOpen, read a prepared statement protesting the lack of notice.
What had really happened illustrated the type of politics that EUnet
encountered.

In each country, it is up to the national UNIX users group to
figure out how to move traffic from the national node out to the rest
of the country. In some countries, a simple dial-up star configura-
tion is sufficient. In others, there is enough traffic that a backbone
linking regional hubs makes sense.

242

Amsterdam

How to provide that backbone is up to each group. In some
cases, the group will simply lease lines and set up a backbone. In
other cases, line-sharing agreements with a national research net-
work might be established. In other cases, a contract can be set up
with some other entity to provide the backbone.

In Switzerland, the Swiss UNIX group (CHUUG) had been us-
ing SWITCH, the national research network, as a_ backbone.
CHUUG and SWITCH got into a squabble over various matters and
CHUUG decided to set up its own backbone. SWITCH set up their
routers to filter out EUnet IP traffic, thus cutting off connectivity
from the 22 EUnet countries to anybody on SWITCH or with a tail
link off of SWITCH.

Fortunately, these childish games are the exception rather than
the rule. In most countries, EUnet happily coexists with the na-
tional research networks. The national nets handle the universities
and the UNIX group handles low-volume UUCP clients, IP links to
small companies, and anybody else who falls outside the official do-
main of “research” as defined in that country.

With my work for the day done, I went out for a couple of beers
with Scott Williamson, the manager of the new Network Informa-
tion Center (NIC) in the U.S. The NIC had been recently transferred
from SRI to GSI, a subsidiary of Infonet. GSI and Scott had a rough
few months as things got going and Scott was now making the tour
of conferences like RIPE and the IETF to convince the community
that things were back in control.

Running the NIC is no easy task, I learned. Just maintaining the
WHOIS database requires an 800 Mbyte database running Ingres, a
condition specified by the Defense Communications Agency, the of-
ficial funding agency for the NIC. Even though a great part of the
traffic into the NIC comes from the NSFNET, not the MILNET, the
NSF had yet to set up its own similar facility, making Scott responsi-
ble for both networks.

One of the big problems was that the NIC had inherited its com-
munications facilities as part of the DCA contract, which specified a

243

Exploring the Internet

64 kbps link to the Internet. A pipe of that size is not enough for an
organization that is at the center of the Internet and the link was
quickly flooded. Only after the NASA Science Internet stepped in
with a T1 line did things start to get better.

Over beers, I learned that Scott is ideally suited for his job. His
background was managing a team of UNIX programmers for vari-
ous Washington tasks, such as a job for the Federal Emergency Man-
agement Agency (FEMA) simulating telephone disasters. More to
the point, though, he is a second-degree black belt in a Korean mar-
tial art and spent three years in the Australian outback. Both experi-
ences breed qualities that are useful when dealing with Washington
bureaucrats.

That night, RIPE had set up a group dinner at a local Indonesian
restaurant. Indonesia was a former Dutch colony and an Indone-
sian dinner provided excellent food, at a price that was affordable
on all manner of budgets, an important consideration given the
wide geographical representation at the meeting.

Over salads in gloppy peanut sauce, preserved eggs in sweet
chili and lemon grass, and fiery beef curries, the conversation
drifted in and out of shop talk. I was too jet lagged and burnt out
to think about networks, so I concentrated on the food.

244

Utrecht

Wednesday morning, I brought my bags down to the train station
and joined a stream of Dutch commuters. It is not unusual for peo-
ple in Holland to live at one end of the country and work at the
other. I was going half-way across the country to Utrecht, a half-
hour away.

The train ran parallel to one of the main Amsterdam canals,
lined with houseboats and barges, then past warehouses and wind-
mills. Passing the suburb of Amstel and the sprawling IBM com-
plex, we started picking up speed. Suddenly, we were in the
countryside, an elaborate system of canals crisscrossing the fields
and forming a series of little islands, connected by bridges and
sluices.

It was market day in Utrecht and, since I didn’t have a hotel
reservation, I headed right into the middle of the town square to
look around for a place to stay. Spotting one, I avoided the tempta-
tion to stop at stands full of cheeses, herring, and baked goods and
got a room.

With a few hours to kill, I brought my laptop into a little café by
a canal, one of a half-dozen places crowded with tables of high-
school and college kids sipping coffees and beers. At 2 PM., I gath-
ered my things and went into the shopping center complex
surrounding the train station to find Kees Neggers, co-managing di-
rector of SURFnet, the Dutch academic and research network.

After seeing so many networks like EUnet in Amsterdam and
AARNet in Australia run with a couple of guys and a Sun, it was
certainly a change to see SURFnet occupying a whole floor of plush
offices. When I asked Kees how many people worked there, he said
only 20, but many of them were responsible for supervising subcon-
tractors.

SURFnet is a company that is owned 49 percent by the Dutch
PTT and 51 percent by the SURF Foundation, a group started in-
itially by the universities, but which has grown to include research
institutes, the research arms of corporations, and other institutions

245

Exploring the Internet

of higher education such as teaching hospitals and vocational
schools.

To explain SURFnet, SURF, and the PTT, Kees went to the white
board and carefully drew the organizational structure, drawing my
attention to the legal status of all the parties and pointing out the
differences in Dutch law between the limited liability partnership,
the foundation, and the association.

As Kees continued his chalk talk, it occurred to me that he had
not said one word about the network itself. I knew that SURFnet
was born as a project of SURF in 1986, had become a real company
in 1989, and I even knew about things like the fact that it had a
capitalization of 3 million guilders (U.S. $1.6 million), and expendi-
tures of NG 14 million per year, of which NG 10 million were cost
recovery services from users and 4 million was development paid
for by the government. What I didn’t know was what the network
did.

Well, it turned out that SURFnet started by using the X.25 data
service from the PTT. After some long thinking and a big RFP,
Northern Telecom got the contract to install and manage a private
X.25 network based on 64 kbps links. This net was based on three
major hubs, with each leg of the triangle having two or three 64
kbps links. Another 16 sites were equipped with switches, each of
these having at least two lines into the network. Altogether, 413
sites had connections.

The network appeared to be your basic FOO over X.25 net.
Most of the services had SURF* names. SURFmail, for example,
was described in the annual report as “using the RFC-822 address-
ing mode as an intermediate step towards the international X.400
standard.”

Recently, the network had begun the process towards a 2 Mbps
backbone. To quell any political problems, five sites were given 2
Mbps links and a “pilot” was run. Some lines ran straight IP, others
ran straight X.25, and some ran a multiplexed mess of the two.
Needless to say, the results of the pilot showed that running IP ona
line was more efficient than running IP on top of X.25. However,
with X.25 identified in many European countries as the “pathway to
OSI” it was important to conduct the pilot.

246

Utrecht

All this technical information I learned in a few minutes from
Kees and by leafing through the annual report. Very quickly,
though, we were back on the subjects of organizational charts, ex-
amining how his network management staff served as contract su-
pervisors to monitor the subcontractors that were doing network
management.

We moved on from SURFnet to pan-European politics, an area
in which Kees is an extremely active player. His first foray was as
the EARN representative for his country. When RARE was formed,
SURF was named as the Dutch national member, and Kees was
named SURF’s representation to the RARE Council of Administra-
tion (CoA).

RARE, the Réseaux Associés pour la Recherche Européenne, was
started in 1986 to promote the use of networks based on open sys-
tems solutions, read “OSI.” This was to be sharply contrasted to
EARN, IBM's proprietary protocol, association, and network, foisted
upon unknowing European countries as part of an alleged plot to
dominate the market. Well, maybe not quite that sharply con-
trasted, but it would not do to underestimate the fury of this par-
ticular religious war.

At that time, only the U.K. really had a national network. Many
people assumed that OSI would be the basis for networking and
that the PTTs would provide the infrastructure on which OSI would
run. That infrastructure would be a commercial offering, univer-
sally available, and available soon.

The focus on the PTT as the service provider led to the assump-
tion of a single national network, and thus RARE is made up of one
member per country. There are also various associate members, in-
ternational members, and liaison organizations, but the basic gov-
ernance of the organization is by the Council of Administration
which has one seat per country.

There is another element of RARE that is worth noting. It gets
money from two sources: contributions from member countries and
from the European Commission. The European Commission com-
ponent is strong enough that they wield considerable clout on
RARE decisions.

247

Exploring the Internet

Most of the linkage between RARE and the Commission has
been through the COSINE project. (Every acronym must be pro-
nounceable and preferably have a double meaning of some sort,
however inane. In the case of COSINE, the expansion yields Coop-
eration for OSI Networking in Europe.)

COSINE was started in 1986 by another group called Eureka, a
research fund subscribed to by a group of countries roughly equiva-
lent to the RARE membership or the European Free Trade Associa-
tion. The COSINE project of Eureka was handed over to the
European Commission to act as project coordinator and they, in
turn, picked RARE as the secretariat.

What all this meant was that Eureka kicked in some money, the
Commission kicked in some more, and a fraction of that money
flowed downhill to RARE. Eventually, so the theory went, money
would flow down even further into a series of market-pull activities
that would prepare and stimulate OSI in Europe.

Much of the research funded by the Commission was in the area
of “pre-standardization research.” Pre-standardization was the
process of preparing a bunch of documents that are then used as
input to the standards process. In the case of COSINE, this meant
establishing the subset of OSI that would best fit the special needs
of Europe.

The functional profiles of OSI were prepared by a series of
working groups. Every country got a member. The chair could in-
vite four “experts,” sort of a standards patronage position. The EC
paid for the travel expenses of each delegate, kicked in for the cost
of donuts at the meetings, and took care of the RARE secretariat.

Paying people to go to meetings in fine places has the sure result
of generating a large demand for meetings. The specification proc-
ess dragged on as people tried to define which portions of OSI
would be used in Europe. The assumption was that when those
profiles were developed, the lure of the common European market
would be enough to attract vendors like flies to a water buffalo.

Vendors and researchers took one of two approaches. A few
went ahead and just implemented the portions of OSI that made
sense in a product. ISODE was a perfect example of this. Most just
went ahead and sold TCP/IP products.

248

Utrecht

After a couple of years of this process, a set of 10 very thick,
very blue volumes of COSINE specifications were issued. These 10
blue books were summarized in a red book, which in turn was sum-
marized in an orange executive overview. I suspect quite a few cor-
porations took that orange distillation and concentrated it yet
further until senior management was hearing summaries such as
“this stuff is great, trust us.”

Once the specifications were complete, RARE developed a CO-
SINE implementation plan. In January 1990, this implementation
plan was signed as a contract between the Commission and RARE.
The RARE COSINE Project Management Unit (PMU) became a real-
ity.

COSINE was designed as a three-year plan with a budget of
ECU 30 million (U.S. $36.9 million). Most of that money, over ECU
12 million, went into IXI, the International X.25 Infrastructure (IX]).
A large chunk, ECU 6 million, went into the PMU.

IXI is an international X.25 network for researchers. Countries
like Greece, Yugoslavia, Ireland, and Portugal all use IXI as their
main path into the European mainland. Even the U.K., for some
strange reason, links itself into the continent with IXI.

The network is simply a star configuration of a set of 64 kbps
lines that link national X.25 networks. Typically, a private, virtual
X.25 network is built on top of the PTT’s own offering. There are
actually two stars, one in Amsterdam and one in Bern. Two 64 kbps
links connect the stars, and each country gets one 64 kbps link into
one of the hubs.

Except for the U.K., that is. The U.K. is big enough that it really
needs two 64 kbps links into Amsterdam. Of course, that means
that you’re pumping 128 kbps of data into a 64 kbps network, but
that’s another story.

In a classic example of how networks will grab whatever re-
sources they can, much of the IXI traffic consists of setting up a
virtual circuit to NIKHEF, which operates a gateway from IXI into
the Internet. While running TCP/IP on top of X.25 is not the most
efficient implementation in the world, when that’s your only choice,
it works fine.

249

Exploring the Internet

The theory behind IXI, and all COSINE projects, was to stimu-
late the market with the initial implementation. At the end of the
three-year trial, some commercial entity would see the inherent
profitability of the enterprise and step forward boldly with an ag-
gressive bid to take the project over as a going concern.

At the end of 1992, IXI would terminate, and there had not yet
been a clamor for rights to run the service commercially. Unless, of
course, there might be some Commission money in it. This had oc-
curred to a few people, and RARE was about to start lobbying for
an Operational Unit.

Meanwhile, a second force reared its ugly head: people needed
networks to do their jobs. Most of those national networks had
started carrying large amounts of TCP/IP traffic, and many RARE
members had to weigh the relative merits of keeping a network run-
ning against the desire to attract Commission money.

RIPE, of course, was the answer at first. RARE got in the act
with the EBONE proposal, in which Kees took a leading role, along
with groups like EUnet and NORDUnet. EBONE took the informal
cooperation from 1991 and turned it into a more structured consor-
tium for 1992.

Since the meeting I had attended in November at the Amster-
dam Zoo, the EBONE proposal had taken shape. Two key groups,
CERN and IBM, had decided not to sign the formal memorandum
of understanding, but would still cooperate in the project.

Once the decisions to join (and how much to contribute) were
made, the resources were turned over to an EBONE Action Team
(EAT) to turn money, routers, people, and lines into a reasonable
network.

A 512 kbps backbone had been decided upon between Stock-
holm, London, Amsterdam, Hamburg, and CERN. Three links to
the U.S. were available, with two operating at 512 kbps and a T1
line from CERN (the EASInet IBM line). In some cases, money had
been used to upgrade existing facilities. In others, existing resources
were simply put into the EBONE pool. 2

The distinction was important because at the end of 1992, ac-
cording to the Memorandum of Understanding that chartered
EBONE, the group would disband and all resources would revert to

250

Utrecht

owners (except for the money of course, which is kind of hard to
give back once you've spent it). No formal company existed and
there was no dedicated EBONE manager.

At the end of 1992, the official plan was to have RARE start an
Operational Unit. The Operational Unit would provide a home for
things like EBONE and maybe even projects like IXI. The Opera-
tional Unit would be set up with capital from RARE members (and
presumably some nice research grants from the Commission to keep
it going). Officially, the Unit would operate as a separate organiza-
tion, but with RARE members contributing the capital, RARE mem-
bers would get the shares, and presumably, RARE would have some
influence in how those shares got used.

Setting up the Operational Unit and operating it as a more for-
mal replacement for EBONE could be looked at one of two ways.
On the one hand, it could be the badly-needed professional opera-
tion that would run the long-heralded pan-European infrastructure.
A few cynics looked at the Operational Unit in another way, seeing
a power grab that could hurt existing operations such as EUnet.

The official raison d’étre for the Operational Unit was contained
in a report Kees handed me entitled “Final Report of the RARE Task
Force on the Establishment of the Operational Unit for the Supply of
Network and Information Services to the R&D Community” with a
bright yellow cover and attractive GBC binding.

This report was a marvel of detail, containing a complete busi-
ness plan for the Operational Unit with a cash flow analysis through
1996, job descriptions for OpUnit officers, and an annex of com-
monly. asked questions and answers, my favorite being ‘““Why are
data networks important for Europe?”

This report explained how a professionally managed backbone
was key for Europe. In fact, it was assumed that you needed to
provide this capability through a single organization, sort of the
way that the NSFNET had provided a single backbone service in the
US.

The flaw in this analogy was that the NSFNET had not provided
the single backbone service in the United States or for the Internet.
The backbone was, as soon as multiple networks came into being, a
combination of resources.

251

Exploring the Internet

The NSFNET, for example, had to cooperate with the MILNET,
the replacement for the ARPANET. NASA already had NSI, Energy
had ESnet, and corporations had their own networks. A single
backbone provider was certainly not the only way to structure Euro-
pean networking.

As I was looking through the business plan for the OpUnit, Kees
handed me another document, a report from the European Engi-
neering Planning Group, which in turn had been set up by the
RARE Council of Administration.

The EEPG document had a grand plan for networking in
Europe. Kees was a key member of the EEPG. The EEPG outlined
a scheme in which research networking in Europe would need, to be
a success, four bodies. An Operational Unit was one of them. A
networking association (RARE), a policy body (people with money),
and a “consultative networking forum” were the other three bodies.

Of course, there were other networking associations in Europe,
but RARE had been doing a pretty good job of setting up agree-
ments. RIPE had become a RARE activity, a status certainly placing
it lower in the hierarchy than an association.

That left EARN. RARE had been engineering an acquisition and
merger of EARN, proposing that the two groups form as a way of
increasing their mutual power. Like a giant game of RISK, with
Europe under control, RARE was taking an active role in the inter-
national scene, participating in places like the Co-ordinating Com-
mittee for Intercontinental Research Networking (CCIRN) and
co-founding the Internet Society.

This wasn’t everything, either. I learned that RARE had helped
form the European Workshop on Open Systems (EWOS) which
would help set up European OSI profiles. RARE was also part of
the European Telecommunications Standard Institute (ETSI), the of-
ficial pan-European telecommunications standards body which was
worrying about issues like ISDN or 2 Mbps X.25 networks and had
joined the European Council of Telecommunications User Associa-
tions (ECTUA).

I left Kees’ office with my head swimming. He had let me know
that he had to leave by 4:06. I had thought the departure time was a
bit odd, but it occurred to me that Kees was in the shopping center

252

Utrecht

of which the train station was a part and that he had probably
timed the walk down to the trains. Just in case he felt like getting a
jump on things, I left at 4:03.

Walking back to town, to keep warm, I went through all the new
acronyms I| had learned, hoping that repeating them would generate
enough hot air to counteract the chill. Needing a cup of coffee and
a blank piece of paper to sort everything out, I ducked in out of the
cold into a little corner coffee shop with windows all fogged up a
definitely counterculture clientele.

Rubbing my hands to warm up, I stepped up to the counter to
order.

“Do you have a menu?” I asked the attendant who ambled over
from a table where he was playing chess.

He pointed down on the glass. I quickly realized that I was at
the wrong counter. This menu specialized in regional specialties.
The left hand side featured different kinds of hashish, ranging from
your basic Lebanese to your more exotic Nepalese. The 10 and 25
guilder columns contained how many grams (or portions of grams)
you got for your money. The right hand side of the menu was de-
voted to fine marijuanas, featuring orange bud, sinsemilla, and the
like.

Not needing to confuse myself even further, I went down the
steep spiral stairs to the coffee bar in the basement. There, amidst
the kids rolling joints, the din of pinball, and the very loud music, I
read through the stack of documents I had collected that day.

253

Bonn

Thursday morning, I sat on the track at the Utrecht station waiting
for my train to Bonn. If you’re a few minutes early, its fun to go up
to the tracks and watch the other trains leave. The 9:23 local was
leaving on my track and I watched the clock turn to 9:23. Exactly
two seconds later, the train was on its way.

One of the questions that always puzzles me is what conditions
lead to a successful infrastructure. The European networking infra-
structure was certainly not as developed as it could be. Likewise,
the telephone infrastructure had always been expensive and not al-
ways on the forefront of technological change.

Yet, the European passenger train system, developed in a highly
fragmented period of highly sovereign governments, was a miracle.
The trains go to every town of consequence and many of none.
Even more remote places are connected to the transportation web by
a series of bus systems, coordinated with the train schedules, of
course.

Staring at the track, hoping for a flash of insight, I realized that
my train was about to leave. A few hours later I arrived at the
Hauptbahnhof in Bonn and pulled out my itinerary of travel reser-
vations. To my horror, I discovered that I was booked in a suburb
of Bonn, Konigswinter.

My instructions to my travel agent had been simple. “No Holi-
day Inns, no suburbs.” Bonn, like Prague, is one of those cities that
never has enough hotels and I was lucky I wasn’t as far away as
Cologne, the provisional capital.

I wandered around the train station for a while trying to figure
out how to get to Konigswinter. I spied a sign for the tram system
with the magic word on it and caught the first one leaving. The
tram was labled Bad Honef and I crossed my fingers that the track
didn’t serve multiple locations.

When I got on, I realized that I didn’t have a ticket. All over,
colorful signs in multiple languages indicated that tickets must be
immediately validated or I would face a DM 60 (U.S. $36) fine.

254

Bonn

Having no idea how to get a ticket, or where I was going, I alter-
nated between peering out the window looking for some appropri-
ate sign and looking nervously around for some sign of the ticket
police.

Konigswinter turned out to be a resort on the Rhine river. My
hotel had a view of four castles up in the hills, was right across the
town square from the church belltower, and my room was right on
the river, looking out at barges and ferries, the tram, a boardwalk,
and people out strolling in the cold air.

I went down to the sun room to work. Old men, couples, and
mothers with children came in periodically to get a cup of coffee, a
beer, or some ice cream. I ordered a cup of coffee, with the waitress
checking to make sure I just wanted one cup.

The next morning, at breakfast, the waitress wanted to know if I
wanted my coffee in a “can or a cup?” Outside it was still dark and
I could make out the fog banks in the hills. Suddenly, the fog
started to glow and I could see the black silhouettes of the castles.
Then it was daylight and time to try to make my way into Bonn.

I took the tram back into the main train station and went in
search of an information booth. I waved a business card in their
face and they sent me to track 2. As I was trying to decipher the
map of the extensive tram system, one pulled in. I suddenly real-
ized that I had been sent to the wrong track, narrowly missing get-
ting on one going the wrong way.

Armed with a valid ticket, I arrived at the stop in front of the
monolithic police station headquarters, walked a few blocks over to
the offices housing GMD, and found Klaus Birkenbihl.

Klaus is the manager of the Netzzentrum fuer die Wissenschaft,
the science networks center in the Institut fur Anwendungsorien-
tierte Software- und Systemtechnik at the Gesellschaft fur Mathema-
tik und Datenverarbeitung mit beschrankter Haftung. In the
interests of saving space, it is hoped the reader will excuse the infor-
mality of referring to the organization as simply GMD.

GMD is a government-owned research institute specializing in
mathematics and computer science. Like INRIA, the French insti-
tute, the institute gets 60 to 70 percent of its funds from the govern-

255

Exploring the Internet

ment, raising the rest through contracts with industry. GMD has
1,200 staff, of which 700 are scientists.

I had come to see Klaus to find out about one more network, the
IBM-sponsored EASInet. I had always assumed that BITNET, and
its European cousin, EARN, were IBM’s contribution to networks. I
heard periodically about EASInet, but did not know anything about
the network nor why IBM should start yet another network. EARN
had started in 1984, originally with six countries but, after Dennis
Jennings heard about it, six countries plus Ireland. The IBM model
was to fund the program for three years, after which it would be-
come self-sustaining. EARN added many countries and did, in fact,
become self-sustaining.

Not only did EARN become independent, but you might even
say the board got a touch hostile, planning a transition to OSI and
cozying up to DEC. The OSI focus was actually a political necessity,
part of a compromise with the PTTs who agreed to loosen some of
their restrictions on clients sharing leased lines in return for a prom-
ise by EARN to migrate to OSI as soon as the protocols became
technically and financially viable. This all happened in 1984 and the
OSI migration took a bit longer than anyone expected. By the time I
got there, most EARN sites were migrating to the BITNET 2 proto-
cols, more commonly known as TCP/IP.

The upshot of the OSI (and later TCP) focus was that it was
quite obvious that EARN was no longer an IBM animal. Creating a
network that was self-sustaining after three years was a great public
service, but didn’t really serve the other goals of maintaining the
IBM presence in Europe and, in the long run, selling more iron.

IBM developed a new proposal as a way of providing an incen-
tive to the big mainframe market in the research community known
as the European Academic Supercomputer Initiative (EASI). Under
the program, IBM plunked down large amounts of heavy metal in
selected research institutions and paid for the cost of 64 kbps links
between participants, the collection of those links forming a fairly
hefty donation of bandwidth. In addition, IBM paid for the cost of
a T1 line from CERN to NSFNET and hired GMD to manage the
network.

256

Bonn

EASInet consisted of 18 sites by the beginning of 1992, with loca-
tions from Spain and Italy in the south, to Stockholm, Amsterdam,
and Hamburg further north. Most of these sites were not exclu-
sively “EASInet” sites, but were on a variety of networks. CERN,
for example, an EASInet site was sometimes referred to as the Cen-
ter for European Research Networking.

In a few cases, the EASInet link was in fact a dedicated 64 kbps
line between two sites. In most cases, however, the IBM money had
been thrown together with other sources. For example, a 256 kbps
line between Amsterdam and CERN was split between HEPnet,
EUnet, and SURFnet. The line sharing arrangements that EASInet
started were an important early example of the model later used in
EBONE.

To handle multiple protocols, many of the EASInet lines used
NET’s IDNX time division multiplexers, with most of the band-
width given over to TCP/IP and SNA and other, smaller, circuits
applied to X.25 and DECnet. EASInet was really bandwidth. In
some cases, GMD actually took care of the lines, in others the job
was split with other organizations. Some of the data flows were
point-to-point sessions, others were part of broader networks such
as HEPnet or the European Internet.

When IBM had agreed to cooperate with EBONE, what they
meant was that the T1 line to NSFNET would become a pathway to
the U.S., and thus to Asia. Even though IBM had not signed the
EBONE document, the link would be available to transit traffic.

Helping coordinate EASInet was only one of GMD’s projects. The
site also acted as a key gateway between Germany and the rest of
the Internet. Germany had long been one of the most ardent OSI
advocates. In 1984, the ministry of research and technology started
up a membership organization, the Deutsches Forschungsnetz
(DFN), the German research network intended to give the German
scientific community the full benefits of OSI.

GMD, part of the federal research establishment, ended up with
a sort of schizophrenic role. On the one hand, they took part in the

257

Exploring the Internet

fervent OSI cult built around the leadership of their funding agency,
helping DFN on conceptual work and software development.

On the other hand, GMD had work to do and actively helped
build EARN, running the German national node since 1987. At the
same time, GMD started building a private X.25 network to link
German national research laboratories together.

This private network, AGFnet, was not OSI (in fact it was SNA),
but at least it contained X.25, the “pathway to OSI,” to make it po-
litically palatable to the bureaucracy. What AGFnet did do was
prod DEN into action, which resulted in a national X.25 network
called Wissenschaftsnetz (WIN or “science network”).

DFN had a notable success in building WIN: they made it af-
fordable. With the fairly strong backing of the federal government,
DEN was able to convince the German PTT to give it preferential
rates and, most importantly, to avoid imposing volume-based
charges. The result was a rate of DM 5,000 per month for a 64 kbps
line. By 1991, WIN had spread throughout Germany and much of
the research community was on the network, including commercial
researchers such as Daimler Benz.

After what Klaus wryly termed “a period of difficulty,” DFN ac-
cepted the premise of multiprotocol networks. In fact, with users
paying for their own usage, it turned out that the vast majority of
sites chose to use TCP/IP or EARN protocols.

There was still some OSI work on the network, particularly since
the federal government was willing to finance such research. X.400
usage, in particular, had significant usage and was growing quickly.
GMD operated application gateways to the Internet and EARN
worlds to keep connectivity.

Hanging in Klaus’ office, high up near the ceiling, was an abacus
with a little mouse perched on top. This odd little arrangement was
the Birkenbihl massively parallel supercomputer, given to Klaus by
his daughter. “Look,” he said, pointing to the blue and red beads,
“it even has a color graphics display.”

I went back to Konigswinter, hoping to warm up with a nice,
hearty German meal, but it was the off-season and all the interesting
places were boarded up. A half-dozen Greek restaurants appeared
to be Gyros-only joints, with nary a drop of taramasalata in sight, so

258

Bonn

I settled on an Italian restaurant. There I was served limp, over-
cooked noodles in a heavy, congealed green sauce.

“That was the worst pesto I’ve ever had,” I told the waiter as I
paid up.

“Thank you very much,” he said with a proud smile.

I finally found a sausage shop open and filled up on a couple
mugs of Konig-Pilsner and a Riesenbackwurst, a greasy sausage
served on a slab of cardboard with a puddle of mustard. Much bet-
ter than the lime-green pesto.

The next morning, checking out of the hotel, I watched the clerk
punch my room number into the computer and we stood staring at
each other with forced smiles while the dot matrix printer chugged
away. The clerk retrieved the printout, then walked over to an an-
cient cash register and punched in the room charges, one by one.
He then picked up a thick sheaf of bar bills and added them in, one
by one.

He took that printout over to a calculator, where he added in my
minibar charges. By then, I was not in the least surprised when he
took that slip over to his desk and figured out the tax, adding it in
by hand.

“Not very modern,” he said with a shrug.

I had to agree.

259

Brussels

I took the early morning Schnellzug to Cologne. With over an hour
to kill, I put my bags into a locker and checked out the bar and café.

It was 8:00 AM., but the bar was standing room only—there
were no chairs but lots of men standing around drinking their pre-
train beers. In the café, there were eight of us and five had ciga-
rettes going, all crammed into a space smaller than a Manhattan
studio apartment. I went back to the bar.

Two minutes before departure, I went up to the track just as the
train to Brussels pulled in. I boarded and spent a dreamy two hours
staring at sleepy villages and isolated farm houses, built of stone
and sitting precariously on hills in the middle of black fields.

We entered Brussels, the center of the densest train network in
Europe. Our track joined others and soon we were picking our way
through a yard dozens of tracks wide. Then, suddenly, we were
inside the cavernous central train station.

I spent the weekend catching up on my reading and looking for
open restaurants in a city where it seemed like even the red light
district closed on weekends. Brussels prides itself on being at the
geographic center of Europe, and thinks of itself as the logical capi-
tal of the European Community, home to the Commission. I think
of it more as the Indianapolis of Europe.

After the sleepy calm of the weekend, I figured my two days
trying to find out about the infamous Directorate Generale 13 (DG
XIII) would be relatively simple and might even yield useful infor-
mation.

This optimistic mood was soon to change. The weather got even
colder, the food got toxic, the taxi drivers started getting into acci-
dents, and, worst of all, I entered the bureaucratic swamp of the
European Commission.

260

Brussels

Monday morning, I walked over to the European parliament build-
ing and up the Rue du Loi, past huge monoliths filled with func-
tionaries, to the Rue de Trévres, home to a portion of DG XIII. The
European Commission security system requires visitors to leave
their passports as identification, but since I had a fax from a DG XIII
PR officer I was allowed through. Security didn’t even look at the
fax to make sure that the letter wasn’t a request to have me perma-
nently barred from Europe.

On the second floor, I was ushered into the office of Peter
Johnston, an official of the F Directorate of the 13th Directorate-Gen-
eral of the Commission of the European Communities. We were
soon joined by another official, also named Peter. There was one
seat free on my side of the little conference table, but Peter 2 took
great pains to clear off a chair and align himself with Peter 1 on the
other side.

With both Peters in place, they began an interchangeable litany
of platitudes, trying to explain everything the Commission was do-
ing from promoting networking to ensuring the future of interna-
tional trade. I had heard about things like COSINE, RACE, PACE,
RISE, and the European Nervous System and naively asked if per-
haps one of the Peters could explain these pieces, perhaps starting
with the RACE project.

There was my first mistake. You see, RACE is a program, not a
project.

“No, no, no,” Peter 1 chided me, explaining how, in the context
of the third framework there were a series of programs such as
RACE, each of which was made up of dozens and dozens of pro-
jects. RACE was no mere project, thank you.

I learned little of substance in my briefings, but later, digging
through over 20 pounds of documents I had amassed, I was able to
put together a cursory picture of the programs and projects and
themes and frameworks that the Peters were so unable to articulate.

DG XIII is the portion of the bureaucracy of the Commission
charged with looking after the areas of telecommunications, infor-
mation technology, and innovation. DG XIII has taken an active role
in industrial policy, funding a great deal of research and develop-
ment, and also handling some regulation.

261

Exploring the Internet

The reason for DG XIII is that Europe has traditionally been a
stronghold of the PTTs, monopoly service providers of postal serv-
ices, telecommunications, telegraph, and even, in many cases, bank-
ing and other sundry services. Although the size of the European
economy is about the same as the U.S., fragmentation of telecommu-
nications has led to conflicting standards, closed markets, and ineffi-
cient service providers.

The Commission was taking a variety of steps as a regulatory
body to try and open up telecommunications, including requiring
PTTs to relax their procurement policies so all European suppliers
could compete. Other regulatory steps included requiring a single
112 emergency phone number for all of Europe and a single stand-
ard for mobile radiotelephony, replacing the 6 different standards
that had sprung up in the 12 EC member states.

In all this regulation, DG XIII worked with the existing PI's,
trying to get them to change their practices. —The Commission was
less than enthusiastic about taking steps that might lead to the PTTs’
demise. In fact, the second prong of the Commission's strategy was
to couple regulation with a very strong industrial policy, trying to
keep the industry competitive.

This was no minor research program, either. In the fields of in-
formation technology and communications, the Commission was
funding R&D to the tune of ECU 2.221 billion from the period 1990
to 1994. (The ECU was worth roughly U.S. $1.25 in 1992).

To understand the Commission, it is best to defer the natural
inclination to ask where the money was spent and what resulted
and instead focus on the process.

The quintessential model for an EC program was Esprit. Started
in February 1984, DG XIII bills Esprit as “the key to reviving Euro-
pean technology.” The premise behind Esprit was matching funds.
The Commission would put up half the money for a research pro-
ject, the participants would put up the other half.

Rather than have researchers suggest some interesting research,
the Commission would issue a call for proposals. To respond, a
group had to have at least two industrial partners from at least two
EC countries. Typically, a successful group would have up to a half-

262

Brussels

dozen members, including research institutions, consulting firms,
telecommunications companies, and computer companies.

Esprit shelled out ECU 1.5 billion and involved 3,000 researchers
in the first phase from 1984 to 1988. From 1988 to 1992, Esprit
shelled out another ECU 3.2 billion, involving almost 6,000 re-
searchers.

Probably the best known result of Esprit is the transputer, an
example of which is the T800 chip, a 10-MIP, 32-bit, RISC processor
useful as a building block for parallel computers or signal process-
ing units. Of course, there are several other chips that the
transputer must compete against, such as those from Motorola, In-
tel, and the Sparc consortium.

Other projects helped train engineers in VLSI design or devel-
oped opto-electrical equipment for video reception at medium dis-
tances without amplification. Esprit officials categorize these as
“major results” and point to “over 500 major results” in the Esprit
program.

I examined the list of major results through the end of 1989.
There were 343 major results, and it is interesting that 43 of these
were in the category of having made a “substantial contribution to
the preparation of international standards.” If prestandardization
activities got their own category and were counted as major results,
it was obvious that all this money was certainly not going to de-
velop new technology or to conduct real research.

Though the bulk of Esprit money went to things like high-defi-
nition TV standards, a substantial sum was devoted to an area of
research projects known as the “Information Exchange System.”
Best known of these projects was COSINE and its IXI backbone.
IXI, connecting 20 different X.25 networks in 17 countries, was listed
as the key infrastructure for Information Exchange.

A COSINE brochure proudly stated that IXI was transferring 25
Gbytes of data per month by the Spring of 1991. The fact that IX]
was several years late in coming was not mentioned, nor was the
fact that 25 Gbytes per month is not a tremendous amount of data.
The German DFN X.25 network, for example, was transferring over
70 Gbytes per month of TCP/IP data over X.25, not to mention sub-
stantial traffic from other protocols such as SNA.

263

Exploring the Internet

There were several other information exchange projects proudly
trumpeted in the COSINE literature, but my favorite was ROSE, Re-
search Open Systems for Europe. The project started in 1983 and
ran for 60 months, involving five European vendors such as Bull,
ICL, and Olivetti. ROSE had the novel goal to promote OSI and
“reinforce standardization work by implementing the standards and
demonstrating them.” Ata gala demonstration at the Esprit 88 con-
ference, the vendors all demonstrated OSI prototypes that included
FTAM, X.400, and the X.28/X.29 PAD services.

With Esprit such a shining success, DG XIII used it as a model
for other industrial policy programs. The STAR program, for exam-
ple, was started to try and get the less developed portions of the
telephone infrastructure jumpstarted. From 1986-91, the Commis-
sion pumped ECU 780 million into places like Northern Ireland,
Portugal, and Greece, with matching funds coming from the mem-
ber states.

These types of government intervention can be quite useful if
the money is put into the right places. The program that really
caught my attention was RACE, Research and Development in Ad-
vanced Communications Technologies for Europe. RACE was sup-
posed to get a broadband ISDN network into Europe and if there
was any place a massive infusion of help was needed, it was the
European telephone system.

As the two Peters explained to me, RACE was focused on the
area of integrated broadband communications (IBC) and was meant
to avoid the problems that ISDN had encountered on its way to
market.

According to Peter 1, ISDN had run into a few setbacks because
the standards were too vague, leading to incompatible implementa-
tions. While the ISDN generation of technology was getting on its
feet, it had taken a bit longer than anyone had planned.

In other words, ISDN was a fiasco, consumed huge amounts of
resources, and produced few results (but did so consistently over a
long period of time). RACE would make sure the next generation
of technology had a smoother roll-out.

If the problem with ISDN had been vague standards, then RACE
would make sure that all the downstream activities from stand-

264

Brussels

ards—the implementation profiles, conformance specifications, fea-
ture subsets, and the like—were formulated in great detail, thus
avoiding incompatibility.

DG XIII had set itself the highly ambitious goal of a B-ISDN
network by 1995 and was investing heavily. Between mid-87 and
mid-92, RACE would involve 294 organizations in 92 projects for a
total outlay of ECU 1.2 billion (U.S. $1.47 billion), of which the Com-
mission was pitching in ECU 550 million.

The RACE literature explained that by bringing all the European
players together in one project, the risk of failure was reduced. Eve-
rybody would work together for a common goal. Putting all your
assets into one investment was a strategy I had not learned in my
class on portfolio analysis, but I was willing to hear the argument
out.

RACE hoped to somehow speed up the standardization process,
a “well-known bottleneck in the exploitation of high technology.”
Rather than reduce the bottleneck by reducing the size of the stack
of paper and focusing on implementation experience, RACE was
going to try and increase the volume of paper produced.

The result would be the integrated broadband communications
architecture, a pile of documents that would specify everything nec-
essary for the network. This suite of documents would, of course,
be adopted by the rest of the world, opening up the vast external
market for European suppliers to exploit.

RACE, one of my glossy brochures exultantly proclaimed, was
“the current focus of attention of all the intellectual and industrial
work on advanced telecommunications in Europe.”

When you talked to the Peters, though, you got the idea that the
real purpose of RACE was to reinforce the dominant position of the
PTTs. ATM cell switching, the technology that forms the basis for
most of RACE and B-ISDN, was to Peter 1 simply a means to elimi-
nate “rate arbitrage.”

Rate arbitrage is one of those code words that summarizes all
that is evil about those who insist on setting up their own networks.
By this philosophy, the only reason that a value added network (i.e.,
anybody but the PTT) can exist is because entrepreneurs take ad-
vantage of differential tariffs for different speeds.

265

Exploring the Internet

If the tariffs did not vary, the philosophy goes, everybody would
buy their services from the underlying network provider, the PTT.
The evil of the Valued-Added Network (VAN) is that these greedy
businessmen skim the cream off the top, diverting funds away from
the PTT that is trying to provide service to everybody. If the profit-
able business accounts go to the VAN, who takes care of the poor
widow in the village?

This line of argument makes a lot of sense unless you ask a
more fundamental question. Do the PTTs have the capability to do
the work in the first place? If not, the cream-skimming argument
breaks down at that first assumption. Stopping others from doing
important work is not an effective strategy for building an infra-
structure.

While I was mulling over the proper way to get high-speed fiber
and ATM switches into place, the two Peters had left the now-bor-
ing details of the RACE program for their newest toy, a program so
new it didn’t even have an acronym.

This new program was in the area of telematics, bureau-speak
for applications that use networks. To the tune of ECU 376 million
from 1990 to 94, the Commission was funding half of the projects in
seven telematic areas, including transportation, health care, libraries,
and government.

I had been wondering what the purpose of Peter 2 was, and |
soon learned that he was heavily involved with the transportation
portion of the telematics program. Peter 1 ceded the floor and Peter
2 began his presentation.

He talked for 10 minutes, at which point I started leafing back
through my notes. During those 10 minutes, Peter 2 had not uttered
one sentence that was evenly vaguely comprehensible. I under-
stood the words, but was unable to parse any of the sentences, let
alone extract any semantic content.

Evidently, there had been a separate program of research on in-
telligent road systems, known as Drive. Drive successfully com-
pleted a total of 71 projects in 1991, presumably with many major
results and at a cost of ECU 60 million. Of course, overhead took 10
percent away, but this was still real money.

266

Brussels

Peter 2 kept referring to the “road transport informatics in the
context of the Drive 2 programme,” but I had a tough time under-
standing what that actually meant. Luckily, Peter 1 realized the
problem and cut in.

“I must insist,” he said, arching his eyebrows, “there is no Drive
2 programme.” Rather, it was more appropriate to refer to the
transportation “theme” within the larger context of the telematics
program.

I still didn’t know what this theme/ program/ project cluster was
actually doing, so I asked Peter 2 to give me an example of one of
the pilot projects or some technology that was actually being used.

He couldn’t. He mumbled on about harmonization of standards
and finally explained, as if talking to a small child, that it was “hard
to pick out a particular report from Drive and say this was used
here,” pointing to the table in front of him.

It was clear to them that I didn’t realize that I was in a room
with policy makers, not some mundane road technician or highway
analyst. Peter 1 gave me a lecture about the vast scope of DG XIII.
It was as if I had asked the Chairman of the Board of EXXON to tell
me how many gas stations he had in Lexington, Kentucky. (Of
course, if you asked the Chairman of EXXON to name any one city
in which he had a station, chances are he might be able to answer
that question.)

I decided to try another tack. The Peters and 40 of their col-
leagues seemed to spend most of their time issuing contracts under
various themes. Was the process computerized?

Peter 1 was insulted. Peter 2 sniffed.

“Of course we’re computerized,” Peter 1 said, waving vaguely
in the direction of his desk, which bore an Olivetti PC and a
Hewlett-Packard laser printer.

“What kind of network do you use?”

“Uh, Novell,” one mumbled uncertainly. The other one looked
down and shuffled some papers.

“And what kind of DBMS do you use to manage your contract
boilerplate library?” I wanted to know.

A look of panic crossed their faces. Peter 1 gave an exasperated
sigh. Peter 2 seemed to think they were using Foxbase. Peter 1 rose

267

Exploring the Internet

and handed me some brochures. Peter 2 excused himself for an im-
portant meeting upstairs.

I took the subway out to the isolated suburb of Beaulieu, where DG
XIII has office space, to meet Jurgen Rosenbaum, an information of-
ficer. We paid our mutual respects and lied about how much help
we would be to each other. I collected another 10 pounds of paper,
then left to try and find someplace to grab lunch in the hour before
my next meeting.

Beaulieu has at least four huge office complexes and dozens
more under construction, and is surrounded by modern looking
apartment blocks. Yet, with all these people, there was not a single
place to eat, or even get a newspaper. This was city planning at its
best.

I got back on the subway and went a few stops into town, hur-
rying to find a lone Spanish restaurant down the street from the
station. After a quick lunch of fish soup (preceded by a 25 minute
wait), I hustled back to the subway to find my next bureaucrat.

Halfway through our audience on Esprit, I felt myself getting
more and more nauseous. I realized with horror that this wasn’t
caused by platitudes but that I was feeling the all too familiar onset
of food poisoning from the fish.

As I rode into town, I felt worse and worse, just making it back
to the warmth of my mediocre hotel before collapsing. I spent the
entire next day in bed, trying to gather enough strength for Wednes-
day’s train ride to Paris.

I left Brussels puzzled by the European Commission. There is
certainly a place for a strong industrial policy, as the Japanese dem-
onstrated with their MITI program and as the U.S. Defense Ad-
vanced Research Projects Agency repeatedly showed with TCP/IP,
UNIX, and a host of other successes. Yet in the one field I felt quali-
fied to judge, the Commission appeared to have been a total failure.
The European Internet was coming into being, but only after fight-
ing active opposition from DG XIII.

268

Paris

On the Euro-City train to Paris, I started to feel a little better. The
conductor looked at my Eurail pass and smiled broadly.

“Ah, un American! C’est tres bon!” he exclaimed with a classic
Parisian accent. This was about as out of character for a Parisian as
it would be to have a New Yorker tell you to “have a nice day.”
Still, it was a welcome diversion from the dour gloom of Belgium.

As the train passed through St. Quentin and headed up the Oise
river, the countryside started getting familiar, the houses looked
French, and even the farms had a characteristic look. I arrived at
the Gare du Nord and stepped into the car of a virtuoso taxi driver
who shot down alleys and back streets, delivering me to the
Montparnasse district for the remarkable fare of FF 70. I gave him
100, the amount such a ride would normally cost.

Flipping on CNN in my room, I was just in time for a commer-
cial featuring “beautiful, bountiful, beguiling, Belgium.” Popping
open a beer from the minibar, I mentally added “bullshit and bu-
reaucrats” to their list.

After watching President Bush’s State of the Union address, |
bundled up and went up the Boulevard Montparnasse to fetch my
92-year old great-aunt, stopping on the way for a paper sack of
roasted chestnuts to warm my hands and popping into Le Chien
Qui Fume for a quick drink.

Over a salad of preserved quail on a bed of dandelions, water-
cress, onions, and gooseberries, I tried to explain to my aunt the
concept of a technical travelogue. She listened carefully, frowned a
bit perplexedly, and told me that she was very proud of me, what-
ever it was that I was doing.

Thursday morning, I took the Métro to the edge of the city, then
a train out to Versailles. My taxi took a detour around the immense
castle that had housed Marie Antoinette and over to the neighboring
suburb of Rocquencourt, where INRIA has a facility.

My 10 A.M. appointment wasn’t there. He had sent me mail, but
when I logged into a borrowed terminal I saw, to my horror, that I

269

Exploring the Internet

had received no new mail messages for over a week. Something
was drastically wrong.

The secretary at INRIA told me where I had to go for my meet-
ing. On the way back into town, I tried to figure out what had
happened. The problem was no doubt in my home PC (or, more
likely, in the way I had configured it).

The domain of Malamud.COM is a registered domain. When
somebody starts to send mail to any of the millions of possible ad-
dresses in the Malamud.COM domain, a record in the Domain
Name System points to my commercial service provider, Colorado
SuperNet.

Colorado SuperNet takes all incoming mail and spools it on a
special UUCP-only account I maintain on their machine. A few
times a day, my PC calls up that account and retrieves all incoming
mail messages. Those messages are sorted, combined with my MCI
Mail, any fax notifications I can grep out of my fax log, and some
miscellaneous status messages, and the whole lot is sent right back
up to the interactive account on Colorado SuperNet that I use while
on the road.

In desperation, I stopped at the train station and gave a call to
my fax number. Since the fax is answered by a fax board in a PC, it
stands to reason that a high pitched tone on the other end would
tell me that my PC was still up and running, and, by logical conclu-
sion, that my house was not a charred, smoking wreck.

Having successfully pinged my house (but not solved the mail
problem), I continued on to La Défense, a complex of convention
centers, shopping centers, exhibit halls, and other public spaces, all
surrounding a futuristic looking arch, kind of a 21st-century rendi-
tion of the Arc de Triomphe. There, next to the World Trade Center,
is the InfoMart, a three story high set of exhibit spaces for computer
vendors.

Most of the vendors are your typical IBM, Bull, and Microsoft
variety, but there are a couple of odd ducks. One is a “house of the
future” exhibit, built to the new European home automation stand-
ards. This house-like display, full of spas, jacuzzis, exercise bikes,
and other things I would never allow in my own house, was a coop-

270

Paris

erative venture of a dozen vendors, all trumpeting the future of do-
motics.

The other shop that doesn’t really fit in is one rented by four
research agencies, including INRIA. There, I met with Milan Sterba,
a young Czech with joint appointments at INRIA and the Prague
School of Economics.

Though officially he had been in the business school in Prague,
Milan had spent most of his time on the large Czechoslovakian pro-
ject to develop an MVS-like operating system to run on the reverse-
engineered System 370 clones manufactured in Bulgaria and Russia.
Milan had spent several years working in his chosen specialty, tele-
communications, developing a VITAM-like telecommunications ac-
cess method.

At INRIA, in addition to real work like reconfiguring Sun work-
stations, Milan had an informal role as one of the focal points for
East European countries trying to get on the Internet. He main-
tained a document detailing current connectivity, chaired sessions at
RIPE meetings, and otherwise helped to spread information around
where it was needed.

The rapid progress in the former Eastern Bloc and Soviet Union
had been truly amazing. As fast as the countries could persuade the
U.S. to process the paperwork for Cisco routers, countries were
plopping in TCP/IP nodes, enhancing EARN connections, and us-
ing UUCP and EUnet to spread connectivity into new places.

Bulgaria, for example, was using a dial-up UUCP link to Am-
sterdam for connectivity. Dial-up was then being used to send mes-
sages and files to the 10 Bulgarian EUnet sites. Dial-up UUCP was
a common first step for many countries, a fact worth keeping in
mind before sending long messages to people on the other end of
slow (and expensive) lines.

Poland was typical of countries with higher levels of connectiv-
ity. Poland had a 9,600 bps leased line between Warsaw and Copen-
hagen and used statistical multiplexors to combine TCP/IP with the
EARN NJE/BSC protocols. Another 9,600 bps leased line between
Krakow and CERN used DECnet protocols as part of HEPnet, but
was converted to IP as soon as an export license for a Cisco router
was approved. A 64 kbps leased line between Warsaw and NOR-

271

Exploring the Internet

DUnet in Stockholm was going operational in 1992 and would
greatly enhance TCP/IP connectivity to the country.

Milan and I talked about the effect that posting standards had
had in Czechoslovakia. Milan told me that, to his knowledge, only
one copy of the ITU Blue Book existed and people spent consider-
able time going to a central facility to consult standards documents.

The ITU and others in the standards cartel had always insisted
that the standards were reasonably priced for those “serious” about _
doing work. Milan confirmed that this position was nonsense and
that many countries in Eastern Europe had particularly welcomed
the ability to access documents they needed for their work.

I found it particularly distressing that the ITU policies were hav-
ing the effect of preventing people in developing countries from ac-
cessing technical standards. After all, one of the purposes of the
United Nations, of which the ITU is a key part, is to promote a
world community. Keeping key documents hidden from those
without money, indeed keeping documents hidden from entire
countries without money, is certainly a convoluted perversion of the
UN mission.

That night, I had dinner in a Venetian restaurant, a classic place
with a dozen tables and the owner acting as the head (and only)
waiter. In his 50s, stout, distinguished looking and impeccably
dressed, the owner showed people to their tables, practicing the art
of being apparently servile while in reality insulting everyone he
could. While ignoring my waiter and eating a mediocre mezzaluna
funghi, I read Don Fernando, Somerset Maugham’s classic essay on
Spain. Maugham was writing about the playwright Lope De Vega,
author of 2,200 plays.

Many had proudly pointed to the size of this document base,
equating De Vega’s proclivity for producing paper with greatness.
De Vega himself had not taken this seriously, remarking that “if any-
one should cavil at my plays and think that I wrote them for fame,
undeceive him and tell that I wrote them for money.”

When you reward people for producing paper, they will do so.
When the European Commission and RARE paid people’s expenses
to go to meetings and make specifications, it is not surprising that
they took so long to make them, nor that they were so voluminous.

272

Paris

Likewise, when you take a stack of OSI documents and put
them next to the RFC series, you can tell pretty quickly which ones
were produced by standards professionals and which ones were
written by engineers who had software to write and networks to
run.

Friday morning, I met Jean-Paul Le Guigner of the Comité Réseau
des Universités, a committee of French universities formed to set up
a national research network, much like EDUCOM in the U.S. and
the vice-chancellors committee in Australia.

We took the subway to Jussieu, changing trains and walking
through the long, intricate tunnels connecting Paris Métro lines,
ending up at the Université Pierre et Marie Curie. Widely acknow-
ledged as one of the uglier campuses ever built, the university was
constructed during the period in the 1960s and 1970s when slabs of
unfinished concrete were considered “modern.”

At the university, we found the office of Christian Michau of the
Unité Réseaux du CNRS, Jean-Paul’s equivalent in the research
world. CNRS, the Centre National de la Recherche Scientifique,
acted as a coordinating body for French research institutions like
INRIA and ORSTOM.

The two groups, the universities and the researchers, had
banded together to form Renater, the Réseau National de Télécom-
munications de la Recherche. Funded to the tune of FF 50 to 60
million (around U.S. $10 million) per year, this network would form
a national backbone of 2 Mbps in 1992 with the core expanding to
34 Mbps in early 1993.

Renater, like the NSFNET in the U.S., was set up as a network of
networks, linking regionals together and France to the rest of the
world. Unlike NSFNET, the interface to the network would be pro-
vided by the telephone company, in this case France Telecom.

France Telecom, under contract to Renater, would offer a point
of presence to which regionals could connect. The regionals, like
the R3T2 network I had seen in Sophia-Antipolis, were locally
funded. In the case of R3T2, for example, local and regional net-

273

Exploring the Internet

works had started the network as a way of attracting industry and
promoting education.

After checking to see if my house was spitting up mail yet,
Christian, Jean-Paul, and I headed over to Christian Michau’s office
and sat around a conference table. We were joined by two other
gentlemen who were somehow joining Renater and needed a brief-
ing on the project.

It was quickly established that having me speak French was a
better strategy than forcing the other four to speak English. Chris-
tian suggested I start with a brief description of my project. Some-
how, while I was researching this book I found that many of my
hosts, though always very hospitable and full of information, never
quite understood the concept of a “technical travelogue.”

I complied, giving a little speech and feeling quite content that I
had remembered enough French to pull this off. Everybody smiled
and nodded gravely, each in turn thanking me. Then, one of the
observers leaned over to Jean-Paul and suggested that perhaps a
good way to proceed would be for somebody to explain exactly
What the American gentleman was hoping to accomplish.

Once this was all straightened out, Christian and Jean-Paul pro-
ceeded to tell me about Renater. In France, research institutions and
universities have a great deal of independence. To make things dif-
ficult, any attempt to establish a national research network would
cut across three different government ministries.

Through a minor miracle, in February 1991 an interministerial
convention was signed which established the authorization for Ren-
ater to begin planning. A pilot committee was formed, which
spawned a technical committee, which created a project team. By
the end of January 1992, a contract with France Telecom was days
away from signing and a pilot network was already in place. Re-
markably quick, given the number of ministries and bureaucracies
involved.

Renater was designed to be a multiprotocol network, but with
an interesting twist. The networks in many countries, such as Ger-
many’s DFN, ran all protocols over X.25 (“the pathway to OSI”).
Running TCP/IP over X.25 was certainly a combination that

274

Paris

worked, but for connecting two hosts together a straight leased line
made more sense.

Renater was set up in a very pragmatic fashion as two networks,
one presenting an IP interface, the other an X.25 interface. At 2
Mbps, the X.25 network would simply layer on top of the Transpac
public X.25 network. The IP network, a totally separate system,
would be a series of Cisco routers on leased lines.

The network would be quickly upgraded to 34 Mbps in early
1993, at least on core routes. It was hoped that at these higher
speeds that ATM-based cell switching would allow the X.25 and IP
networks to coexist on the underlying substrate.

Presenting an X.25 interface solved several problems. OSI could
run on the network, so all parties in Renater were able to boldly
proclaim that theirs was an OSI network which incidently happened
to support “immediate needs of existing traffic.” The X.25 network
also gave a platform on which DECnet and SNA traffic could run.

Michau and LeGuigner were decidedly pragmatic about their
work, trying hard to avoid political battles and concentrate instead
on getting the network up and running. They strenuously avoided
religious decisions they didn’t have to make, using mechanisms like
study groups to look at questions of SMTP versus X.400.

What I found most interesting was the close working relation-
ship with France Telecom. The telephone company seemed to view
Renater as an opportunity instead of a threat, using the research
community as a place to test technology and building expertise for
future commercial rollouts.

A 5-year contract worth roughly U.S. $10 million per year had
been signed with the telephone company, giving them a substantial
incentive to provide a working network. In addition, a contract pro-
vision was added that any tariff decreases in the future would be
used to upgrade the bandwidth of the research network, thus guar-
anteeing a fixed amount for the telco and at the same time protect-
ing the research community.

How had Renater managed to avoid the need to make OSI their
only protocol? Michau had one of the better answers to the ques-
tion of OSI I had seen.

275

Exploring the Internet

“Many countries in Europe are beginning to realize that OSI is
not X.25,” he explained. Instead, he looked to standards like OSI to
provide solutions to a range of problems, particularly at the applica-
tion layer.

If you accept OSI as a set of services instead of some monolithic
religion, useful protocols like X.400 and X.500 can be applied imme-
diately instead of waiting for an entire grand design to come into
being. In fact, you could, with this view, run X.400 on top of
ISODE, which in turn can run on top of TCP/IP, and still provide
OSI service.

Defining OSI as a series of applications took Renater out of the
debate. After all, Renater was simply providing an infrastructure, a
core backbone, and it was up to the users to decide what to run on
that backbone.

Running an X.25 interface kept Renater in the OSI game. Even
for international links, Renater was willing to use X.25 links. While
acknowledging that “some choices are not technically perfect,” it
was Clear that Renater would consider projects like the 2 Mbps IX]
extension. After all, the European Commission favored X.25 and the
Commission had lots of money. It was clear that the larger Euro-
pean countries like France would quickly catch up to the U.S., put-
ting in high-bandwidth backbones, setting up regional networks,
and running multiprotocol environments.

That night, I picked up my great-aunt for a quick dinner. Over
glasses of port and a dish of snails soaked in garlic and butter, we
talked about her childhood in Russia, her emigration to Montreal
and New York in the early part of the century, and her escape from
the Nazis in the war, walking over the border to Free France with
her 6 year old daughter in the middle of the night. She had been in
France for 60 years, even receiving the Legion of Honor for her
work at the Pasteur Institute, where she conducted research and
later helped numerous visiting scientists get settled. After walking
her back to her apartment, I went to my seedy hotel to pack and
trade insults with the snooty night manager.

276

Washington, D.C.

February 1, having visited nine countries in January, I happily
landed at Dulles airport, taking my carry-on luggage straight
through customs and into my waiting rental car. My hotel was a
minor miracle, cheap and comfortable. I went straight down to the
bar in the basement and settled in with the New York Times, a bacon
cheeseburger, and a bottle of Brooklyn Lager, a beer as “vital and
diverse as the city itself.”

Sunday evening, I met Tony Rutkowski, who had quit his job at
the swamp of the ITU and moved back to Washington to work for
Sprint International. Although we had exchanged lots of e-mail
messages, it was still nice to talk in person. While e-mail is certainly
convenient, person-to-person is still a communication medium that
has a much higher semantic bandwidth.

Monday, I made an early morning reverse commute out to the
suburbs, thumbing my nose at the parking lot heading into the city.
In fact, it was so early when I got to Reston that I stopped into the
local Sheraton for breakfast.

“We are glad to be able to contribute to the start of your day,”
the menu proclaimed. “We bid you a wonderful day.” Thinking I
might have misread my itinerary, I pulled it out to make sure |
wasn’t in California.

After my power breakfast of a bagel with lox, I wandered over
to the headquarters of the Corporation for National Research Initia-
tives (CNRI), the institutional home of such network luminaries as
Drs. Vinton Cerf and Bob Kahn.

My first meeting was with Vint Cerf, the Chairman of the In-
ternet Activities Board (IAB) and an engineer whose involvement
with TCP/IP dates back to the very first ARPANET nodes. Vint is
famous for his ability, despite a huge stack of lofty responsibilities,
for keeping up a detailed knowledge of the protocol suite at all lay-
ers.

Vint Cerf is the kind of person who can switch from briefing a
congressional aide to an in-depth discussion of dynamic routing

277

Exploring the Internet

protocols, public key cryptography, or the inner details of multi-me-
dia messaging. I have never seen him attend a working group
meeting at technical forums like the IETF without making at least
one substantive contribution.

Technical issues, though still demanding, now have to contend
with organizational questions, issues that suck up more and more of
his time. The most pressing issue was the future of the IAB and the
IETF, two organizations that had been remarkably successful but
were institutionally adrift.

The IAB started as a technical advisory committee to the De-
fense Advanced Research Projects Agency, but when the ARPANET
died, the IAB was no longer an official body. Through technical
leadership rather than governmental fiat, the IAB continued to
guide the development of the TCP/IP protocols in the Internet com-
munity.

The IAB had to bear the liabilities of its DoD origins in a
changed world. It didn’t help, of course, that the IAB remained an
all-American group until 1991, when Christian Huitema from France
was asked to join. (Not to mention the fact that in 1992, the IAB
remains exclusively composed of white, western males.)

Lack of accountability and lack of international representation
made it increasingly difficult for the IAB to exercise leadership. The
standards-making activities of the IAB and the IETF also raised is-
sues of due process, antitrust liability, copyright, and a host of other
issues difficult to resolve without a more formal organization.

Vint Cerf’s solution to these problems was to announce the for-
mation of a new professional association, the Internet Society. The
IAB and the IETF would become activities of the Internet Society.

The Society managed to attract quite a few other projects as
well. The INET conference which evolved from the Larry Landwe-
ber seminars would become the annual meeting of the Society. Tony
Rutkowski would run the publications board and hoped to produce
a newsletter of topical information and a journal of archival mate-
rial.

Launching an activity of this magnitude is never easy, and by
February 1992, Vint Cerf had already spent many months getting
the project off the ground. A key activity was selecting the initial

278

Washington, D.C.

trustees and drafting bylaws. Though the Society would be mem-
ber-run, the first board would appoint itself. Then, in 1993, the first
third of the trustees would be up for election. Within three years,
the entire 21 person board would have been elected by members.

Self-selection of the first board and the decision to have the first
official meeting at INET 92 in Kobe, Japan was causing some grum-
bling by American engineers, who were used to having IETF meet-
ings in accessible U.S. locations.

One of the key architectural assumptions of the Internet Society
is that it must be truly international to succeed in giving legitimacy
to the standards-making activities of the IAB. As such, Cerf and
many others felt that holding the first meeting out of the U.S. and
appointing a truly international board were crucial to its success.

A set of bylaws and other procedures, control by a group of
trustees, and other formal structures were also a change from the
IETF, which had run its first meetings by shouting out policy at ple-
nary meetings. Even in 1992, the IETF ran as a pretty loose sort of
organization.

More elaborate procedures were also crucial if the activities lead-
ing to a standard were ever challenged. Standards promote interop-
erability, but can also be challenged as a potential antitrust violation
if it appears that there is collusion to restrain trade.

The big challenge was how to change the process to put in some
form of accountability, yet still maintain the effectiveness and flexi-
bility of the IAB and the IETF. Too much process, and the whole
thing would degenerate into just another standards body, pumping
out paper and enjoying many fine lunches and dinners.

The Internet Society had been simmering in the six months since
the announcement of its formation at INET 91 in Copenhagen. De-
spite the low profile, 800 members had joined. Though this was a
small fraction of the 95,000 members of the Association for Comput-
ing Machinery (ACM), Cerf was encouraged and saw it as a good
start. The board was also falling into place, with members from Ja-
pan, Australia, and several European countries joining the American
members.

Next to Vint Cerf’s office is that of Bob Kahn, the founder and
president of CNRI. Kahn was a Senior Scientist at BBN responsible

279

Exploring the Internet

for the design of the ARPANET, then moved over to DARPA as Di-
rector of the Information Processing Techniques Office where he in-
itiated the largest computer R&D program ever undertaken by the
federal government, a billion dollar Strategic Computing Program.
He still plays a crucial role in the Internet, with CNRI coordinating
the 5 gigabit testbeds he helped to establish.

A former MIT professor, Kahn is described by most people who
have met him as an extremely quick study. Hand him any topic and
in a few minutes he will have something intelligent to say about it.
A classic long-range thinker, he devotes much of his time to the
question of how to build an effective infrastructure.

We talked about the “look and feel” of an infrastructure—what
somebody needs to know in order to use it. Cars and roads, for
example, require a fairly extensive knowledge base, ranging from
how to use the pedals to how to read a street sign. Other infrastruc-
tures, ranging from electrical systems to education to banking and
insurance, all have their own particular look and feel.

The issue for the Internet was providing a growth path for this
infrastructure, which would support a pervasive, global Internet.
With 4 to 7 million people already on the Internet, how would it
grow to support an order of magnitude more?

It was interesting to look at what types of organizations were
getting plugged into the Internet. Small users were rapidly getting
on to some portion of the network, even if it meant something as
simple as getting an MCI Mail account and learning how to use a
modem and address messages.

Likewise, very large (or very high tech) users were getting
plugged in, linking their campus networks to regional providers.
They had the staff to manage their own networks, extending the
Internet out to the users.

What seemed to be missing altogether, though, was a way for
medium-sized enterprises without a high tech component to get on
the network. Even getting a simple Novell network up and running
was causing more stress than necessary.

I continued my day of meetings and greetings, ending up the
evening having dinner with one of my editors, Stephanie Faul.
Over a dish of crunchy jelly fish salad, we talked about her latest

280

Washington, D.C.

projects in the world of freelance. In addition to her usual stint act-
ing as a ghost editor for the AAA Magazine, Stephanie had just
landed a corporate writing gig, helping to organize a seminar with
the prescient theme “The Future is Tomorrow.”

Tuesday was spent hiding from the cold, doing research in the
Hawk and Dove, one of a half-dozen Irish bars on Capitol Hill. Af-
ter paying a call at the Office of Technology Assessment to brief a
staff member on the Bruno project, I took the metro under the mall
to the offices of the National Science Foundation.

Steve Goldstein, the manager of international programs for the
NSFNET, had invited me to his home for dinner and we would join
his carpool from the NSF. He had just come back from a WAIS
symposium and brought me a set of the conference materials. We
both agreed that WAIS could easily prove to be the Lotus 1-2-3 of
the Internet, providing users with a simple, intuitive way to get real
information.

Steve and his life-partner Isabelle were both members of a gour-
met cooking club and my timing was lucky enough to get the left-
overs. While we waited for Isabelle to get home, Steve heated up a
pork roast stuffed with mushrooms and a dish of fennel bulbs in a
delicate cheese sauce while we talked about the NSF and the
NSENET.

The NSF has taken its charter to help U.S. science and engineer-
ing as a broad mandate which ranges from hands-on science muse-
ums for children all the way to funding research into teraflop
computers.

The NSFNET is a perfect example of the broad interpretation
NSF takes of its mandate. Instead of dedicated networks for super-
computer centers, the agency built a system of general-purpose re-
gional networks connected by a backbone—a true infrastructure.

Steve Goldstein described the NSF funding philosophy as strate-
gically placed “drops of oil,” hoping to leverage comparatively
small expenditures to achieve a broad impact. The drops of oil phi-
losophy was necessary because the agency’s budget was fairly

281

Exploring the Internet

small. Program directors have to husband their money carefully
and overhead is kept down to a paltry four percent.

The international program of NSFNET, for example, makes do
with a budget well under a million dollars per year, a minuscule
amount when you consider how high telephone tariffs are and what
an important role NSF played in building a global Internet.

To spread the money out, NSFNET usually funds half-circuits.
Most international circuits are actually operated by the two PTs
involved, each making up the tariff for the outgoing line. Typically,
NSF only picks up the cost for the U.S. half-circuit, which are often
significantly lower in cost—anywhere from 20 percent to factors of
2, 3, and even 4 times less. In some cases, the international partner
will pick up the cost for both half-circuits. NORDUnet, for example,
paid for the whole line to the U.S. for some time until the NSF was
able to pitch in.

In the case of the “fat pipes,” the high bandwidth circuits over-
seas, there is usually a consortium of agencies that pay. The link to
the U.K., for example, has the NSF, DARPA, and NASA sharing the
costs in one direction and the Joint Network Team and the Ministry
of Defense sharing the other side.

In addition to sharing costs for general-purpose infrastructure,
NSF is able to get a lot for its money because of simple competition.
Sprint, for example, is the manager of international connection serv-
ices for NSF, and won that contract because of a very agressive bid
response. Sprint didn’t do this out of altruism—the experience it
gained with NSF was invaluable in capturing other international
business.

NSF encourages regional solutions as a more efficient way to
connect the world. To Europe, for example, the NSF helps fund the
lines to INRIA in France and NORDUnet in Stockholm, both major
gateways for their respective regions.

The agency also uses line sharing to buy larger pipes. The
NSFNET backbone, for example, does not actually lease telephone
lines. Instead, NSFNET contracts for a T3 service which runs on an
underlying network owned by ANS. That underlying network car-
ries traffic both for NSF and for ANS commercial customers, with
packets from the two sources freely commingling.

282

Washington, D.C.

Line sharing is applied to international links as well. NSF joined
the EBONE consortium as an informal “supporting organization,”
the same category chosen by CERN and IBM. Supporting in this
case meant that NSFNET and EBONE would be connected, but that
NSF would not sign the memorandum of understanding.

Line sharing in EBONE meant that traffic from AlterNet and
NSFNET might share the same lines. Government regulations pro-
hibited NSF from funding commercial traffic, and this was the root
cause of many of the appropriate use restrictions on the Internet.
Allowing traffic from commercial services to share a link with
NSFNET, provided the commercial service paid its own way, was an
important step forward.

Acceptable use policies were set up in the era of the physical
networks when NSF leased 56 kbps lines. The idea was that NSF
should not be subsidizing traffic generated by those outside the
community it meant to serve—the scientific and engineering re-
searchers.

This policy was getting harder and harder to enforce, particu-
larly if the enforcement mechanism was to bar all traffic originating
from certain broad groups from traversing key links. With NSFNET
simply paying for service, it no longer really mattered if packets
fraternized, so long as the educational and research communities
got their money’s worth.

Wednesday morning, I walked a few blocks north to Dupont Circle
to the home of Bob Barad, founder and sole employee of Baobab
Communications. Bob runs the Baobab bulletin board, a Fidonet
node specializing in African development.

His house is full of old maps of Africa, and blankets, baskets,
and masks collected during his two years in the Peace Corps in Si-
erra Leone and on many subsequent visits to the continent. After
his two years in swamp rice development in Sierre Leone, he came
back for a law degree, then returned to Togo on a Fulbright scholar-
ship.

283

Exploring the Internet

Following that he spent a couple of years at the World Bank, an
experience he diplomatically called “an extension of my education.”
Bob started spending his spare time dialing into bulletin boards. He
got an account on PeaceNet and found he could start talking to peo-
ple in Africa. He was hooked.

Soon he had hung out his shingle as a freelance development
consultant, a business that seems to draw heavily on his Fidonet
expertise. He had quickly tired of being a user of other peoples’
boards and wanted to set up his own node.

Bob had been corresponding with a digital nomad named Mike
Jensen, a man who travels the world with a laptop, setting up Fi-
donet nodes. Jensen blew into town one day late in 1989, stayed at
Bob’s house for a day, and Baobab was in business.

Fidonet is a network which can be thought of as the PC equiva-
lent of the UNIX USENET system. It provides mail, file transfer,
and news feeds, all managed by an informal volunteer hierarchy of
city, regional, and national hubs. The free software and the cheap
platforms it runs on have attracted the attention of many human
rights, environmental, and other public interest groups.

Being able to set up a network anyplace there was a PC had
proved valuable in Bob’s business. He had just completed a job
putting in nodes between Chad, Mauritania, and a group in Ar-
lington, Virgina. The Virginia group was a subcontractor on a U.S.
AID program to provide a famine early warning system.

Satellite images of the weather are periodically broadcast by the
European Space Agency. Data would be downloaded and proc-
essed, resulting in graphic images of the affected areas that would
be placed in files totalling roughly 70 kbytes. These files would be
transferred to floppy disks and sent by DHL to Chad and Maurita-
nia for use by regional workers.

This was a situation ripe for automation. By elimination of the
weekly DHL runs, the payback from the use of computers could be
immediately quantified, a situation very different from those trying
to justify the more nebulous benefits such as “increased interaction”
resulting from general-purpose e-mail.

The Fidonet software was successfully installed and the first files
transferred. The system worked so well that one site immediately

284

Washington, D.C.

requested the previous week’s file, which had not yet arrived by
DHL.

The transfer of data uses the zmodem protocol on a direct call
between the African sites and Arlington. Direct polling to the end
destination is not unusual in Fidonet. While it is possible to hand
off messages to a local gateway for percolation through the network,
high priority data usually cuts through the hierarchy. To send mail
to Africa, for example, Bob placed a biweekly call to the London
hub, a site only a hop or two away from the African destinations.

The number of bulletin boards like Baobab in operation all over
the world is truly incredible. Merely listing the boards in ASCII for-
mat makes for a file of 1.44 Mbytes. One of the great challenges of
the Internet will be to give those BBS systems an easy way to inte-
grate into the global network, providing better connectivity than just
mail delivery, but doing so in a way that is cheap and easy.

I returned to my hotel to check out, only to discover that the
window to my car had been broken in the overnight parking ga-
rage. I filled out forms for the parking garage, the police, Hertz,
and American Express in order to get my keys back and sit on a pile
of broken glass, sucking wind on the drive to Dulles Airport.

285

Cleveland

Having never been to Cleveland, I was firmly convinced that I
didn’t want to go. To my surprise, my visit was delightful. My
hotel was a 1910 mansion, renovated in 1988 to be warm and to
support modems and CNN. Dinner was just a stroll down the block
to another renovated mansion, where I made short work of fresh
rainbow trout, grilled with a pungent caper and garlic sauce. The
conversation around me was, not surprising given the proximity to
Case Western Reserve University, full of graduate-student angst and
plans for monumental careers in the “real world.”

Thursday morning, I met Tom Grundner, the founder of the
Cleveland Free-Net and an apostle of populist computing. Grund-
ner became a BBS hacker in 1984 when, as an assistant professor of
family medicine, he was looking for a way to deliver community
health information to the public. His St. Silicon bulletin board was
such a success that, as Tom relates it, “it ate my career.”

St. Silicon blossomed and by 1986 had become the Cleveland
Free-Net, a multi-user service running on a UNIX platform with
functionality similar to CompuServe. Users could chat with each
other, could go through builetin boards, and send mail.

There were, however, some big differences from CompuServe.
For one thing, the service was free to the user. Grundner draws an
analogy to television, where commercial networks fill one niche and
the Public Broadcasting System (PBS) fills another. Grundner sees
the Free-Net movement as the PBS of videotex.

In Cleveland, the network quickly expanded as more and more
volunteers agreed to act as moderators for different topic areas.
Lawyers, doctors, librarians, veterinarians, and many other profes-
sionals quickly joined. Rollerskating and Sci Fi SIGs came into be-
ing, until there were more than 400 different topic areas.

While people were using the system, keeping it going was cost-
ing money. It was kind of hard to claim that this was tenure-track
research for a professor of family medicine, so Tom looked around
for another institutional home.

286

Cleveland

He struck a deal with Case Western Reserve University’s com-
puter group. They were in the process of spending U.S. $10 million
for an all-fiber campus network but, according to Tom, “no one
gave a thought what to put on it.” Cleveland Free-Net became a
university project that also served the community, running on five
Sun servers maintained by the department.

Tom got a salary, a couple of big offices, and a place in the bu-
reaucracy far away from daily control of the Free-Net. Of course, he
still had the ultimate control weapon, the ability to cause a loud fuss
if things went too far off track.

So what to do? In 1989, he started the National Public Telecom-
puting Network (NPTIN) and started pushing the Free-Net concept
in other cities and countries. It was to the offices of the NPTN that
Tom had brought me.

Both rooms were jammed with shelves, tables, postage meters,
and ratty old couches, making these offices no different from count-
less public interest groups all over the country. This was obviously
the kind of place where lots of volunteers spent their evenings.

It was morning, though, and the place was deserted. I sat down
next to a life-size balloon with a skeleton painted on the front.

“Meet my staff,” Tom said wryly, while he flitted around the
room checking Macintoshes. He lit a cigarette and sat down in an
easy chair to tell me about his populist computing movement.

The model for a Free-Net was based around a local organizing
committee. This group of volunteers would organize and run the
Free-Net, raising money wherever they could for a multi-user sys-
tem, a copy of UNIX, and some modems and phone lines. The com-
mittee would then get on its collective knees and plead with the
nearest college with a router for Internet access.

Software to run the Free-Net comes from Case Western Reserve
in Cleveland. Long available for only U.S. $1, the University had
gotten stars in its eyes and raised the price to U.S. $850 for non-
profits, and much more to corporations.

In addition to the software, Tom’s NPTN group provides what
he had dubbed “cybercasting services.” His first cybercasting suc-
cess was convincing USA Today to allow him to distribute at low

287

Exploring the Internet

cost an online, electronic version of the newspaper. USA Today got
supplemented with a variety of other news feeds.

When the Supreme Court started its Hermes project to distribute
opinions electronically, it selected 12 groups ranging from the Free-
Net and UUnet to UPI, Reuters, and Mead Data. Other information
on current events (“teledemocracy” in Free-Netese) comes from the
Congressional Memory Project. Every week, NPTN takes three Sen-
ate bills and three House bills and types in a two- to three-para-
graph summary of the bill and how officials voted. The database
builds every week, allowing people to start scanning the voting re-
cord of incumbents during reelection campaigns.

Cleveland Free-Net has grown over the years to 40,000 regis-
tered users, with 1,000 to 1,500 unique logins per day. Other Free-
Nets are operational in Cleveland, Cincinnati, Peoria, and
Youngstown. There is even a rural Free-Net in Medina County,
southwest of Cleveland. An old, donated IBM RT running AIX is
maintained at a 100-bed community hospital in the county seat. The
local agricultural agent is on the system, as are librarians and other
professionals in town.

Medina County is typical of most midwestern rural counties. It
is perfectly square and the county seat is in the middle of the
square. A call from anyplace in the county to the seat is considered
a local call. Farmers, practically all of whom have PCs to run their
farms, have a strong incentive to avoid long rides into town.
Grundner felt that Medina County could serve as a model for the
entire Midwest.

Financing Free-Nets is an interesting proposition. In Cincinnati,
Cincinnati Bell sponsors the system. Many of the other sites operate
on a shoestring. Grundner offers two financing models to the local
organizational committees. On the pay-as-you-go model, the local
Free-Net pays for services. A USA Today feed, for example, is based
on the number of lines and will cost a typical Free-Net U.S. $1,000 to
$2,000.

The other model has the Free-Net gather the names and ad-
dresses of users and ship them off to Grundner. Grundner, in turn,
sends out tear-stained letters pleading for contributions. Of the five

288

Cleveland

existing Free-Nets and the five others about to go into existence,
three opted for the pay-as-you-go plan.

Whenever Tom could get a plane ticket abroad, he tried to
spread the Free-Net concept to other countries. Often, though, he
spreads the Free-Net concept by electronic mail, as in the case with
his correspondance to the Helsinki Free-Net or to Richard Naylor in
Helsinki.

Grundner was even invited to give a seminar in Singapore,
sponsored by the National Computer Board. I found this a bit
strange as Singapore’s government, and especially the NCB, is
pretty much the antithesis of the populist ideas Grundner repre-
sents. Interestingly, Grundner was invited back to Singapore, this
time not by the NCB but by the Singapore Microcomputer Society.

The aspect of NPTN that was the most fascinating was Academy
One, a K-12 computer literacy project based at Cleveland Free-Net
that uses the Internet to try and include kids from all over the
world.

Academy One sponsors periodic special events, such as the
TeleOlympics. Any school with an Apple II (or any other cheap sys-
tem), a modem, and Telnet prompt could participate in the
TeleOlympics.

Four events were picked, including a 50 meter dash, standing
and running broadjumps, and a tennis ball throw. The kids went to
the school yard and ran the events, then came back in and posted
their results.

The day started in New Zealand, then moved on to California,
Cleveland, and ended up in Finland. The Free-Net kept a leader
board of individual results, class averages, and a running news
commentary with late breaking bulletins like “Mrs. Jones and her
4th Grade Class Break into the Lead.”

A more sophisticated event was the space shuttle simulation.
University School in Shaker Heights had a full scale mockup of the
interior of the space shuttle, courtesy of a NASA program, complete
with tape loops of the Earth from space running on TVs located
behind the portholes.

With children’s games, the kid with the toy gets to name the
game, so University School acted as mission control, running 24-

289

Exploring the Internet

hour simulations of a shuttle run. Other schools got to play other
roles.

A school in California was selected as the alternate landing site.
An alternate landing site has to furnish weather data to mission con-
trol, so the kids would go out and gather data on temperature, hu-
midity, and barometric pressure. The data would go to mission
control, which would respond with the current shuttle status.
Status information was used to plot the current location on a map.

For all the Academy One events (and for Free-Net as a whole),
the focus is on accessibility. This means cheap and easy to use.
Tom looks with great skepticism at things like gigabit testbeds, pre-
ferring instead to convince a librarian to start an online book club.

I dropped off my rental car and went to the airport restaurant
for a nondescript meal, spiced up by an article in the New Yorker
about a group of Tibetan refugees in Scotland. In a strange form of
cross-cultural exchange, the Scottish haggis would be used as a base
for the traditional Tibetan momo dumplings. Poems by Robert
Burns and the Dalai Lama were read at these affairs. One of the
Tibetans had tactfully suggested to his countrymen that the momos
were even more delicious when doused with liberal quantities of
hot sauce.

290

Chicago

With great relief I arrived at O’Hare airport, knowing that my next
plane ride would bring me home. I drove out the I-88 toll road,
past AT&T’s Bell Laboratories, to the little town of Warrenville.
When I first moved to Warrenville in 1969, the town of a few thou-
sand people was in the middle of farm country.

Robert Wilson, the legendary high energy physicist had gathered
many of the best experimental physicists in the world to 6,800 acres
of prairie to make the largest accelerator ever built. Over the years,
the accelerators they constructed became capable of smashing pro-
tons against antiprotons at the incredible energy of 1.8 Trillion Elec-
tron Volts (TEV).

Collisions of subatomic particles at high energies help reveal in-
formation about quarks and leptons, the fundamental forms of mat-
ter. Acting like a giant microscope, the Fermilab accelerator used
higher energies to create and discover particles unobservable at
lower energies.

In the mid-1970s, the accelerator provided evidence of the exist-
ence of the upsilon, a subatomic particle composed of a bottom
quark and an anti-bottom quark, a discovery that provided strong
support for the Standard Model. In the 1990s, the lab was hot on
the trail of the top quark, the last quark yet to be observed experi-
mentally.

When I moved to Warrenville, though, there was nothing but
prairie where Fermilab now stands. My father and the other physi-
cists did their work in a suburban office complex until they could
start using houses in a subdevelopment on the site that had been
purchased under eminent domain. Old farmhouses from the coun-
tryside were moved together into a cluster, to house visiting scien-
tists. The subdevelopment was painted bright primary colors, and a
chef was brought in to staff the village cafeteria (although the food
was lousy, at least there was a chef).

At that time, Fermilab was just beginning to coalesce. While
everybody slaved away building the accelerator, Wilson continued

291

Exploring the Internet

to put in those unique features that made the lab such a special
place. A herd of buffalo was brought in and a project was started in
the middle of the 4-mile circumference main ring to restore it to the
original prairie grasses. One of the lab buildings was even con-
structed as a geodesic dome, with the panels of the dome made out
of panes of plastic, sandwiching thousands and thousands of alumi-
num cans, collected in a recycling program by the neighboring
schools.

My visit to Fermilab on Friday had been set up over the past
two months via electronic mail. Around Singapore I had received a
message from a lab public relations engineer asking if we might
make my visit a two-way exchange of information.

Of course, I wrote back, two-way is good. By the time I hit Am-
sterdam, I had another message from the lab wanting to know if I
would be discussing the future of networking on my visit. With
another 100 messages to plow through, I dashed a hurried (but puz-
Zled) acknowledgment.

Arriving at the lab, I saw that the little two-way exchange had
blossomed into a full-blown public seminar with the title “The Fu-
ture of Networking.” A nice topic, no doubt, but one best left to
astrologers and marketing analysts.

I asked around to see if perhaps some physicist had spoken on
“The Future of Cosmology” in a recent seminar, and perhaps I was
just part of some series. No such luck, so I took advantage of the
rule that states the one with the white board gets to set the agenda
and changed my topic to a more manageable one.

After my talk, I sat down with some of the Fermilab networking
staff, including the managers of the U.S. High Energy Physics net-
work (HEPnet). HEPnet ties physicists in dozens of countries to the
major world physics laboratories, such as Fermilab, CERN in Ge-
neva, NIKHEF in Holland, KEK in Japan, and many, many, other
sites.

HEPnet started as a physical network, a microwave link between
Lawrence Berkeley Laboratory and the Stanford Linear Accelerator
Center (SLAC). It was based on DECnet protocols, a natural choice
considering the overwhelming use of PDP and VAX computers in
the physics community.

292

Chicago

The network quickly grew, using the big laboratories as hubs to
run tail circuits to regional (and sometimes not so regional) research
universities. CERN and Fermilab were two of the more aggressive
hubs, bringing some of the first network lines into places like China,
South America, and Eastern Europe.

Connecting the hubs together were the lines that made up the
HEPnet backbone. While some of the tail sites were strictly DECnet,
many of the backbone sites began using multiprotocol routers, let-
ting HEPnet finance a part of the line and having the other user
communities fund the rest.

Many of the universities at the ends of the tail circuits began
joining the Internet, and the HEPnet lines became one of many pos-
sible paths, or often just a virtual network layered on some other
service provider.

Gradually, HEPnet grew from a single protocol network to a
multiprotocol one, and from a physical network to a shared line use
model. The shared line use just made HEPnet a convenient label for
the bundle of resources that the high energy physics community
was purchasing for its users.

In the U.S., HEPnet was partially swallowed up by a larger net-
work run by the Department of Energy, the Energy Sciences net-
work (ESnet). Rather than lease many lines between labs for
individual communities, ESnet would furnish the backbone.

Although HEPnet relinquished the role of backbone, it still fur-
nished quite a few tail sites, particularly medium-speed links to
places that ESnet did not serve, such as Brazil, or those sites that
need special high-speed dedicated capacities, such as the physics
departments of many large universities. In addition to managing
the dedicated links, the HEPnet group was trying to move towards
providing higher level services, such as running a mail gateway and
starting video conferences.

In addition to the HEPnet WAN links, the Fermilab Computer
Division had to manage a vast array of internal resources, including
what was reputed to be the largest VAX Cluster in the world.
Workstations are a dime a dozen, and large Amdahls provide local
mainframe support.

295

Exploring the Internet

The network architecture at Fermilab was originally, as can be
expected by the DECophilic physicists, set up as a huge extended
Ethernet running protocols like DECnet and LAT. Over time,
though, UNIX-based workstations made TCP/IP an important part
of the network.

Routers were introduced to segment the traffic into subnet-
works. With some of the Local Area VAX Clusters (LAVCs) having
as many as 60 workstations as clients, segmenting the network cer-
tainly made management easier.

The biggest concern, though, was how to continue migrating to
higher and higher bandwidth LANs. Fermilab is the kind of place
that will always be using the latest equipment in all sorts of custom
configurations. To accommodate FDDI and any future LANs, the
laboratory laid bundles of 48 single and multimode fibers around
the ring.

All that fiber around the lab meant that in addition to one or
more lab backbones, the Computer Division could take a fiber to an
experiment or workgroup, thereby creating rings on demand. Like
many sites, Fermilab was trying hard to go through the exercise of
laying fiber in a large area only once.

After a quick lunch, I went into the building housing the Linear
Accelerator (LINAC), the first of the seven accelerators linked to-
gether at the laboratory. Behind racks and racks of electronic equip-
ment, I found the office of Peter Lucas, one of the designers of the
accelerator control system.

Controlling a series of accelerators involves monitoring and set-
ting a huge number of devices on subsystems ranging from vacuum
pumps to refrigeration units to superconducting magnets. In the
case of Fermilab, the accelerator control system must handle over
45,000 data points, some of which operate at very high frequencies.

Most devices are managed by CAMAC crates, a venerable back-
plane architecture that can hold up to 22 cards per crate. CAMAC
crates are controlled by serial links, which in turn plug into a com-
puter, usually a PDP 11.

294

Chicago

The computer can poll each of the cards on the crate to gather
and set data. If a card needs to generate an interrupt, it can set a
single “look at me” bit on the crate. The computer then polls each
of the cards to find out which ones need attention. The basic con-
trol flow on the front end was thus a PDP 11 to a CAMAC crate to a
card to the device being controlled (although the PDPs were sched-
uled to be swapped out at some point for »VAXen).

In the early 1980s, when this system was first being put together,
VAXen were just being released. A series of 3 11/785s, later supple-
mented by an 8650, acted as database servers, keeping track of in-
formation on each of the data points, such as which CAMAC and
PDP that data point was accessible from.

The last part of the system were pVAXen (originally PDPs) act-
ing as consoles. The basic operation would be for a console to log a
request for a certain data element at a certain frequency. The data-
base server would define the elements and the PDP 11s were re-
sponsible for delivering the data.

So how to connect all these systems together? In 1986, when the
current network was being designed, the great Ethernet versus To-
ken Ring debate was raging. After extensive deliberations, the lab
decided the predictable latency of the token ring was needed for the
main ring control system.

DEC, of course, was firmly in the Ethernet camp, but lack of
product support by the vendor doesn’t faze experimental physicists.
They turned to their electrical engineers, who built token ring cards
for the Unibus used by the PDPs and VAXen.

Running on top of the token ring, the lab put their own Accel-
erator Control network (ACnet). ACnet allows any console to moni-
tor and manage any device, a capability that makes operators at
other accelerator centers stuck with special purpose consoles green
with envy.

ACnet is a protocol in which a console logs a single request,
which can generate a stream of data. A request might be for a cer-
tain data item at 1 hz, which would yield a stream of data spaced
one per second. In addition to logged requests, certain events trig-
ger alarms, sending the front-end PDP into the database to decide
who should be notified.

295

Exploring the Internet

The token ring choice worked, but it turned out that there really
wasn’t that much difference between it and Ethernet. Not only that,
DEC kept coming out with new pVAXen, each sporting a newly in-
compatible variant on the Q-bus. It was evident that Ethernet had
to be incorporated into the network somehow.

An Ethernet was added and VAXstations started to pop up
around the laboratory as consoles. To link the Ethernet with ACnet,
the lab designed and wrote an Ethernet-to-token ring bridge run-
ning in software on one or more of the VAXen.

The ACnet was still token ring, but the addition of the bridge
meant that authorized users from all over the lab had access to ac-
celerator data. One of the VAXen was even set up with the X Win-
dow System, allowing X servers to access the data.

Since X runs just fine on top of the Internet, especially if you are
communicating between two core sites, it is even possible to run the
accelerator remotely. On one occasion, lab employees at a confer-
ence in Japan pulled up the accelerator control panels. To prevent
any possibility of accelerator hackers, the lab set up the system so
that remote terminals could only read data items.

This architecture proved to be quite flexible, easily expanding to
meet the needs of the Tevatron and other newer subsystems. PDPs
controlling CAMAC crates could easily be added if new data points
needed to be controlled. Over time, CAMAC crates became a bit
old-fashioned and newer devices simply had a VME crate directly
on the token ring, alleviating the need for the front-end PDP.

Even the token ring decision turned out to have been wisely
made. Networking at the lab grew like a prairie fire, with worksta-
tions going in as fast as physicists could unpack them. This is not a
user community that will wait for official installers to unpack boxes.
In fact, if pieces are missing, they are just as liable to manufacture a
replacement themselves as they are to call up the shipper and com-
plain.

After walking me through the architecture, Peter Lucas took me
over to an X Window System console and showed me some of the
1,000 application packages that had been written for the accelerator.

All the screens were color coded, providing visual clues on
which items in a complex form led to other screens and which items

296

Chicago

could be changed by the user. Clicking on a line item with the ID
for a CAMAC crate, for example, would result in a graphic repre-
sentation of the map, with each card and its ID depicted. Clicking
on one of the cards would show the data items kept by the card and
its current status.

All these programs were written at first for PDPs, in a day when
that level of user interface was real, heavy-duty programming.
These were the kind of programmers who battled their way through
RSX, FORTRAN 2, and TECO, languages that make me shudder
whenever I think of them.

When the superconducting magnets in the main ring generate pro-
ton-antiproton collisions at 1.8 TEV, there needs to be a way to col-
lect data describing the event and store it away for further analysis.
The detectors used in high energy physics are as complicated as the
accelerator itself.

I drove around the main ring to D Zero (D@), the quadrant of
the ring containing a detector that had been under construction for
over nine years. The instrument had consumed most of the time of
400 physicists and the U.S. $50 million camera was almost ready to
be inserted into the ring.

D@© was a six-story building, just large enough to house the de-
tector. When it was operational, the detector would put out 100,000
data points per collision, and there would be a collision every 3.5
microseconds. With roughly 250 kbytes per collision, this results in
a data rate of 75 Gbytes per second.

This is quite a bit of data, especially when you consider the fact
that the experiment would run 24 hours per day for six to nine
months at a time. One of the great hopes of the DO detector was
that it would provide experimental evidence of the existence of the
top quark, the only one not yet empirically observed.

Top quarks are very rare, and it was quite likely that only one
(or less) would be found per day. In a day of operation, the data
acquisition subsystem on the detector thus had to somehow filter

297

Exploring the Internet

through 6.48 petabytes (6.48 million gigabytes) looking for the inter-
esting 250,000 bytes.

D@ was pandemonium. With the pins being pulled in a few
days to roll the detector in, desks were littered with piles of empty
styrofoam cups and coke cans, boards were everywhere in various
stages of assembly or disassembly. A pager went off every few min-
utes, directing people to where they weren't.

I found Dean Schaumberger, a professor from State University of
New York at Stony Brook who agreed to give me a quick briefing.
We walked through the data acquisition system, concentrating on
the data and ignoring the physics. To decide which events to keep,
an event mask provided the first level of filtering, allowing the
physicists to hone in on a few dozen attributes that characterize in-
teresting events. Once the event mask was tripped, boards housed
in 79 VME-based front end crates would digitize the data on the
detector. The goal of the mask was to reduce the data flow to 200 to
400 events per second.

Since it was possible that two interesting events could occur
back to back, the front-end boards were double buffered. After the
two buffers were full, though, events that tripped the mask had to
be discarded until space became available again.

The 79 front-end crates were distributed among eight data ca-
bles, each cable providing a 32-bit wide data path running at 10
Mhz, yielding a data rate of 40 Mbytes per cable. With all eight
cables running in parallel, the data flow reached 320 Mbytes per
second.

The next level of analysis was provided by a farm of 50 pVAX
4000s, each machine running at the rate of about 12 times a VAX
11/780. Each pVAX had 8 custom boards, one for each of the 8 data
cables. The data cables thus came out of the VME Crates and
snaked through each of the pVAXen allowing any machine to collect
all the data on one event.

The pVAXen, running the ELN real-time operating system, were
also connected to a scheduling VAX, a machine that decided where
on the farm there were available cycles to process the next event.

With the major pieces in place, Dean walked me through the
data path. When the mask got tripped, data would go into the

298

Chicago

front-end boards. If the data bus was clear, the token generator (an-
other p VAX at the end of the data bus) would issue a token with the
current event number.

Each of the eight data cables would get a token. On any one of
those cables, if a VME crate saw a token, it was able to capture the
token and send as many bursts of data as it needed, releasing the
token downstream when it was done to allow the next crate to
dump its data. When all eight tokens got back to the token sched-
uler, the next event could proceed.

When the data reached a np VAX on the farm, it would undergo
scrutiny by a variety of programs designed to filter interesting
events from the subatomic chaff. The goal at this stage was to win-
now down the data rate another 200 to 400 times.

Events that met the second level of filtering would emerge from
the p VAX farm and go to a VAX 6000 to be spooled to tape. I saw a
rack with 11 Exabyte tape drives on it, and Dean informed me that
the experiment had well over 50.

This was not excessive when you consider that one event per
second for six months would be close to 4 terabytes of data. Even-
tually, a physicist would spend up to a half hour with each interest-
ing event, looking for clues. Some events that looked particularly
promising were also spooled on the VAX 6000 for immediate analy-
sis during the run.

The next morning, I went with my father to see SciTech, a hands-on
science museum for kids he had spent the last two years building.
SciTech had managed to obtain a cavernous old post office in the
middle of nearby Aurora.

The museum specialized in the very big and the very small,
teaching kids about light, sound, astronomy, and even particle phys-
ics. Every exhibit was hands-on. To explain quarks, for example,
SciTech was modifying slot machines, replacing lemons and cherries
with rare quarks and adjusting the probabilities to match. Pulling
the lever to create a particularly rare particle would result in a prize.

299

Exploring the Internet

All around were prisms, tuning forks, echo chambers, comput-
ers, and dozens of other devices making up the exhibits. Signs were
next to each exhibit and teenage “explainers” would walk around
telling people what they were seeing. A group of girl scouts trun-
dled down to the basement to help build a new exhibit. Upstairs,
some six-year olds were playing in a science club.

Starting a museum is never easy, and SciTech had managed to
surprise everybody with its rapid growth. Still run on a shoestring,
the museum was able to get enough grants, admissions, and mem-
bers to keep expanding the exhibits and start filling up the huge old
post office.

I admired the exhibits, then ran to catch my plane. O’Hare was
not as bad as usual, so I had a little time to kill. Walking down the
moving walkway, I came to a stop behind a little girl and her
mother.

“Move over, honey,” the mother said, “this is an airport and the
nice man has to hurry and catch his plane.”

The little girl smiled sweetly and moved over, just in time for
me to get off the ramp and head into the nearby cocktail lounge.

300

Boulder

On the flight back to Boulder, I struck up a conversation with the
prosperous looking yuppie on my left. He worked for Trammel
Crow, the quintessential Texas real estate developer. We talked
about the current fiscal performance of Compag, his Vail vacations,
and the relative profit potential of commercial and residential devel-
opment in the greater Houston area.

Over a second drink, I asked this seemingly normal modern
businessman what sort of activity brought him to Boulder. A
merger, perhaps or an acquisition? Dividing virgin forest into poor
imitations of Southern California?

“T’'ve quit my job to open a Great Harvest outlet in Madison,
Wisconsin, ” he explained, referring to the popular whole wheat and
granola bakery located in a Boulder shopping mall.

Back in Boulder, I brought a box of Parisian chocolates over to
my neighbor, a freelance wholesale carpet salesman and the proud
owner of at least a dozen vehicles. The vehicles gave the place a bit
of a hill-country look, but I didn’t mind since it made it harder to
tell that my house was deserted most of the time.

“Do you ever suffer from writers block?” he wanted to know.
“How about jet lag?”

“Um, sort of,” I said with a lack of conviction, pretty sure this
wasn’t just an idle question.

“Here,” he said, handing me several packets of powders. “I’ve
added a new line,” pointing proudly to the pouches of Wow!™ and
Focus!™ (“Nutrients for the Brain”), products of Durk Pearson &
Sandy Shaw®, the founders of Omnitrition International and the in-
ventors of Designer Foods™.,

Pearson and Shaw have built quite a nice little business taking
the hopes we all share for intelligence and vigor and peddling them
nationwide in a hierachical sales scheme of life extension that has
taken in such world-famous businessmen as Jerry Rubin. I thanked
my neighbor and told him I'd be in touch, remembering the advice

301

Exploring the Internet

my mother gave as I left home to never fight with the woman who
feeds your cat when you are going away soon.

Soon, in fact, was an understatement. I had less than two weeks
to dump the text I had accumulated, sort out my finances, and get
back on the plane for my third round-the-world. First on my list
was to bring my primitive PC in for a new disk drive, the cause of
its failing to function as my mail system for the last few weeks.

Next on the list was responding to a couple of hundred mes-
sages from people trying to understand why they couldn’t get ITU
standards anymore. Bruno was dead and engineers all over the
world wanted to know all the grisly details. Most agreed that the
situation was ridiculous and couldn’t understand why the ITU
would reverse such an obviously useful service.

Most touching was a message from a blind engineer in Australia
who had hoped that he could work with the ASCII text, piping it
out to /dev/audio so he could hear the texts. The message re-
minded me of Milan Sterba, who told me about the pitiful number
of copies of the standards that existed in Czechoslovakia. I for-
warded both bits of information to Dr. Tarjanne at the ITU, inquir-
ing if it was truly the policy of a major UN agency to deny access to
the physically handicapped and those in developing countries. A
cheap shot perhaps, but that’s politics.

In the middle of the week, while trying to come up with 12
pithy columns for Communications Week, I checked my mail to see an
urgent note from Mike Schwartz. David Farber, a grand old man of
networking, was giving a lecture in an hour and a half and nobody
in Computer Science seemed to have been aware of the event.

When Mike found out from out-of-band sources he promptly
broadcast a message to those of us who would realize that this was
a worthwhile way to spend an afternoon, and I promptly rear-
ranged my schedule.

David Farber had actually been on my mind that morning. One
problem with trying to write a technical travelogue is that you can’t
possibly see everything. I wasn’t going anywhere in Africa, Central
America, or South America, for example. I wasn’t going to any of
the supercomputer centers. I wasn’t even going to visit Merit, the
network operations center for the NSFNET.

302

Boulder

And, I couldn’t figure out how to squeeze in visits to try and
meet people like Ira Fuchs and David Farber, people instrumental in
a host of key projects ranging from ARPANET to CSNET to BITNET
to the NSFNET.

It was my luck that Boulder is close to enough ski areas to at-
tract frequent guest lecturers, including David Farber. When I got
to the lecture hall, it appeared that the computer science department
wasn’t as awed as I was. Only Mike Schwartz came from the de-
partment, the rest of the professors being unable to climb down
from their towers in time.

Farber started by explaining how the gigabit testbeds came to
be. People like Farber, Gordon Bell (the father of the VAX) and Bob
Kahn from CNRI had spent several years working to get the project
funded by NSF and DARPA. The genesis for the testbeds came
when researchers started asking what would come after Broadband
ISDN and ATM cell switching.

In particular, the researchers wanted to know what applications
would run on these testbeds. Visits to places like Bellcore and Bell
Labs didn’t yield any answers: the telephone companies didn’t see a
market for gigabit networks and weren't paying attention to them
yet.

Even if gigabit networks existed, the researchers saw a funda-
mental problem: the current technology would not evolve gracefully
into gigabit or multi-gigabit speeds.

Switching technology, for example, had progressed up an evolu-
tionary chain from T1 to T3 to the 622 Mbps speeds of Broadband
ISDN. These switches had become increasingly complex, and many
felt they had reached the edge of the technology. What DARPA
called a “paradigm shift” would be needed to move to the next
stage.

Government leadership was obviously in order here, and Kahn
and Farber made a proposal to NSF for the gigabit testbeds. The
proposal had a few unusual features to it. For one, the testbeds
would depend extensively on participation from industry, and in-
dustry would have to pay its own way.

Money from NSF and DARPA would go to pay for the academ-
ics and program overhead, and this leverage would be used to bring

303

Exploring the Internet

in industrial participants. To prove that industry would indeed par-
ticipate, Farber and Kahn produced letters from IBM, DEC, Xerox,
Bell Labs, and Bellcore.

The other unusual twist to the proposal was that although gov-
ernment would start the program, the operation of the project
should be left in the hands of what Farber called “an organization
with a faster metabolism.” Ultimately, this management group
ended up being CNRI.

After a series of white papers, proposals, and intensive lobbying,
the project was set up with U.S. $15 million in funding. Bob Kahn
hit the road and used that seed money to raise another U.S. $100
million in facilities and staff commitments from industry.

All the resources were put into a pot and reorganized into the
five gigabit testbed projects. In some cases, additional participants
(particular telephone companies to provide fiber) were added to
round out a testbed.

Most of the testbeds were set up to look at a combination of the
underlying technology and applications. The CASA project in Cali-
fornia and New Mexico, for example, would run three test applica-
tions over wide-area gigabit channels. |

The Aurora project, the testbed that Farber had joined, was a bit
different, concentrating on the underlying technology. The chief
participants were a group from MIT led by David Clark, Farber’s
group at the University of Pennsylvania, and teams from Bellcore
and IBM. Several telephone companies were also participating to
provide up to 2.4 Gbps of wide-area bandwidth.

Typical of the research flavor of Aurora was the work being
done in the switching substrate. Bellcore would look at cell-based
switching using ATM technology. IBM would run an entirely differ-
ent datagram substrate based on their PARIS project.

Farber and Clark were looking at the upper levels, the protocols
that ran on the fat pipes. Both felt that current protocols, such as
the TCP/IP suite, might not work on the next generation of net-
works.

Most protocols were based on many modules that exchanged
messages, with protocols layered on other protocols. While this was

304

Boulder

architecturally convenient, the result was that software layers would
endlessly transform data, copying it from one buffer to another.

The immediate problem was thus not on the network. A gigabit
was more than most hosts could handle. The problem was not only
in the CPU or the programming model, either. The problem was
the network software. Clark was advocating application-layer fram-
ing, collapsing the protocol stack and giving the application much
better control over network operation. He saw many of the flaws in
today’s networks as the result of modules making decisions in areas
such as flow control that didn’t make sense in the context of the
upper layer user.

Farber was proposing an equally radical solution, doing away
with the message abstraction of current protocols and moving to-
wards a totally different way of looking at the network. Instead of
messages, Farber wanted to look at the network as a memory mov-
ing mechanism, a backplane.

Programmers on a single host, or even a multiprocessor, deal
with memory as the main abstraction. Calling a subroutine by
name, for example, is a simple way to refer to a segment of memory.
If the network implements that same abstraction, programmers can
continue to work within the model they were used to and the net-
work simply becomes an extension of that model.

With its roots in an early research project conducted by Farber
and his student Paul Mockapetris, the concept first really took shape
in a 1985 Ph.D. thesis by Gary Delp, another one of Farber’s stu-
dents. The system was called Memnet, and consisted of a bunch of
PC ATs connected to a modified memory interface and a cache, con-
nected in turn to a 600 Mbps token ring.

The LAN acted as a way for requesting, moving, and writing
segments of memory. The caches on other hosts could keep copies
of segments, alleviating the need for a network operation on fre-
quently accessed pages.

Memnet demonstrated that a LAN could indeed be a simple ex-
tension of memory. It operated within the flat address space of DOS
and enforced tight consistency guarantees on pages of memory
through mechanisms such as only permitting one writer per seg-
ment.

305

Exploring the Internet

By 1988, the Memnet concept had been moved up to software,
running on an Ethernet LAN. A modified version of Sun’s UNIX,
developed by Ron Minnich, another Farber student, allowed several
hosts to share memory. The software was actually put into produc-
tion use at a supercomputer center, being a useful mechanism to
provide parallel processing on certain classes of applications.

It was clear to Farber and his students that Memnet was one
example of a general class of networks. To extend Memnet to a
wide area environment, some changes would have to be made. A
ring topology or a simple bus was not appropriate, and constraints
like the flat address space would certainly have to be relaxed. Most
importantly, looser consistency guarantees would be needed if the
system would successfully scale to become a “national memory
backplane.”

Farber and his students were actively working on an implemen-
tation, called GobNet, that would eventually work on the Aurora
testbed. On a gigabit testbed, the software certainly showed prom-
ise. The latency of a 3,000-mile gigabit network is roughly equiva-
lent to a local page fault on a machine, meaning that the network
could appear to the host as the equivalent of a memory segment
that had been paged to disk. |

While page faults were natural on a host, they were also some-
thing to be avoided. Virtual memory systems would preemptively
cache memory segments to try and have the proper segment waiting
when it was asked for. This was also going to be the key to good
performance on a wide-area Memnet.

The switches in GobNet would have the caches. If a host re-
quested a segment not in the cache, the cache would have to go and
find the requested object. In a token ring or Ethernet, finding an
object is simple: you ask for it.

In a WAN, though, the topology may be quite complex and find-
ing the correct path is the real challenge. GobNet would use a
flooding algorithm, similar to what a bridge would use in trying to
decide where a host lay on an extended Ethernet.

For the first request on a segment, the request would be flooded,
being sent out every available path on the network. Eventually, one
hopes, the page would come back on a certain path. Each of the

306

Boulder

switches on that path would set up a segment to host binding so
that future references to the segment would be easier to find.

Flooding a network to find an object and then setting up a bind-
ing for future references turned out to be a strategy that was good
for more than just memory segments. Researchers at Bellcore were
using the same technique in personal communications systems to
find the current location of a portable phone anywhere in the coun-
try.

The view of the Internet as a national memory backplane was
exciting, but what impressed me the most was the overall scope of
the research being conducted in the gigabit testbeds. It was clear
that this was money well spent and that we would all benefit from
the incredible collection of talent that had been assembled.

After the lecture, with my column deadlines looming, I raced to
the Trident, a coffee shop and bookstore, to try and find some pithy
sayings to write. Munching a Blueberry Tahini bar, I walked around
the bookstore hoping to get inspiration.

I picked up a copy of The Secret Books of the Egyptian Gnostics but
couldn’t find any parallels to my current subject. When The Aro-
matherapy Workbook didn’t prove to be any help, I sat down and
tried to grind out words of MIS wisdom while around me people
layed out tarot cards and discussed the many paths to personal ful-
fillment.

307

example ol a qe iol class aw ryt
ah: saniplalgd sap ane eat
ie so opg lle¥o of awe pont sd_ont aft
sone aainith oatediey Sidewia au tah ety nee.
reps, shane y Lhe, blac, AY AeAR bows jaca, Hoye. ys
e car noistegees come had ln Mah "9 oa oak
of, ‘basen Lanciie 4 esctifk eb Fon yor Min sun a A
itiqrganos boll bas. 108 ol ooiedord bias FOUR. infaph t a i
hawete ba PBW (hy, ae a L MUG sordald. & ape ans uit a as NS
toed, Chr abt testhe -aovenigen: ts Vag AEM a ° eae
Sec chsiignn La st mega ott pa ae S392; I t to a ah ae
“tite cad ndanlcyterrH> 0. Bt Taller, XR P
SUB ey elds, ee : Ly ghedy WARE: a PE. YN Pale rok prAores CGY"
siaiap ‘3 ~ ae KUOT abiivy: guobalw elm io abuow 10 aoe af hat
“htt ban Qn296,, oA. TREK, wage ot bewepransts bas it ioxal Yok > £

rae OT) were and ar.
ting to, be awoke Virtus) sistigs #9 nyatem wend preemp pt we ly”
Cache-mermwiry 3 Cyrene i iy ana have U 2 proper resmnerit walt
VV ) etl rf wer asked fon. ‘This was alue. ROIs 10 _ the «Ky, a
perutonarnt on 8 Wwidewra Mimuet. ‘ME 7
te fwiiches In Goole: wosdd. have the caches, i ay ae
quested a segment notin the cache, the cache would Hianve teen
find the requested ehjant diy a token, Ting, oF Echemet, “findis ee
oblgot is siniples you wk fordt |. gi ome 9," Ti
In a WAIN, theugh, the apology may be parton ace
ing the correct path tw the teal challenge. “Gobet,
fooding alyoritua, similar ty what 0 bedge ch visi | ne |
sich Ses See a pel shear _

EA)

nal
4

2 qaaet Snenc
mn of ante Ly Claes

Round 3

In Search of a Standards Haven

ei? |) ae! Ray

Pier a am ce

ty 7 rr | pe Le

r of le ee Au a pei Wa,
} dai al ay au ilk sn vi ie “ q
f ny W

jr OK: i . inte Ming ttt se ie
rf eae eben ths

ee ee) nu | | Unn ites js : ve Se a
_ | im ey) Wy . ‘4 iat sm py.
eat sentry aa .

can . in os Gone i ua en Pt

A ray i + ais iD /

‘aay wel debate 5 a8 foes? a nt

Pre

ts Ne
7 rae :

: an = ‘

erry ar (iis rales
an iy vO ee mh ne Pian

i ; i uy ue pe i ) 2 a f

Marina del Rey

Sitting by the lagoon in Marina del Rey, I sipped a Pilsner Urquell
and watched the planes at LAX, two taking off followed by two
landing in an infinite repetition. In addition to the world’s largest
man-made small boat harbor, Marina del Rey is the home of the
Information Sciences Institute (ISI), a research group that has played
a key role in the development of the Internet.

Monday morning, I got in my car to find ISI. Since the Institute
is part of the University of Southern California, I expected to arrive
at some ramshackle house on the edge of campus, but instead found
myself on the 11th floor of a modern office building, right on the
edge of the harbor.

With 200 staff members and most of its funding from the De-
fense Advanced Projects Research Agency (DARPA) and other gov-
ernment agencies, ISI plays an important catalyst role in several
areas. The MOSIS service, for example, is a way for researchers to
get small quantities of custom VLSI circuits. MOSIS gangs together
designs from many researchers onto one mask, breaking fabrication
costs down to manageable levels.

My goal, though, was not to learn about MOSIS but to try and
learn more about the many networking projects IS] undertakes.
First stop was to find Dr. Jon Postel, Director of ISI’s Communica-
tions Division. Jon was reading his mail when I came in, a never-
ending task when you get over 100 messages per day. He looked
up from his two-finger typing and gave me a warm welcome, evi-
dence of his long residence in California.

Postel is one of the best-known figures in networking. The edi-
tor of the RFCs titled “Internet Protocol” and “Transmission Control
Protocol,” he is certainly the person most often cited in the footnotes
at the back of research papers. In addition to writing key protocols,
Postel is the editor of the Request for Comment (RFC) series, and
has thus been responsible for the documentation of the TCP/IP pro-
tocol suite.

311

Exploring the Internet

Jon had set up a demanding schedule which we promptly shot
to hell. All the projects were so interesting, I kept dallying to find
out more. Even then, I saw only a fraction of what ISI had to offer,
not even getting to talk to people like Clifford Neuman, codeveloper
of key network services such as Prospero and Kerberos.

We started talking about the networks to which ISI belongs.
Several years ago, Postel sent a mail message around proposing that
a network, now called Los Nettos, be started. The very same day,
Susan Estrada of the San Diego Supercomputer Center proposed the
establishment of CERFnet.

Both networks went forward and they illustrate two different,
yet complementary approaches to providing network service. Jon
describes Los Nettos as a “regional network for Blue Chip clients.”
Los Nettos connects 9 big players, including UCLA, Caltech, TRW,
and the RAND Corporation. The network assumes that the clients
are self sufficient and thus has little overhead. As Jon puts it, “we
have no letterhead and no comic books.”

By contrast, CERFnet is a full-service network provider, one of
the most aggressive and best-run mid-level networks. CERFnet has
aggressively pursued commercial customers and helped to found
the Commercial Internet Exchange (CIX). CERFnet provides a wide
range of user services, including comic books, buttons for users, and
other introductory materials.

Both models are necessary and the two networks interconnect
and work well together. T1 lines from UCLA and Caltech down to
the San Diego Supercomputer Center give Los Nettos dual paths
into the NSFNET backbone. Having the two types of service
providers gives customers a choice of the type of service they need.

ISI is also a member of two other networks, DARTnet and the
Terrestrial Wideband (TWB) network, both experimental systems
funded by DARPA. TWB is an international network, connecting a
variety of defense sites and research institutes, including UCL in
London.

TWB is a semi-production network used for video conferences
and simulations. The simulations, developed by BBN, consist of ac-
tivities such as war games with tanks, allowing tank squadrons in
simulators across the country to interact with each other so com-

312

Marina del Rey

manders can experiment with different strategies for virtual destruc-
tion.

DARTnet, by contrast, is an experimental network used to learn
more about things like routing protocols. To learn more about
DARTnet, I walked down the hall to find Robert Braden, executive
director of the IAB. Braden was buried behind mountains of paper,
including a formidable pile of RFCs.

DARTnet connects eight key sites such as BBN, ISI, MIT, and
Xerox PARC. Braden described the network as “something re-
searchers can break.” Although applications such as video confer-
encing run on the network, the main purpose is experimentation
with lower layer protocols.

Each of the DARTnet sites run Sun computers as routers. Peri-
odically, network researchers, such as the elusive Van Jacobson,
modify the code in those routers to experiment with new routing or
transport protocols.

One of the key research efforts is supporting different classes of
traffic on one network. Video, for example, requires low delay vari-
ance in the data stream, a requirement not necessarily compatible
with large file transfers.

DARInet researchers are examining ways that resource reserva-
tion can be accommodated in a general-purpose network. Connec-
tion-oriented protocols, such as ST, share the underlying
infrastructure with the connectionless IP protocol. IP multicasting
and other newer innovations are also implemented in the network,
allowing bandwidth-intensive video to be sent to several sites with-
out necessarily duplicating the video stream.

The next stop on my tour was a visit with Gregory Finn and
Robert Felderman, two of the developers of ATOMIC, an intriguing
gigabit LAN under the direction of Danny Cohen. ATOMIC is
based on Mosaic, a processor developed at Caltech as the basis for a
massively parallel supercomputer.

One of the requirements for a message-based, massively parallel
system is a method for passing messages from one processor to an-
other. Mosaic had built this routing function into the processor. ISI
decided to use that same chip as the basis for a LAN instead of a
supercomputer.

313

Exploring the Internet

The prototype Gregory and Robert showed me consisted of
three Suns connected together with ribbon cables. Each Sun has a
VME board with four Mosaic chips and two external cables. It is
interesting to note that each Mosaic chip runs at 14 MIPS, making
the LAN interface board significantly more powerful than the main
processor.

The four channels on the board have a data throughput rate of
640 Mbps. Even on this rudimentary prototype, 1,500-byte packets
were being exchanged at a throughput rate of over 1 Gbps. Small
packets were being exchanged at the mind-boggling rate of 3.9 mil-
lion per second.

The system I saw simply had host interface boards connected
directly to each other. The same Mosaic chip is being used as the
basis for a LAN switch, an architecture that seems to provide a vi-
able alternative to the increasingly complex ATM switches.

The Mosaic-based supercomputer consists of arrays of 64 chips,
configured on boards 8 square. The ATOMIC project was going to
use those arrays as the basis for a switch, connecting host interfaces
(and other grids) to the edge of the array.

Messages coming into the grid would be routed from one Mo-
saic chip to another, emerging at the edge of the grid to go into a
host interface or another grid. Source routing headers on the mes-
sages were designed to take advantage of the hardware-based rout-
ing support in the processors. In effect, 896 MIPS of processing
power on a grid would sit idle while the routing hardware moved
messages about.

What was so intriguing about ATOMIC was the simplicity and
elegance of the architecture. While ATM switches had become ever
more complex in order to support more users or higher speeds, this
appeared to be an architecture that could scale evenly.

Even more intriguing was cost. The individual Mosaic chips
and the 8 by 8 grids were being made in large quantities for the
Caltech supercomputer and costs would descend to very low levels
if ATOMIC ever went into production. The tinker-toy approach to
connecting grids together meant that a LAN switch could easily
grow at small incremental costs.

314

Marina del Rey

ATOMIC is implemented as a LAN under the Berkeley sockets
interface, allowing TCP/IP to run over the network. There is no
reason, however, that ATOMIC couldn’t be made to send fixed-
length packets of 54 bytes, making the technology well-suited to
ATM and B-ISDN, as well as more flexible packet-based architec-
tures.

With only a few researchers, ISI showed that gigabit switches
didn’t necessarily have to involve huge development teams or com-
plex designs. By piggybacking on the Mosaic effort, the ISI re-
searchers quickly prototyped a novel solution to a difficult problem.

Just down the hall from the laboratory housing the ATOMIC
project is the ISI teleconferencing room. Multimedia conferencing
has always been a highly visible effort at ISI, and I met with Eve
Schooler and Stephen Casner, two key figures in this area, to learn
more about it.

Traditional teleconferencing uses commercial codecs (coder-de-
coders, sort of a video equivalent to a modem) over dedicated lines.
ISI, along with BBN, Xerox, and others, has been involved in a long-
term effort to use the Internet infrastructure to mix packet audio
and video along with other data.

The teleconferencing room at ISI is dominated by a large-screen
Mitsubishi TV split into four quadrants. In the upper left, I could
see Xerox PARC, in the lower right, the room I was in. In the Xerox
quadrant, I could see Stephen Deering, a key developer of IP multi-
casting, holding a meeting with some colleagues. In the upper right
hand quadrant, I could see the frozen image of a face from MIT left
over from the previous day.

Working on multimedia networking began as early as 1973
when ARPANET researchers experimented with a single audio
channel. To get more people involved, Bob Kahn, then director of
DARPA’s Information Processing Techniques Office, started a satel-
lite network. In time, this network evolved into the Terrestrial
Wideband (TWB) network.

Over the years, the teleconferencing effort settled into a curious
mix of production and research. Groups like the IAB and DARPA
used the facilities on a regular basis to conduct meetings, with TWB
logging over 360 conferences in the past three years.

315

Exploring the Internet

On the research side, people like Eve Schooler and Steve Casner
were experimenting with issues as diverse as the efficient use of un-
derlying networks to high-level tools used by meeting participants.

At the low level, one of the big constraints is raw bandwidth.
Video channels can take anywhere from 64 to 384 kbps, depending
on the brand of codec and the desired picture quality. Codec re-
searchers are examining strategies to squeeze better pictures out of
the pipes, ranging from better compression to clever techniques
such as varying the amount of data sent depending on the amount
of motion.

Teleconferencing, because of high codec costs, had been based
on connecting special, dedicated rooms together. With the cost of
codecs coming down and even getting integrated into workstations,
the focus is rapidly switching to office-based group applications, al-
lowing people to sit at their desks in front of their multimedia
workstations.

While desk-based video was still a ways down the pike, audio
was maturing rapidly. ISI has been working with people such as
Simon Hackett in Australia to come up with a common protocol for
moving audio around the network.

After this whirlwind morning of research projects, Jon Postel
took me to lunch and I returned to learn more about the key admin-
istrative role ISI plays in the Internet.

First stop was the famous Joyce K. Reynolds, Postel’s right arm.
In addition to announcing new RFCs to the world, Joyce manages
the Internet Assigned Numbers Authority (IANA), the definitive
registry of constants like TCP port numbers, SNMP object identifi-
ers, and Telnet options. IANA included all administered numbers
except for Internet addresses and autonomous system numbers, an
area in which ISI acts as a technical advisor to the NIC.

Although Joyce’s job in RFC processing and IANA functions
would seem to take all her time, she also manages the IETF User
Services area. For several years, she had been pushing the IETF to
try and make using the network as important a priority as making it
run. Finally, User Services got elevated to a real “area,” and Joyce
led the effort to publish tutorial documents and introductory guides.

316

Marina del Rey

Right next to Joyce’s office is that of Ann Cooper, publisher of
the Internet Monthly Report containing summaries of X.500 pilots,
IAB meetings, IETF and IRTF meetings, gigabit testbeds, and other
key activities.

One of Ann Cooper’s functions is administrator of the U.S. do-
main. While most U.S. names are in the commercial (.COM) or edu-
cational (.EDU) name trees, the rest of the world uses a
country-based scheme. The U.S. domain is an effort to bring the
U.S. up to speed with the rest of the world.

My own domain name is Carl@Malamud.COM. While this is
nice for me, placing myself this high on the name hierarchy starts to
turn the Domain Name System, in the words of Marshall T. Rose,
“from a tree into a bush.” In fact, Rose, being a good citizen of the
network, has registered himself under the mtview.ca.us subdomain.

While most U.S. members of the Internet are still under the
older .COM, .GOV, and .EDU trees, the .US domain is beginning to
catch on. In fact, enough people want to register that ISI] has dele-
gated authority for more active cities to local administrators. Erik
Fair, for example, the Apple network administrator, has taken over
the task of administering names for San Francisco.

My day at ISI was finished just in time to go sit in traffic. I had
proved myself to be a non-native of California by booking my flight
in the wrong airport, so I wheeled my car past three Jaguars head-
ing into the Marina and started the 3-hour drive south to San Diego,
watching the snow-capped mountains poking their heads over the
Los Angeles basin fog.

317

San Francisco

Tuesday morning, I took the shuttle from my hotel to the airport to
await the arrival of the flight from Tokyo so I could meet George
Abe, a manager of Infonet, a commercial, global IP network. The
driver asked us for our airlines.

“United, Domestic,” one person shouted.

“United, International,” chimed in a second.

A couple in the back looked at each other in bewilderment. The
lady elbowed her husband.

“Uh, United, Hawaii,” he said.

I met George in an airport lounge. Incredibly energetic after a
15-hour flight, he bounced in, whipped out his view graphs, and
started to tell me all about Infonet.

Infonet was originally started as a time-sharing company owned
by Computer Sciences Corporation (CSC), offering application pro-
grams on a series of Univac 1180s running their own proprietary
operating system. To access the mainframes, by 1971 the company
had started offering an X.25-like packet service in competition with
companies like Tymnet.

In the mid-1980s, George Abe worked as a member of a 3-man
team to upgrade this proprietary network to X.25. Then, CSC de-
cided packet networks were not their cup of tea and decided to di-
vest. Infonet managers were able to convince a large number of
PTTs around the world that they needed a way to compete against
global giants like IBM and AT&T. The sales pitch worked and In-
fonet became 100 percent owned by telephone companies, with ma-
jor shares held by MCI, the German Deutsche Bundespost, and the
French Transpac, with smaller stakes help by eight other PTTs.

By the late 1980s, the network had grown and was based on
multiplexors located in 34 key locations. Attached to these core
nodes were X.25 switches, PBXs, and other interface devices, all con-
nected together by international circuits.

In 1990, Abe spent a week in Finland with Juha Heinanen, and
the network soon added Cisco routers along with the X.25 switches,

318

San Francisco

allowing Infonet to provide IP service. Like the Finnish PTT service,
the IP service was a way of connecting corporate networks together.

The first question from MIS managers when George described
the service invariably was “Are you connected to the Internet?”

At first, George would hem and haw, sheepishly admitting that
they weren't connected with the rest of the world.

“Good!” was the usual reply. Large corporate MIS staffs were
scared stiff of the security implications of the Internet, imagining
hordes of hackers breaking into their systems.

Increasingly, though, users were beginning to demand some
form of interconnection, at least mail connectivity if nothing else.
Infonet was using PSI as a service provider, although I noticed that
George Abe had only an MCI Mail address.

The solution to access to the broader Internet was similar to the
one I saw in Finland. Inside of Infonet, a router at the border of the
PSI domain has access control lists. If a customer of Infonet pays a
fee, traffic for that particular Internet address can make it through
the router. Otherwise, even though there is data link level connec-
tivity with PSI, the router acts as a firewall, keeping out IP level
traffic for those who don’t want it.

What was interesting about Infonet was not the few hundred
customers it had signed up by the end of 1991, but the corporate
backers. In addition to its shareholder PTTs, Infonet had enlisted
most of the other monopoly carriers around the world as sales affili-
ates. In many countries, therefore, a company that wanted interna-
tional IP service would have to deal with Infonet.

I wasn’t quite sure what to make of this company, but made a
mental note that it was certainly worth watching. I left George Abe
to run a crucial errand down at Interop.

Ole Jacobsen, publisher of ConneXions, had asked me to help set-
tle an Acronym Dispute (AD). Barry Leiner, a former DARPA pro-
gram manager, had submitted an article which discussed, in part,
the International Standards Organization. Ole, with his eagle eye,
had tried to explain that ISO really stood for the International Or-
ganization for Standardization, but Leiner insisted that a dyslexic
acronym could not possibly be correct.

319

Exploring the Internet

I brought Ole the cover sheet from an ISO standard, a worthless
piece of paper that cost me U.S. $10. Sure enough, the International
Organization for Standardization had its name proudly embossed
on the front. Using the document to silence Barry Leiner was per-
haps the first time this particular standard had found a useful appli-
cation.

320

Tokyo

Friday morning, after a breakfast of green beans, cocktail franks,
and a salad, I met Kenji Naemura, at the time a vice president of
Nippon Telegraph and Telephone who has since become a professor
at Keio University. We sat in the back of a car with shiny white seat
covers and drove 50 km south of Tokyo to NTT’s Yokosuka research
laboratories.

Naemura began his career in the 1960s in Tokyo using a machine
developed by NTT that preserved the Illiac I interface but had a
different architecture, possibly the world’s first clone. Naemura de-
veloped the hardware architecture for NIT switches, then spent two
years at Champaign-Urbana working on the Illiac IV.

When he got back to Japan, the focus moved from switch archi-
tectures to communications software, and Naemura began playing a
central role in the Japanese delegation to ISO. He was involved in
the very first OSI meetings where the seven-layer Reference Model
was developed. Although the Japanese delegation would not have
been unhappy to have been referred to as the seven samurai, Nae-
mura felt that perhaps the archiecture should have a less rigid strati-
fication of the protocol architecture.

Naemura was bringing me to Yokosuka to learn more about
NTT’s vision of what the telephone network would look like in the
next 25 years. Dubbed VI&P, for “Visual, Intelligent, and Personal,”
the vision had been developed over several years as a broad consen-
sual effort within the corporation. The previous shared vision had
been the installation of narrowband ISDN around the country, and
with that effort well underway, NTT, the largest corporation in the
world, wanted to give its employees a sense of common purpose.

Basically, VI&P outlined a B-ISDN, all-fiber telephone network
with 620 Mbps channels into the home and one or more terabits of
capacity inside the network. NTT was targeting a 15 year develop-
ment effort, with large-scale deployment in the early years of the
21st century.

321

Exploring the Internet

Much of the VI&P effort dovetailed into current products. For
example, NTT had recently announced a 230 gram portable tele-
phone. One of the VI&P goals is for a 50-gram personal phone, cou-
pled with an intelligent network to find the phone wherever the
person might be.

We started our tour by looking at some of the prototypes, devel-
oped for press briefings, management demos, and other shows. Al-
though there were a half-dozen systems, the one that I found the
most interesting was called “teleview.”

The heart of teleview was image processing software to extract
objects from a video feed. The prototype had a camera trained on
an NTT office and used ISDN to send the video to a Yokosuka
workstation. Based on movement of people and the location of
desks, the computer could determine if people were sitting at the
desks.

My immediate reaction was that this was highly Orwellian, al-
lowing supervisors to track the length of bathroom breaks. While
toilet tracking was certainly one of the possible applications, there
were others.

For example, a touch screen on a workstation had a grid, with
one square for each employee. If the employee was at the desk, the
square was green. Touching the square automatically dialed the
telephone. Of course, if a dog were sitting in the employee’s chair,
the square would also be green, but the phone might not get an-
swered.

We touched a button and a poor harassed employee of NTT ig-
nored it for a while, then finally picked it up.

“Just testing, thanks a lot,” my tour guide said. The employee
didn’t bother turning around and I got the impression that he
wasn’t as thrilled with teleview as the researchers were.

The marketing video for VI&P showed a couple of other
teleview applications. Applied to a parking lot, teleview could di-
rect drivers directly to empty spaces. At an amusement park, the
system could count the number of people in line, displaying the ex-
pected delay for each ride on monitors.

After a walk through various labs, looking at ATM switches,
HDTV systems being fed with 150 Mbps lines, and prototype equip-

322

Tokyo

ment for distributing fiber from poles to homes, we drove back to
Tokyo to NTI’s Kasumigaseki showroom to look at ISDN equip-
ment.

Japan has very aggressively deployed narrowband ISDN, pro-
viding service in two thousand service areas distributed over all the
islands. Both basic rate and primary interfaces are available, and |
met several people with basic rate service in their homes.

The showroom was full of slick equipment ranging from G4
digital fax to dedicated floppy disk transfer devices. Videoconfer-
encing stations, ISDN pay phones, and portable video phones were
all available. Although the cost for equipment and service is not
cheap, NTT had managed to attract over 60,000 basic rate lines by
the end of 1991.

Throughout my visit, people kept pointing to picture phones,
and it was obvious that they hoped this device would spur personal
ISDN use. While picture phones didn’t appear to me to be the Lo-
tus 1-2-3 of the ISDN market, I did see one device that had real
promise, the ISDN Karaoke Machine.

Using 2 64-kbps B channels, a large database server could send
audio down one pipe and a video image with song lyrics and a
bouncing ball down the other. The system kept the bouncing ball
synchronized with the audio track. Here was a way to turn every
room into a karaoke bar.

Dr. Naemura dropped me off at NTT’s R&D headquarters where
I met the three scientists who were translating STACKS into Japa-
nese. Shigeki Goto, Ken-Ichiro Murakami, and Hisao Nojima came
walking into the conference room armed with huge stacks of paper.
For the next two hours, we went through a long list of what they
politely referred to as “questions.”

Most of the “questions,” of course, were examples of typos, mis-
calculations, and other errors on my part. Not only had they scruti-
nized my text on a word-by-word basis, but in most cases had gone
back to the original protocol specifications to check all data. It oc-
curred to me that perhaps we should translate the Japanese version
back into English for the next edition.

323

Exploring the Internet

Saturday, Jun Murai had asked me to come to the monthly WIDE
meeting. With a few hours to kill before the meeting, I went to the
Akihabara “Electric Town” to look at the latest palmtop computers.

An intriguing discovery was the Takeru Club, a vending ma-
chine for selling software. In the center of this 4-foot high machine
was a touch screen to select software. On the right were receptacles
for bills and coins. On the left were floppy drives for three different
diskette formats, and even a slot for dispensing laser discs.

Restraining myself from buying software, and tearing myself
away from a 14-channel programmable controller, I caught a series
of trains down to the Yagami campus of Keio University. I got a can
of hot coffee from the vending machine and sat down to wait for
Jun.

When we arrived at Keio, we found a room full of 40 volunteers,
students, and scientists, hard at work discussing internetworking is-
sues. Research results were presented, the latest routing tables were
discussed, and new plans for WIDE were formulated. This group
gathered once a month for an all-day Saturday session and had pre-
viously met twice a month. Monthly meetings by these volunteers
are supplemented every year at two camps, where all 65 WIDE re-
searchers and students collect in the middle of Mt. Fuji, wiring
themselves into the network with Ethernets and ISDN.

I gave my lecture, a rambling discourse on resource discovery,
then read my mail and went back to Tokyo and wandered around
the Ginza entertainment district. In a liquor store, I decided against
buying a bottle of Lemon Pie and Cream and went next door to a
random sushi joint for a dinner of uni and kazunoko. On the way
back to the hotel, I stopped into the Atlantis Blue bar, a fairly new
establishment proudly sporting “Since 19XX” on its sign, figuring
they deserved my money for using wild cards on their date.

Sunday morning, suffering from newspaper withdrawal, I went in
search of the International Herald Tribune. In Japanese, one asks for
the paper as “the Geraldo,” but a search of a half-dozen hotels in
my area yielded nothing.

324

Tokyo

After two hours of walking, I was getting hungry, so I followed
the railroad tracks to the next station. Under the tracks was a little
sushi restaurant, no bigger than your average closet. A dozen stools
were set around a circular counter. In the 2 by 3 foot space in the
middle of the counter was the chef. A conveyer belt full of plates
circled this island and customers simply grabbed plates as they
came by.

Over the belt was another level which held tea glasses, chop-
sticks, jars of radish, and, in front of each stool, a spout for dispens-
ing hot water, and a bucket of tea bags. If you didn’t see what you
wanted, you shouted out and the chef would make it and hand it
out. After a half-dozen dishes, I forked over 480 yen (U.S. $3.84),
amazingly cheap for Tokyo.

Fortified with octopus, I spied a bookstore, found my Geraldo,
and went back to the hotel to await my dinner appointment that
night at Jun Murai’s house.

Jun lives in Seijogakuen, the Japanese Hollywood. I boarded the
wrong train and ended up in Sagami Ohno, the equivalent of going
across Manhattan by way of New Jersey. The trains were full of
commuters, many wearing white surgical-style face masks worn
when sick to prevent the spread of disease.

After a few more wrong turns, I finally ended up at Jun’s house.
The house, on the same plot with his parents, is unusual by Japa-
nese standards. Apart from its large size, it features built-in Eth-
ernet cabling, wiring for 5-channel Dolby stereo, and a truly
impressive area of electronics gear. A little table near the entrance
held a stack of 12 remote controls for the VCRs, televisions, CD
players, and other evidence of many, many trips to Akihabara.

Over a dinner of salsa and margaritas, smoked salmon wrapped
in horseradish and California wine, and roast pork with champagne,
we talked about—no suprise here—computers. Jun keeps his finger
right on the pulse of the latest technology and seemed to already
have used most devices that I had only heard about.

Monday morning, I found a Yashinoya shop and had a bowl of
fatty beef over rice garnished with a raw egg. Totally enamored by
this time with Japanese food and electronics, I pulled out the Japan
Times to look for a place to live. A three bedroom apartment—“with

325

Exploring the Internet

telephone”—was available for a mere 1 million yen (U.S. $8,000) per
month, but the six month deposit would have put a real crimp in
my hardware acquisition plans. A nice room of 42.05 square meters
was only 200,000 yen, but I wasn’t quite sure how I could fit my
computers into a space that small, let alone find room for a bed.

At 9 AM., I met Tomoo Okada from Fujitsu for a trip down to
the new Tokyo city hall, at 60 billion yen (U.S. $500 million) reputed
to be the most expensive building ever built. We took the train to
Shinjuku station, which handles a mind-boggling 3 million passen-
gers per day, and walked out the west gate to the city hall.

The complex consists of two towers, 34 and 48 stories high, and
a 7-story city council hall, all clustered around a 5,000-square-meter
outdoor Citizen’s Plaza. All the buildings are made of a lattice of
light and dark granite, which, along with the reflective glass makes
the complex look like a huge integrated circuit.

We started on the first floor in the main control room for the
large tower. The room is divided into regions, one for each of the
major subsystems. Each region has a series of video screens and
keyboards. The elevator panel, for example, shows the status of all
20 elevators and has emergency shutoff switches for each one below
each screen. |

Another set of screens shows the status of all 800 card readers in
the complex. All employees have magnetic cards, and these are
used to clock in in the morning, to open conference rooms, and even
to start up computers. At the far end of the room, a technician was
using a mouse to click on the diagram of one of the floors. He se-
lected a room, and with his mouse adjusted the humidity setting
from 40 to 50 percent.

We took an elevator, equipped with extra buttons at waist level
for handicapped access, up two floors to a public videotex room.
Here, citizens can pull up and view or print demographic data,
schedules of cultural events, and city council meetings, as well as a
wide variety of other types of information. That same floor features
a publications sales outlet, a citizen counseling area, and other pub-
lic services.

On the 13th floor, we entered an office area 108 meters long
housing part of the MIS group. This floor, like all the other office

326

Tokyo

floors, has an “OA Room,” which houses PCs for secretaries as well
as the wiring closet.

The complex features an optical backbone with 7 km of fiber
running FDDI. Access to this backbone is through 100 bridges lo-
cated in office complexes, machine rooms, and other locations. At-
tached to the bridge is a star coupler which is used to distribute
fiber out to the workstations. These local loops run Ethernet and
add another 295 km of fiber.

On the 9th floor, we looked down from the viewer’s gallery into
the 2-story room housing the Disaster Prevention Center. This con-
trol room is used in the event of natural disasters, such as earth-
quakes, typhoons, or large fires.

In the middle of the room are two 200-inch video screens. Video
feeds from helicopters and other sources can be displayed on the
screens, as well as computer simulations which show the expected
spread of fires. On either side of the screens are other displays that
show the current weather conditions, the reporting status of civil
disaster teams, and other information needed to determine strategy.

Next, we went up to the 10th floor, devoted to seven 3090-class
mainframes from Fujitsu, Hitachi, and IBM. Another machine room
in the other tower has another five large systems, for a total of 3,000
square meters of machine rooms.

Each machine room has a fault-resilient IBM System 88, featur-
ing dual processors, dual disk drives, and dual buses. The two ma-
chines act as hot backups for each other, providing a second level of
protection. The System 88s are used for environmental monitoring,
using a token ring to control and read information such as power
status, humidity, and temperature. In case of an exception condi-
tion, the environmental monitoring system could initiate an orderly
shutdown, dial telephones for system programmers, and sound
alarms.

Each of the machine rooms features a special earthquake-resis-
tant raised floor. In case of large shocks, the steel lattice will absorb
and dampen vibrations, rolling the entire machine room in the op-
posite direction from the building.

Impressive as all these systems were, I was most intrigued by
the analog data jack on the ISDN pay phone in the lobby. The next

327

Exploring the Internet

morning, I called up Jun Murai and asked him to set me up a local
login, then headed back to Shinjuku.

I set up my office in the lobby, balancing my fax modem and
laptop on the counter. Ten minutes later, 1 was reading my mail in
Colorado, using the talk program to bother my systems manager,
and even Kermiting my files down to my laptop to work on later.

Periodically, people would sidle over to try and figure out what
this crazy Gaijin was up to. Each time I spotted somebody looking
over my shoulder, I would point excitedly at the screen and say
“America!” This was usually enough to evoke a polite smile and a
hasty retreat.

At Narita airport, I set up my laptop in the lounge to work on
the files I had downloaded. Sipping a scotch and munching dried
squid, I noticed that people kept their distance.

328

Seoul

My previous visit to Korea had been a six-hour transit stay just be-
fore the Olympics and I had been used as training material for a
half-dozen different security teams getting their procedures ready
for the onslaught of tourists and athletes. My current visit was con-
siderably less high-strung, although there was still an awfully large
number of security personnel in evidence. It was a traffic violation
control day and each intersection had one or two cars pulled over
for traffic violations and the expressway had police vans every few
hundred meters, waiting for potential perpetrators to pass.

Wednesday morning, I took a cab out to the Korea Advanced
Institute of Science and Technology (KAIST) to meet Kilnam Chon, a
professor in the computer science department. KAIST is the center
of one of Korea’s R&D networks, the System Development Network
(SDN) which, for some reason, is also known as Hana. SDN has
hubs at the two major KAIST locations in Seoul and at the Daeduk
Science Park.

Fifteen universities and research organizations connect to the
two hubs at speeds ranging from 9.6 to 56 kbps. A 56 kbps line to
the University of Hawaii links this TCP/IP network into the In-
ternet. SDN supports OSI applications such as FTAM and X.400,
but usage was declining rather than increasing. Five years ago, in
the middle of a large OSI push, 20 percent of traffic was X.400, but
the number faded to insignificance as users switched over to SMTP-
based mail handlers.

There are two other TCP/IP networks in Korea. While SDN
supports itself by fees from members, the other two are government
supported. The Korea Research Environment Open Network
(KREOnet) is sponsored by the Ministry of Science and Technology
and was originally established to provide an access path to a Cray.
KREOnet was linked to SDN in two locations, providing a fairly
seamless Internet.

KREOnet also maintained a 56 kbps link to CERFnet in San Di-
ego, but it appeared that most Internet traffic moved across the

329

Exploring the Internet

PACCOM link to Hawaii. Though the Hawaii link was saturated at
times, for some reason the CERFnet link had been relegated to a role
as a backup.

The third network was the Korea Research Network (KREN),
sponsored by the Ministry of Education. The network linked eleven
BITNET sites in Seoul to Japan via a 9,600 bps line. KREN also
linked nine universities together with 9,600 bps TCP/IP lines and
provided a path into the rest of the Korean Internet.

After a morning briefing on the networks, Professor Chon in-
vited me to choose between hamburgers and “lousy Korean food”
in the KAIST dining room. Lousy was an overstatement, but it cer-
tainly would not be described as spectacular.

What made the lunch palatable was the conversation. Professor
Chon was the chair of the JTC1 Korean Committee, the key commit-
tee for OSI work, and was also active in the CCITT standards arena.
He had heard that the Bruno project had been killed and grilled me
for the reasons.

Professor Chon had been planning on replicating Bruno for Ko-
rea and we discussed whether Korea really needed the ITU’s per-
mission. As an individual, one could argue that I had in fact
needed their permission, assuming you take the copyright assertion
at face value.

A country, on the other hand, was in a different situation. After
all, the ITU had come into being when countries like Korea had
signed the ITU treaty. A key aspect of international law is that any
right not explicitly delegated is retained by sovereign states, and
one had to wonder if the ITU could prohibit Korea from distributing
standard documents to its citizens.

One could even read the ITU treaty as encouraging this behav-
ior. The treaty required that all members take all possible steps to
achieve the broadest dissemination of ITU recommendations.
Though online distribution of standards was not specifically men-
tioned, it was certainly not explicitly prohibited.

I explained my concept of a “standards haven,” where a country
would ensure that standards were available online to its citizens.
Not for anonymous FIP, mind you, since that would muddy the
issue of sovereignty. Once the standards were online in one country,

330

Seoul

though, it would be a simple matter of cutting tapes for other coun-
tries.

“Let’s do it!” Professor Chon said. I cautioned that there might
be political fallout, with the ITU possibly objecting to the Foreign
Ministry, a common bureaucratic home for the official delegate to
the ITU. In Korea, though, the ITU representative happened to be
in the Ministry of Telecommunications. To my delight, the Ministry
had already gone on record as actively supporting online distribu-
tion of standards and had passed the requirement for such a fileser-
ver down to Korea Telecom, the PTT.

Lunch started to taste better and I finished my kim chi. Korea
was looking more and more like it might become the world’s first
standards haven.

The conversation then shifted to other standards documents,
particularly OSI. There were two types of OSI documents that inter-
ested Professor Chon: working group documents and the final prod-
ucts, the International Standards.

Working group documents constituted enormous piles of study
papers, submissions, drafts, technical corrigenda, and other docu-
ments in such profusion that simply moving paper around to par-
ticipants had become the key bottleneck in the standards process.

ISO had formed a working group on working group procedures,
but the working group was still bogged down in formulating the
procedures under which it would operate. The best that they had
come up with had been to use floppy disks to exchange ASCII-
based files.

The secretariat for JTC1 was, unfortunately, ANSI. Any attempt
to provide a file server for distribution of working documents
would require the cooperation of the ANSI secretariat and it was
highly unlikely that this notorious group of luddites would want
any part of it. ANSI had been the single most vocal critic of the
Bruno experiment, sending their objections straight to Pekka Tar-
janne.

Online distribution of working documents would certainly im-
prove the ISO process, but I wasn’t so sure that I wanted to improve
it. It would be much more interesting to get the product from that
process available to those mere mortals who didn’t have the time or

331

Exploring the Internet

money to fly to exotic locations and stay in expensive hotels. People
in developing countries, students, and others without extensive re-
sources should at least be able to read the international standards.

ISO standards, centrally produced and all in a common format,
were a prime candidate for scanning and optical character recogni-
tion. A single format and high-quality documents means that the
OCR software can quickly learn and can produce a fairly accurate
rendition of the standards. Modern OCR software can go so far as
to save markup information such as the font size or the location,
making semi-automatic conversion into a language such as SGML
very feasible.

The only hitch was that distribution rights went from ISO to the
national standards bodies and ISO appeared to have inserted a
minimum selling price for international standards. For draft inter-
national standards, though, there was no minimum selling price. In
other words, a national standards body could give away a DIS but
not an IS.

In most cases (but not all), the draft standard was virtually iden-
tical to the final product. The definition of a draft standard is that
there are only minor technical corrections to be made. It would cer-
tainly be nice to distribute the final product (the idea of giving stu-
dents an inferior product certainly grated), but this might be a
loophole that would force some change.

The goal of setting up standards havens was not to get countries
into the standards distribution business. Instead, I was much more
interested in getting ISO and ITU to see the light and begin the
process of distribution themselves. After all, it would make so
much more sense to put a series of FIP (or FTAM) servers around
the world and release all documents over the Internet.

Professor Chon set me up to visit Korea Telecom the next day to
pitch my idea, and I went back to town to take care of unfinished
business. First on the list was a new power cord. The lounge at
Narita Airport had only 2-prong straight outlets and, desperate to
get some work done, I had finally snapped the ground prong off the
cord for my laptop.

Luckily, Korea used the same outlets as the U.S., including the
same grounded outlets. Unluckily, however, getting the concierge at

332

Seoul

the hotel to send me in the right direction was a bit of a challenge.
Arranging a visit to the North Korean border would have been a
snap, but computer components drew a blank. |

I made typing motions until the concierge shrugged her shoul-
ders and wrote the name Se-Woon Sang Ga on a piece of paper. I
handed the paper to the cabby who shot me a puzzled look,
shrugged his shoulders, and took off. Ten minutes later, he pulled
in front of a run-down 5 story concrete building.

Turned out this was where people rebuild old video games.
Room after room was filled with old motherboards, technicians
squinting at oscilloscopes and old men playing Go. The halls were
crowded with rebuilt machines and school children came flooding
in to play free games, testing the machines.

I wandered past a store that sold nothing but joy sticks, another
that had ribbon cables, several video tube stores, and, inexplicably, a
sculptor making plaster busts of the dearly departed. Finally, I
found a shop with a few PCs running Tetris. I made jabbing mo-
tions at the wall, ran my hands in a snaking motion towards a PC,
and reached for the power switch, making flicking motions.

One of the kids, realizing that I probably wouldn't just disap-
pear, got up from his game and I repeated my charade. He nodded
and reached up to the top shelf and handed me a power cord. I
forked over my 3,000 won (U.S. $4) and went happily back to the
hotel.

The next day, I took a cab way out to the outskirts of Seoul to the
research laboratories of Korea Telecom. The Umyon Dong building,
I found out sitting in Future Hall, was the most intelligent building
in Korea.

It certainly looked smart. Everything gleamed, the halls were
empty, and there were lots of video screens all over. Ushered up to
the second floor, I met Moon-Haeng Huh and Joo-Young Song, two
senior officials of Korea Telecom. They showed me around the
building, pointing out CATV servers, the FDDI backbone, and even
fingerprint recognition equipment for entry to high security areas.

333

Exploring the Internet

The place didn’t even have light switches or temperature controls in
the rooms: you used the telephone to call the automated control
server, keying in the temperature you wanted for a room.

After some ginseng tea, we discussed the idea of a standards
server. The labs, with an IBM 3090, Pyramid superminis, VAXen,
Sun servers, and lots of other equipment, certainly had the re-
sources. They also had the mandate. We agreed that this looked
like a likely project and I left agreeing to send in a formal proposal.
I couldn’t wait to get home to get this started, but I had a few more
stops to make first.

334

Taipei

I handed over TW $1,000 (U.S. $40) to the cab driver and checked
into my hotel, asking for a 4:45 A.M. wakeup call. It was 9 PM.
While waiting for my baked dried squid with chili sauce to come up
from room service, I read about SEEDNET, half of the Taiwanese
Internet.

As with many countries, the Internet was split among multiple
ministries. Education ran TANET for the universities and the com-
mercial SEEDNET was run out of industry. The namespace was
split between the two groups.

What was interesting about SEEDNET (but certainly not brought
out in the paper I read) was that SEEDNET used a block of 249
Class C addresses. SEEDNET was connected to JVNCnet, which in
turn was connected to the NSFNET backbone.

Having a country take 249 separate Class C addresses was an
interesting illustration of the problem of the IP address space. Class
B addresses were being rationed because of exhaustion of that por-
tion of the address space, but giving out multiple Class C addresses
could put an additional strain on the routers, already working hard
to keep up with close to 5,000 known network addresses.

At the time of my visit to Taipei, SEEDNET was physically con-
nected to JVNCnet, but the SEEDNET world was not announced
over the NSFNET backbone to other regionals. If you attempted to
connect to SEEDNET from JVNCnet, the routers would know how
to route the packet. If you tried to do so from another regional,
packets would disappear.

The SEEDNET problem was certainly just a temporary one, but
it showed the strains that were beginning to appear on the routing
infrastructure of the Internet. Older networks, such as the MILNET,
had even more problems.

The MILNET routers could only handle routing information for
3,750 networks. Since the NSFNET had over 5,000 networks, the
MILNET administrators had to decide which routes to accept and
which to ignore. The decision had been that all international net-

335

Exploring the Internet

works that were not associated with peer military organizations
were cut off inside the MILNET. The Australian AARNet, for exam-
ple, could no longer communicate directly with hosts inside the
MILNET.

Cutting off people who probably wouldn’t talk to you is cer-
tainly a rational response to the problem of saturating the Internet.
The problem, however, was that this didn’t solve the long-term
problem of scaling the Internet. The Internet was doubling every 7
to 10 months and there were projections that the routing table size
could easily grow by an order of magnitude if nothing was done.

Address space resolution was a big issue at all the IETF meet-
ings. One solution which was proposed was to do away with the
distinction of address classes and instead delegate arbitrary blocks
of addresses to the regional networks, which would in turn delegate
blocks to their clients.

Using blocks of addresses and delegation on the hierarchy
meant that, with suitable changes in the routing protocols, a single
network prefix and mask could be used to refer to what otherwise
might have been several hundred (or even thousand) entries in a
routing table.

I fell asleep trying to figure out why it had to be so hard to put
together an integrated global Internet. There was obviously a need
for many types of networks: the day of “the” network had long
passed. Yet, this diversity meant that the network was starting to
fragment and splinter into subsets of connectivity.

The next morning, dozing on the airplane to Hong Kong, I was
awakened by the flight attendant who presented me with a large
box with the imprint “Name Card Holder” on it. Looking around, I
saw all the other business class passengers stuffing their gifts into
carry-on bags.

I opened mine to find a handsome slab of wood with a large
blue and green ceramic duck mounted on it. Presumably, the busi-
ness cards would be inserted in the beak. While this functional yet
attractive desk accessory would certainly have made an impressive
addition to my office decor, I felt that my seatmate might make bet-
ter use of it. I handed my gift to him and he whisked it into his bag
before I could change my mind.

336

Hong Kong

Hong Kong was hell. I got in at 9 in the morning and had 12 hours
to kill before my flight to Bombay. My first stop was out to Sha Tin
to visit Chinese University. Eric Lo Hing Cheung started me up a
visitor account and let me read my mail, telling me about the latest
status of Internet access. To nobody’s surprise, the system had
moved forward at a remarkable rate since my last visit and most of
the Hong Kong universities were up and running.

Eric and Dr. Michael Chang were both gracious as always, but I
had a problem. My back went out and I could barely move. I
travel very light when it comes to clothes and other luxuries, but all
the books I insist on bringing with me brings the weight up to the
level that is enough to seem light but can strain the back when you
take a different plane every day. Sure enough, I felt like lying on
the floor at Chinese University, but that would have been less than
polite.

I took a cab to the main island to find a place to lie down. The
fare card in the taxi had a list of all the surcharges, including HK $4
(U.S. $0.50) for each “piece of luggage, animal, or bird.” At the bot-
tom, though was an addendum which said that “wheelchair or
crutches, carried by a disabled person free.” I tried to think of the
legislative history that must have gone into such a clause.

I spent the rest of the day with my friend Harry Rolnick, the
writer. Harry is not the kind of person who minds if you spend the
afternoon lying on his floor and I hoped that my back would re-
cover.

It was deadline day for Harry, a weekly columnist for the South
China Morning Post. It being March 6, he decided that the Michelan-
gelo virus would be a likely topic. Harry is not your most techni-
cally astute type—this is somebody who still uses Wordstar—but
has a great sense of humor.

We spent the next hour trying to dream up new viruses. The
civil servant virus was easy enough: it simply dims the screen and
does nothing. The Hong Kong waiter virus has possibilities: you

337

Exploring the Internet

can’t find the thing, but it will pop up occasionally and snarl at you.
And, of course, there is the expatriate manager virus, which won't
go off until mid-morning, when it displays “what a night, what a
night” on the screen and then goes to lunch.

While Harry finished his column, I laid on the floor and looked
up at the walls. The walls were certainly interesting. When he is
not reviewing concerts or restaurants, Harry does things like travel
to Korea to write feature pieces about kim chi or travel through the
mountains of Pakistan to get a little exercise. His walls show the
results.

On one wall was a framed copy of the Pyongyang Times from
North Korea. “Comrade Kim Jong Il Inspects Kwangbok Depart-
ment Store” blazed the headline. On the other wall was a 1991 cal-
endar from Kampuchea Airlines, certainly a rare item.

To help me pass the time, Harry handed me down a copy of a
North Korean-English phrasebook he had acquired on a recent visit.
Verb conjugation was illustrated by the following sequence:

“We fight against Yankees”

“We will fight against Yankees”

“We have fought against Yankees”

My favorite section, though, was the one on admirations, point-
ing out ways to say nice things about people.

“What a wise leadership.”

“Fancy abolishing taxation!”

“This is a success for the people indeed.”

This phrasebook made me want to spend some time in North
Korea. After all, where else can you find such an interesting permu-
tation of the language?

I started to count the number of cities I had been in the past 5
months. Flat on my back, nursing a Mekhong whiskey to ease the
pain, I wondered if it was worth it. There had to be an easier way
to earn a living.

Then, Harry walked in and showed me some of his clips, includ-
ing his story on kim chi, the fiery Korean pickled cabbage, made by
burying cabbage and peppers for the winter to ferment. Perhaps
there was an easier way to earn a living, but sitting in a cubicle

338

Hong Kong

hacking UNIX kernels would never expose me to the wonders of
kim chi.

I dragged myself downstairs, caught a taxi, and gritted my teeth
all the way to the airport, dragged my way through customs, and
finally got on the plane. Even the dumpling stuffed with yam ice
cream and served with a delicate, blue peanut sauce didn’t ease the
pain.

339

Bomba

At 2 AM., I crawled off the plane in Bombay feeling miserable. Brit-
ish Airways whisked me through the diplomatic line and carried
my luggage out to the curb where I grabbed a shuttle to the closest
hotel I could find.

Twelve hours later, after an hour under a hot shower, I finally
felt up to a taxi down to Juhu Beach, south of Bombay, where I lay
in my hotel room for the next two days recuperating. Lying on the
floor, I spent my time reading marriage ads in the Times of India and
watching Hindi films. Outside, the beach was teeming with people
riding horses, watching the horse races, or gathering around acro-
bats doing flips. Looking down, I could spot a row of Kalaghatta
stands advertizing the cold, ink-black drinks known as the “poor
man’s Coke” and thought to myself that a beer would be nice. Just
then, the phone rang.

“Mr. Carl, this is room service. Would you like some tea or
something to drink?” The idea of telepathic room service appealed
to me. I began to feel much better.

Monday morning, Dr. S. Ramani, director of the National Center for
Software Training, sent a huge old empty Tata bus to collect me. We
rattled our way through the swarming vehicles adorned with “Horn
OK Please” bumper stickers, past the shack which specialized in re-
winding old fans, to the NCST headquarters.

My day with Ramani was a whirlwind of activity. Fascinated by
how to use networks to help his country, Ramani rapidly jumped
from ways to help the Indian Antarctic expedition to how to control
infectious diseases to railway reservation systems to ways to dis-
seminate Indian news to graduate students overseas.

He would leap to the white board and draw a diagram, then call
out for his secretary to get certain staff members, then he would

340

Bombay

grab first one, then two telephones and start placing calls to officials
and corporate executives for more information.

“Ramani, here,” he said, calling the managing director of the
Press Trust of India, a wire service with over 1,700 employees. “I
have some ideas to discuss. Can you have dinner tonight at 9?”

The managing director had just gotten back from a trip to Delhi
to meet with the Prime Minister, but dinner was arranged.

Meanwhile, we discussed ERNET, the Indian research network.
The backbone was built around hubs in four cities—Bombay, Ma-
dras, Bangalore, and Delhi—which were connected to each other by
9,600 bps leased lines running TCP/IP. Other sites all over the
country used UUCP into these hubs.

Over 70 sites in ERNET shared a single 9,600 bps satellite link to
UUnet for international access. Needless to say, I felt more than a
little bit guilty using Telnet to read my mail back in Colorado.

Low bandwidth, both domestic and international, was partly
due to India’s economic situation and partly to some technical de-
tours towards satellites instead of terrestrial lines. Unlike some
other countries I had visited, the underdeveloped Internet infra-
structure was certainly not due to the lack of qualified engineers or
a demand among users. India has one of the largest cadres of scien-
tists in the world and the computer scientists I met at NCST were
highly capable and fully aware of current developments in the In-
ternet.

The economic situation was certainly a prime stumbling block in
putting together a real research network (let alone a commercial
service offering). The rupee had only just started to become con-
vertible when I visited India and foreign exchange was at a pre-
mium. The U.S. half-circuit for the 9,600 bps line had been funded
by the United Nations Development Program (UNDP).

Even within India, getting phone lines was no mean feat. Lines
cost roughly the same as in the U.S., but average salaries in the U.S.
are an order of magnitude higher than in India. The telephone in-
frastructure in India was notoriously underdeveloped, with electro-
mechanical exchanges (i.e., something has to move to complete a
circuit) still in operation in many locations. The wait for a residen-

341

Exploring the Internet

tial line could easily take several years. Businesses pay a substantial
surcharge to cut the wait to weeks or months.

In rural areas, lines are so unstable that running UUCP over
them could lead to endless retries. Even when the lines are opera-
tional, overhead of 500 percent is not uncommon. In fact, in some
areas the lines were so bad that ERNET used a novel “floppy-based
e-mail system.” It turned out to be significantly more cost effective
when the line got bad enough to simply spool mail to floppies and
hire a courier to bring the diskettes down to the nearest hub. One
system manager in Kanpur was once observed going through this
biweekly ritual, neatly labeling floppies and stacking them into a
pile.

“These guys have no other business,” he grumbled, “all they do
is keep sending mail through the day and night.” Ramani showed
me the network management utility which showed 30 to 70 mes-
sages exchanged per courier run with one of the FloppyNet sites in
Kanpur.

Even in Bombay, local lines are not immune to problems. Dur-
ing the summer monsoon, 80 to 100 inches of rain can fall in 80 days
of the monsoon season, knocking out service for 7 to 10 days per
year. For this reason, the main international gateway had been lo-
cated 20 km south of NCST in downtown Bombay. This machine,
Sangam, was located in the Air India building and was just a kilo-
meter from the telephone company’s point of presence, making it
fairly immune to disruptions during the monsoon.

‘20 kilometers south of NCST, in downtown Bombay’.

A poor telephone infrastructure led Ramani and others in the
Indian academic community to look to satellites, particularly the In-
dian manufactured INSAT series. The satellites gave high band-
width, used Indian technology, and needed only a dish on the roof
instead of depending on local infrastructure.

The original ERNET project plan from 1984 envisioned linking
the 8 large academic and research institutions together with 4.5 me-
ter dishes. For 1.5 million dollars, this infrastructure would provide
multiple data paths, based on, of course, OSI protocols. By 1988, the
project had evolved to specify 64 kbps data paths for the transmis-
sion of voice, data, and fax.

342

Bombay

The voice requirement would satisfy needs for voice broadcast
services for seminars, and would allow network management to
have their own voice channels. The data channels would be estab-
lished using the X.25 protocols and very coarse time slots of 1 sec-
ond or even more would be provided. The data channels were thus
suitable only for file transfer and electronic mail.

Although the project plans developed on schedule, the satellite
kept getting delayed. By 1988, though, some Indian students re-
turning from the states started agitating for Telebit modems, capable
of running at fairly high speeds over voice-grade lines.

Anil Garg and some other NCST staffers began playing with
UUCP dial-up transfers. The first links were in Bombay between
NCST and a sister institution, the Indian Institute of Technology
(IIT). At the time, the only phone available at the IIT computer sci-
ence department was located in the office of the chairman.

To transfer mail, Anil would call up the chairman of the depart-
ment, apologize for the interruption, and ask him to place the phone
on the modem. If the phone was needed for voice sessions, then
e-mail would have to wait. Eventually, a set time slot on the tele-
phone was allocated and, assuming calls got through, e-mail started
to flow regularly.

Over time, the terrestrial network started to grow. By 1991, with
a little prodding from Vinton Cerf, an advisor to ERNET, the net-
work was based on leased lines and Cisco routers. In early 1992,
Ramani was in the midst of careful negotiations with the Depart-
ment of Telecommunications to use a 64 kbps chunk of a new 140
Mbps fiber line to Delhi and maybe even to get a piece of the digital
microwave links to Bangalore and Madras.

Meanwhile, the satellite project was finally getting off the
ground, but its purpose had to be reevaluated. The terrestrial
ERNET was obviously much more suitable for interactive applica-
tions and with 64 kbps terrestrial lines coming in the foreseeable
future, Ramani was searching for a use for the satellites. Ramani
was hoping that the satellite WAN could supplement the terrestrial
network, perhaps being used by applications with a broadcast focus,
such as the distribution of news.

343

Exploring the Internet

Over a lunch of aloo bhindi and carrot halwa in the NCST can-
teen—equipped with a PC and custom software, of course—Ramani
explained his attempts to automate the infectious disease units of
the local hospitals.

Bombay hospitals get several hundred admissions per day for
diseases ranging from dysentery to measles, polio, and hepatitis.
An estimated 20,000 deaths per year are caused in Bombay by dis-
eases that are preventable by vaccination, boiling drinking water, or
other simple means.

When an admission occurs at a hospital, the admissions officer
fills out a slip and places it in a pigeonhole for one of Bombay's 23
wards. The slips are picked up and, within 24 hours or so, make
their way to public health officers in the wards who take appropri-
ate actions. In the case of measles, for example, the health officer
might go to a child’s building and look for unvaccinated playmates.

Ramani had been conducting briefings and otherwise pushing
people to computerize the process. He was hoping that a small
number of computers could easily cut out the 24 hour delay. More
importantly, aggregate information could be quickly examined for
trends. A cluster of hepatitis cases, when displayed on a map
would quickly indicate contaminated drinking water in a neighbor-
hood. The information could be used to make special efforts in
those neighborhoods to try and convince people to boil their water,
at least for a while.

That afternoon, I F[Ped my mail down to a local account to
avoid tying up the international link, then went downstairs to give a
lecture on the politics of standards. The audience of several dozen
people were all active in implementing standards or using them as
part of their work. This was certainly a sympathetic audience and
everybody laughed when I explained the theory the standards pota-
toes advanced that anybody “serious” about implementing stand-
ards could certainly afford to participate in the process. Somebody
came up to me after the lecture and explained how tough it was to
get foreign exchange allocated for buying documents, let alone tak-
ing trips to conferences and standards meetings.

After the lecture, I went back to my hotel and pressed (having
finally reprogrammed myself) the elevator’s up button to go down

344

Bombay

to the coffee shop. The coffee shop was playing a rousing polka
over the intercom, but I didn’t see any kielbasas on the menu so I
ordered some samosas and sweet lime water instead.

That evening, Dr. Ramani picked me up and we went to a local
restaurant to meet Gourang Kundapur, managing director of the
Press Trust of India (PTI). Over a dinner of brains masala and tan-
doori cauliflower, we gossiped about Indian politics, a topic even
more complex than the politics of standards.

Tuesday morning, my driver took me from Juhu to Bombay to see
the railway reservation system at Victoria Terminus, one of the main
train stations and a Bombay landmark. Smack in the middle of rush
hour, we dodged goats tied to shacks on the side of the road, push-
carts loaded with potatoes, and cows that ambled along content in
the knowledge that they alone were safe on the roads.

My driver took what might charitably called a fairly aggressive
approach to driving. He clearly felt that the green light was an indi-
cation that he should have already cleared the intersection. Even
stuck behind several dozen cars, he would keep his hand on the
horn. The side mirrors were carefully folded flat against the car,
giving at least 3 inches of extra maneuvering room.

Arriving at the Central Reservations building, I entered a very
large room filled with an incredibly dense mob waiting for tickets.
Fighting my way up to the information counter, I was handed a
form before I could even open my mouth, then was promptly swept
along with the crowd.

The Indian Railways system is one of the largest in the world,
moving 10 million passengers a day in over 6,000 trains. Tickets for
these trains come in 7 categories (e.g., express or local), 32 kinds of
quotas (e.g., foreigner or defense official), 100 types of discounts
(e.g, veteran or handicapped), and 7 classes of reservations (e.g.,
first class or first class air conditioning).

The old, non-computerized system had, although I had a hard
time picturing it, been known for incredibly long lines, which pre-
sumably meant that the lobbies must had somehow fit in even more

345

Exploring the Internet

people than I saw. To make a booking, you needed to be in the
right line for the particular class of tickets for the particular train
you wanted to take.

Demand has always far outstripped supply for Indian trains, so
it was not unusual to have 1,000 people clustered around one clerk,
while the next counter was empty. An 8-hour wait in line only to
find there were no seats left was not unheard of.

Assuming you were able to fight your way up to the top of the
line and get the clerk’s attention, the process still had no guarantees.
Complicated rate calculations, manual ledger books, and a host of
paperwork led to lots of errors. When you got to the train, you
might easily discover that there were duplicate bookings or that a
supposedly sold-out train had dozens of empty seats.

To find out about the new and improved system, I met with
officials from CMC, Ltd. the government-backed system integrators
that had computerized the reservations process. CMC had set up
five separate reservations systems for each of the five main regions
of India, centered in Bombay, Delhi, Calcutta, Hyderabad, and Ma-
dras.

Four of the five systems were based on VAXen. For some rea-
son, Hyderabad had some small CDC systems. Bombay was typical
with two VAX 8650s and a 6310 clustered together. All the code was
written in FORTRAN and even the database management functions
had been locally developed.

The Bombay system handled 7 different reservation centers. Vic-
toria Terminus was the largest, with 80 terminals, and the nearby
Churchgate station had 20 terminals. Muxes and modems linked
remote terminals to the cluster. A satellite office in Ahmedabad, 500
km away, linked 45 terminals over a series of eight 4,800 bps lines.
All told, around 200 terminals handled 50,000 to 60,000 transactions
on a typical day, with peaks up to 90,000 in the busy season.

For customers, the system meant that a reservation for any train
in any class could be made in any line. Just as importantly, far
fewer errors were made, since fares were automatically calculated
and it was much harder to double-book seats.

Oddly enough, there was a downside. On the old system, it was
possible to gauge your approximate chances of getting a ticket by

346

Bombay

the size of the queue and your ability to step over people and get to
the front. On the new system, it was much harder to estimate your
chances, since you competed with people all over the city. Monitors
were posted at the stations indicating how many seats were avail-
able on the more popular runs, but the Chief Commercial Superin-
tendent told me he was certainly receiving complaints from people
unable to get tickets.

The huge excess demand for tickets made scalping a profitable
business. Ticket sales were limited to 4 to 6 per person to prevent
scalping and each ticket had imprinted on it the name, age, and sex
of the passenger. An ID was not required to ride the train, but con-
ductors attempted to make sure that the holder of the ticket had at
least some resemblance to the data on the ticket. The main result
was that scalpers had adopted more sophisticated inventory man-
agement techniques.

While the seven main reservation centers in the Bombay region
were online, smaller stations still used electromechanical teleprinters
to communicate with Bombay. The messages would be received on
slips of paper, a clerk would key in the data to the computer and
send a message back to the station which would issue the ticket.
Recently, data from a few teleprinter sites had been fed straight into
the VAX. Hooking teleprinters to the VAX allowed properly format-
ted messages to be automatically answered. Not quite an interac-
tive terminal, but much quicker than the old system.

This same message switch was being used to connect the inde-
pendent systems for each region. When I visited, each center had a
few terminals for each of the other regions, allowing a Bombay pas-
senger to book passage on a train originating in Calcutta. Hooking
those lines into a local VAX instead of a terminal would provide a
way for one system to be a virtual teleprinter to a remote region.

After my briefing, my hosts insisted on showing me their ma-
chine room, which we entered via the washroom. Afterwards, I
rode back out to Juhu, passing a line three blocks long waiting to
enter the temple of Ganesh, the god with the elephant head.

347

Exploring the Internet

Wednesday, Ramani sent one of the NCST jeeps to pick me up. The
jeep was prominently adorned with “Govt. of India” on the front
and back. NCST is not technically a part of the government, but
Ramani explained that the labels made parking significantly easier.

I spent the day at the edge of the Ramani cyclone, periodically
leaving to check on my FIP jobs, which were downloading source
code for utilities that I thought might be useful to NCST. By the
end of the day, netfind, perl, traceroute, and WAIS had all made it
over the link.

The only thing I really wanted to bring down was the code for
Cleveland Free-Net. A Free-Net is typically linked to Cleveland
over the Internet and has a local bank of modems for public access.
Ramani and I couldn’t see how NCST could meet this requirement,
but wanted to see if there was some way that the Free-Net concept
could be applied in India.

Unfortunately, Free-Net wouldn’t let us do that. Public access or
nothing. People in India rarely had PCs and modems at home, and
adding substantial traffic to an already saturated 9,600 bps link just
didn’t make any sense.

WAIS had no such arbitrary administrative restrictions, however.
While ideally a WAIS client accessed servers all over the world, it
was certainly technically feasible to run an isolated WAIS world on
an Ethernet, providing a local information environment. Running
WAIS with the NCST LAN was useful as a way for building up
local expertise and seemed to fit right in with another current pro-
ject.

NCST was taking the wire service from the Press Trust of India
and feeding it into a VAX. There, the news feed was automatically
broken up into individual items and fed into a local USENET news
group. Every Friday, one day’s worth of news was posted globally,
providing information on Indian politics, sports, and culture to
graduate students and professionals overseas.

Once the news hit the VAX, applications like WAIS might be an
interesting method for searching and reading the news. NCST pro-
grammers were also developing their own sophisticated applica-
tions for sifting through information, such as a stock and news
monitor that was built on SCO’s Open Desktop.

348

Bombay

All these fancy user interfaces made me curious how the news
got produced. The next day, I went back into Bombay to visit the
headquarters of the PTI. We drove past a police officer on a chauf-
feur-driven scooter and ended up at Flora Fountain in the center of
India’s banking district.

The Uco Bank building, where PTI was located, was a few
blocks away from where my driver left me off. He waved his arms
vaguely in the proper direction, and I headed off down the street.
Every block or so, I would stop a policeman or a rickshaw-wallah
and say “Uco Bank Building” a few times.

I happened to be clutching a few random notes in my hand and
every time I stopped to ask for directions, the person would grab
the paper out of my hand and scrutinize it for several minutes. The
notes had nothing to do with Uco Bank, but after a while the person
would look up and wave me on down the street.

This method apparently worked because I soon saw a sign for
the Uco Bank Building. I found it interesting that no matter which
paper I had in my hand, it would be grabbed and examined. With a
chuckle I thought of Cliff Lynch, one of the more active members of
the library automation community. Cliff is famous for always
clutching a stack several inches thick of business cards and scraps of
paper. It would have taken him all day to get directions.

I went up the rickety stairs to the PTI offices and met Gourang
Kundapur. Over cups of sweet coffee, he told me the history of PTI.
Formed as a cooperative for small and medium newspapers in the
early part of the century, the service was taken over by Reuters be-
fore the war. When the British left in August 1947, Reuters also
pulled out and PTI was formed as a non-profit cooperative.

The news is gathered by over 400 journalists spread in 135 of-
fices throughout the country. Before automation began, stories were
typed on a teleprinter and punched to paper tape. The paper tape
was fed in and the data went over a 50-baud line to one of four
main centers.

In 1980, when Ramani began acting as a consultant to PIT, the
entire network of 100,000 km of 50 baud lines was based on manual
switching. The data came into a regional center and a tape was

349

Exploring the Internet

punched. A subeditor edited the story, and a new tape was be pro-
duced.

The tape then went to the transmission room, where an operator
flipped switches to indicate which lines should be active, and the
story was sent out. If the story had wide enough distribution, re-
gional centers got the data, punched a tape, and sent it out to their
local clients.

The obvious places to computerize were the four regional of-
fices. PDP 11 systems were installed with custom code to terminate
the teleprinter lines, acting as a switch and also providing an online
editing environment for editors.

Connecting teleprinter lines turned out to be no mean feat. Each
line transmitted at a slightly different speed. Custom boxes were
developed that adjusted the speed to an exact 50 baud, manually at
first and later automatically. The code from the PDP system was
later ported to Xenix, with the Xenix system acting as an editing
environment and a backup to the message switching function of the
Pe.

While the regional offices were fairly well automated, Mr. Kun-
dapur explained that it was not quite as easy to get rid of the tele-
printers or even the 50 baud lines. The 50 baud lines were available
to the press at 1/6 of the cost of normal lines, and, while it was
possible to drive the lines at higher speeds locally, this would not
work in remote areas.

People were another consideration. Electromechanical printers
were noisy and dirty, but they had been used so long that people
knew how they worked. Old-time journalists were not quick to
adapt to electronic teleprinters (i.e., a dot matrix printer and a key-
board), although the mechanics tended to love the new systems as
they required air conditioning.

For larger sites, teleprinters had been replaced with PCs or Atex
systems. A custom device was built by a PTI subsidiary which ad-
justed the incoming voltage, changed the speed to 300 bps, and con-
verted 5-bit BAUDOT code to ASCII, allowing most computers to
easily accept a teleprinter feed.

PTI didn’t have the most hi-tech system in the world, but it il-
lustrated how to work in a technically challenging environment and

350

Bombay

still get work done. The system was continuing to move forward,
with the central code being ported to an 80486 running UNIX and
TCP/IP to link regional centers being investigated.

I got back to NCST in time for a quick lunch of baingan ka
bhartha, a delicious eggplant dish similar to the Middle Eastern
baba ganouj, accompanied by flat chapati bread. That afternoon
and the next day I spent giving lectures on topics as diverse as
WAIS and high speed networking and meeting with NCST students
and staff. Saturday, Ramani and I got together to talk about stand-
ards.

Ramani was chair of a key committee in the Bureau of Indian
Standards (BIS). One of the challenges in India was training large
numbers of engineers, both for developing local systems and to
stimulate the growing software export industry.

Training engineers meant giving them the information they
needed and that meant giving them access to standards. BIS had
been grappling for several years on how to distribute information,
and online standards was high on their list. Ramani’s committee
had already decided that dissemination of information was a key
priority.

Would Ramani be interested in posting standards for the Indian
community? “In a minute,” Ramani said without any hesitation at

all.

This was the second place that felt that training people was
much more important than playing international politics, and it was
obvious that many other countries would post standards if the data
were available.

It was obvious to me that I should get off the road and get busy
preparing data, a task I couldn’t do very well from an airplane. Be-
fore I could go home, though, I had one more stop to make.

I reported to Bombay’s international airport for my 3:30 A.M.
flight to London, Chicago, and Madison, Wisconsin. After seven
stamps on my boarding pass, three in my passport, a special line
and three stamps to clear my computer, a half-dozen luggage
checks, and several lines with no apparent purpose, I was on my
Way.

351

Madison

While killing time in the lounge at Heathrow waiting for my plane
to Chicago, a group of three couples sat down next to me and
started playing dominoes. Obviously on a tour of some sort, they
played their game and babbled on about some market they had vis-
ited in Cairo.

A few minutes later, a young lady bearing a name badge with
the title “Travel Coordinator” came running up, all out of breath.

“T almost didn’t make it,” she told them, “I was on the wait list.”

“Oh dear!” one of her charges commiserated. “Was your lug-
gage too heavy?”

Monday morning, I felt great despite my 12-hour jet lag and 26-hour
trip. This was my last stop, my 56th city since INTEROP 91 Fall,
less than 6 months ago, and I was SSA looking forward to ter-
minating this little journey.

I had come to Madison to learn how the Internet had gone from
a DoD-centered ARPANET to a collection of autonomous systems in
the early 1980s. ‘To learn this ancient lore, I had come to meet Larry
Landweber, one of the key figures in the early development of the
Internet.

In India, some of us had been sitting around over a beer specu-
lating on the topic of Larry’s age. He was a fairly prominent theore-
tician before his entry into networking, and the consensus was that
Lah-Ree (as his name is pronounced in much of the world) must
easily be in his 60s and close to retirement.

When I mentioned this little anecdote to Larry, he quickly estab-
lished that he was well under 50 and had no immediate plans to
retire. It’s always good to start an interview by putting your subject
in a pleasant, relaxed state of mind.

We walked down to the computer room in the Computer Sci-
ence department for a quick tour. Computer Science at Wisconsin is

352

Madison

large, with over 35 faculty, and extremely well equipped. The ma-
chine room is huge and includes such sports computers as a Think-
ing Machines CM5 and a now ancient 32-node HyperCube, each
node equipped with a 300 Mbyte disk drive.

As we walked past old VAX 11/750s, kept in place so the uni-
versity auditors didn’t reclaim machine room space, we stumbled
across two brand new Silicon Graphics computers. Larry called
over the computer room manager, who was busy trying to get some
equipment packed and shipped back to the manufacturer.

“Whose SGI machines are those?” Larry asked.
“Those are yours, Larry,” the manager said with a patient smile.

At the corporate offices of Me, Inc., the arrival of a new RS-232
cable is enough to stop work for a celebration, so I couldn’t imagine
having so much equipment you forget about two Silicon Graphics
machines.

Outside the machine room, Larry logged onto a little terminal
next to a coke machine and typed in his request for a Diet Coke,
which soon came slamming out of the slot. I commented on how
clever it was to rig up the machine to a host.

Turned out the reason for this automated soda server was fairly
practical. The campus administration had passed some sort of rule
that would have removed the machine from its convenient location
outside the machine room. By hooking it up to the network, a mere
soda server became a full-fledged example of computer science re-
search. The machine stayed.

Back upstairs, we sat down for a history lesson. Larry’s first
exposure to networking came in the summer of 1977 when he was
sitting in a bar with a few other theoreticians lamenting the lack of
communications among themselves. They dreamed up the idea of
TheoryNet for exchanging e-mail and, luckily, an NSF official was
sitting with them.

“Give me a proposal,” he said.

Next thing he knew, Larry had a grant for U.S. $136,000, which
for a theoretician was a lot of money, far surpassing any previous
grant he had received. Soon, theoreticians all over the country were
happily banging away on their Texas Instruments paper terminals,

353

Exploring the Internet

exchanging missives on NP completeness and computational com-
plexity.

In 1977, Larry became chair of the department and started look-
ing around for ways to improve computing facilities. He began
casting covetous glances at the ARPANET, but Wisconsin didn’t
have nearly enough defense contracts to justify its inclusion in the
elite dozen or so nodes.

Wisconsin certainly wasn’t the only school not on the ARPANET,
and Larry started looking around for ways to get the have-nots on
the network. He was primarily interested in Computer Science de-
partments, a logical focus given the level of penetration of networks
at the time.

Kent Curtis was director of the Computer Science section of NSF
and was quickly sold on the idea of a CSNET. A small group of
computer scientists was convened in May 1979 to discuss the idea.
First on the agenda was to see if maybe they could just get into the
ARPANET. While the ARPANET had plans to grow, it wasn’t clear
that it would grow so fast that it would include them. Just in case,
Kent Curtis encouraged the group to submit a proposal.

In November 1979 the proposal for CSNET was submitted and
went out for review. The results were disastrous. Reviewers
thought TCP/IP was far too complex for end-users, networking
should be left to the ARPANET professionals, and who did these
amateurs think they were anyway?

Not exactly a vote of confidence. Kent Curtis looked at the re-
views, thought about it, and sent Larry back with U.S. $49,000 to
make a study. By June 1981, a second proposal for CSNET was sub-
mitted and the reviews were again negative.

Kent Curtis looked at the reviews, and called an advisory panel
together. They didn’t like the idea, either. Curtis kept on support-
ing the idea though. What made CSNET possible was the skillful
lobbying by Kent Curtis inside the government and the cooperation
of DARPA. Vinton Cerf had attended a CSNET planning meeting
and indicated that DARPA looked on the project favorably and
would be willing to work with CSNET on ways to interconnect.
January 1980, after some fairly intense lobbying, NSF awarded U.S.
$5 million for five years to the CSNET project.

354

Madison

Armed with money and the blessings of NSF and DARPA, a
group of five people was put together to manage the project. A
classic exercise in distributed management, CSNET was run by
Larry Landweber, Peter Denning, Dave Farber, and Tony Hearn,
along with Bill Kern, the NSF project manager. The project began
the task of putting together a comparatively low-cost solution that
would allow a sharp increase in the size of the Internet.

One of the immediate effects of the NSF award was to get Madi-
son and the other key development sites onto the ARPANET. These
sites—Madison, Purdue, RAND, and Delaware—formed the core of
CSNET. For the rest of the CSNET world, there were three levels of
connectivity: exchanging mail, using a service host, and full TCP/IP
connectivity.

For exchanging mail, CSNET used a technique known as Phone-
Net. PhoneNet used the MMDF software developed by Dave
Crocker and Dave Farber to transfer messages over telephone lines,
much in the same way UUCP did.

There was a big difference from UUCP, however. Mail which
arrived at the PhoneNet relays at RAND and the University of Dela-
ware could continue on into the ARPANET. DARPA had agreed
that the CSNET sites, a group of hosts outside of their direct admin-
istrative control, could exchange mail with ARPANET sites. This
was thus the first example of two autonomous systems connected
together with gateways. Although the initial exchanges were e-mail
only, this was soon expanded to IP connectivity.

The agreement between ARPANET and CSNET to allow connec-
tion was a carefully worded agreement, prohibiting sites in the
CSNET from forwarding data originating from the ARPANET to
“non-authorized users.” Most importantly, though, the agreement
delegated authority to the CSNET to administer its own network,
omly requiring that the members of CSNET abide by the rules of
acceptable use of the network.

The issue of charging was also carefully dealt with, specifying
that there would be no charges between the autonomous systems:
you accept my traffic and I'll accept yours. Lack of settlement pro-
cedures has remained a tradition in the Internet even today. The
Commercial Internet Exchange, for example, uses a system of flat

355

Exploring the Internet

fees based on the size of pipes, not on the amount or attributes of
the data flowing through those pipes.

PhoneNet allowed message exchange within CSNET and to the
ARPANET, but didn’t give people access to services like Telnet and
FTP. To provide this higher level of connectivity, CSNET provided a
service host at Madison.

Users were given individual accounts and used X.25 or dialin to
log onto the service host. From there, they had full access to the
ARPANET. The service host also held a collection of useful docu-
ments and some nameserver software to help users find electronic
mail addresses.

PhoneNet and the service host caught on like wildfire. By De-
cember 1981, the RAND and Delaware relays were up and running
and Purdue, Princeton, and the NSF were PhoneNet clients. By the
summer of 1982, 24 sites were on PhoneNet and two years later
there were 84 sites. The population of PhoneNet peaked in 1988 at
170 sites.

Getting sites fully integrated into the ARPANET with IP access
was a bit tougher. At the time, X.25 was viewed as the way to set
up widely dispersed hosts and the common carriers were telling
people that in the future, X.25 networks would become cheap and
leased lines expensive.

There were two well-known public data networks at the time,
Tymnet and Telenet. Tymnet didn’t return phone calls from CSNET,
so Telenet became the candidate network for interconnecting CSNET
hosts at the IP level. (Note the difference between the Telenet X.25
network and the Telnet virtual terminal service.)

All the X.25 networks at the time were optimized for the holy X
trinity, X.3, X.28, and X.29. These standards defined how an asyn-
chronous terminal could connect to a Packet Assembler/Disassem-
bler, which in turn set up virtual circuits to hosts on the X.25
network.

Optimizing for terminals made it tough on hosts. Telenet used
128-byte packets and a virtual circuit could only have two packets
outstanding. This certainly slowed down potential throughput, but
to make matters worse, the data link level of LAP-B enforced a limit
of 7 outstanding packets for all virtual circuits sharing a link.

356

Madison

What this meant was that a host with a 9,600 bps link to Telenet
could only get 1 to 2 kilobits of throughput on one virtual circuit.
This was a best-case scenario, and during peak periods congestion
reduced the throughput even further.

Douglas Comer of Purdue took on the daunting task of trying to
integrate X.25 and TCP/IP. Comer and his group developed a solu-
tion using multiple X.25 circuits, one of the first examples of path
splitting. Comer’s group also developed the software that set up
and tore down virtual circuits in a manner that was transparent to
IP.

Comer’s group mapped datagrams over virtual circuits. The ap-
proach of layering connectionless traffic over connection-oriented
circuits was considered controversial at the time. Another group at
University College London took a different approach, using protocol
translation to turn TCP circuits into X.25 circuits. Comer’s group
went off and hacked, and came back with a working IP over X.25
demonstration between Madison and Purdue.

Even with this clever software, X.25-based networking never re-
ally caught on in the U.S. Not only was performance poor, but the
cost for running TCP/IP over X.25 could be astronomical. Remem-
ber, these networks charged by the packet and a single character
typed for a Telnet session could consume 4 packets (send character,
acknowledge, echo character, acknowledge).

Even though the cost was high, a few sites wanted IP access
badly enough that they were willing to pay. After Purdue built and
tested the software, Rice was the first X.25 customer in the fall of
1984, and was soon joined by DEC’s Western Research Laboratory,
and BBN. Later, as it became clear that leased lines were more eco-
nomical, the Purdue group developed a low-cost leased line net-
work called Cypress, which was offered as an alternative access
path in CSNET.

BBN also became the site for the new Coordination and Infor-
mation Center (CIC) and became the support center for CSNET.
The service host and MMDF relays were moved, and the CIC main-
tained the document database and helped new users get started.

Although NSF funded CSNET, they also made sure that a clause
was inserted in the project plan that the network should be self suf-

357

Exploring the Internet

ficient at the end of five years. At the end of five years, it was in
fact self sufficient, charging universities U.S. $2,000 to $5,000 per
year and industrial sites such as DEC and IBM $30,000 to partici-
pate. Self-sufficiency of CSNET was used by the NSF as an impor-
tant justification for funding future projects such as the NSFNET.

In addition to self-sufficiency, CSNET helped establish a tradi-
tion in the U.S. of projects of limited scope. Instead of expanding
CSNET to be a backbone, to serve other communities, and to
achieve world peace, CSNET remained focused on one problem.
Over time, other solutions such as regional networks became more
attractive and by 1988 the number of CSNET hosts was beginning to
decline. By 1989, CSNET had merged with BlITnet and by 1991,
PhoneNet was dead.

CSNET was not just limited to the U.S. It formed an important
way, along with BITNET, to spread the Internet overseas. The inter-
national spread of the Internet was directly linked to an annual se-
ries of meetings that came to be known as the Landweber Seminars.
I was interested to learn that Larry was not even present at the first
Landweber Seminar.

In September 1982, Peter Kirstein of University College London
convened a “meeting on U.S.-European Academic Network Collabo-
ration.” The meeting was small, but included representatives from a
half-dozen countries.

At this meeting, Germany described their DFN project, the Scan-
dinavians described active projects in Norway and Sweden, and the
U.K. described the Coloured Book. There was a CSNET presenta-
tion, and CERN described the early stages of what would become
HEPnet.

In other words, this was a collection of people who were actu-
ally doing something about getting together to exchange informa-
tion. One of the most useful functions of such a meeting was what
occurred in the halls and at dinner where synergy between different
groups quickly led to new projects.

In 1983, Kirstein and Landweber jointly organized another semi-
nar in Oslo at which Larry announced that CSNET was going to go
international. Needless to say, the question of where it would go
international first occupied much of the discussions at the breaks.

358

Madison

In February 1984, Israel became the first international member of
CSNET. A carefully worded agreement was signed stating that data
originating from the ARPANET would not be forwarded. Interna-
tional access, after all, had the potential to become a political issue.

The Israel PhoneNet connection was rapidly followed by Korea,
with Professor Kilnam Chon aggressively pursuing connectivity. By
September 1986, PhoneNet access had spread to Australia, Canada,
France, Germany, and Japan. Israel, Korea, and the Japanese NTT
soon got IP access.

The Landweber Seminars became firmly established as a place
for people to meet and things like CSNET connectivity, gateways,
and other projects to get started. In 1984, the Seminar convened in
Paris, followed in subsequent years by Stockholm, Dublin, Prince-
ton, Jerusalem, and Sydney.

Princeton, in 1987, was the breakthrough meeting. The invita-
tion-only seminar had reached 100 people and the attendance list
was certainly impressive, including most of the people active in
spreading the Internet.

By Sydney, the meeting had grown so big that the seminar for-
mat began to outlive its usefulness. It was decided that year that
the seminar should metamorphosize into a conference, and the next
year INET 91 was held in Copenhagen. At INET 92, in Kobe, Japan,
the conference turned into the annual meeting of the Internet Soci-

ety.
0

After a nap to shake off jet lag, I drove across town for dinner with
Larry and his wife, Jean. Also there was Tony Hearn, one of the
CSNET founders, who was driving through on a college inspection
tour with his son and had stopped for the night in Madison.

We had a nice dinner of vegetarian lasagna and steaming home-
made rye bread, finished by strawberries with a soy powder cus-
tard. When the second bottle of wine was cracked, a 1969 Pinot
Noir, Tony Hern sniffed his glass appreciatively and smiled mischie-
vously.

359

Exploring the Internet

“Gee, this would taste great with meat,” he teased. “Honey,”
Larry said, turning to his wife, “do we have any cow left?”

“1 think we have a slab in the freezer.”

I began to fade from jet lag and went back to my hotel, anxious
not to miss my early flight the next morning back to Boulder.

360

Boulder

It was Tuesday afternoon, and I was home. I wandered over to the
Ideal Market, the kind of place that has a “customer-of-the-month,”
to get myself a bagel with lox and cream cheese.

“Have a nice day,” the bag person said to me as I walked in the
front door.

“Have a nice day,” the deli person said, handing me my bagel.

“Have a nice day,” said Dennis, the chi-chi butcher, as he va-
cated his table to go back to selling ground buffalo and turkey brat-
wurst.

That did it. Enough was enough. I went over to the pay phone
and called my realtor. Minutes later the formalities were complete
and my house was on the market.

Before I could leave Boulder, however, I had to finish this book.
Mind you, this was not out of some sense of duty or an overwhelm-
ing desire to clean up loose ends. Half of my advance was payable
when I turned in a manuscript suitable for publication, and I
needed the money to move.

I did what many Boulder residents do during those rare mo-
ments when they have to concentrate, and went to the Trident Cafe.
Sitting down, I looked around.

At the next table, a man with unfocused eyes was flipping
through a book of star charts, trying to find astrologically similar
periods in the past to shed light on the future. In a booth nearby, a
guy was dealing out Tarot cards, stopping periodically to stare at his
destiny and grumble to himself. The table next to him was spread
out with materials for advanced Shambhala training, and next to
that three people were engaged in some strange form of group mas-
sage.

With so many people looking for inner truth and sharing their
feelings, I decided to try and list the good qualities of OSI. After
thirty minutes of protocol meditation, however, I looked down at
my piece of paper. It contained only one item.

“We have to live with it,” my list read.

361

Exploring the Internet

The goals of the OSI effort are certainly laudable. The concept
of open systems is about as hard to argue against as any other lofty
ideal. But, saying that being against OSI is to be against open sys-
tems assumes there is only one road to truth, beauty, and interoper-
ability.

Quite simply, OSI has not achieved anything close to its goals.
Too much complexity, lack of hands-on experience, and far too
many fine lunches and dinners had hindered the creation of a lean,
mean, finite-state architectural machine.

Just because OSI is not ideal, however, doesn’t mean that it can
be ignored. OSI has been adopted by many governments as a low-
est common denominator of connectivity, taking its place as the logi-
cal successor to IBM’s RJE and Bisynch protocols.

OSI has become one of those challenges that keep system inte-
grators and consultants employed. OSI is just one of those ugly
things you have to deal with when putting together a system, no
different than COBOL-driven ISAM files, sendmail configurations,
or VIAM routing tables. In fact, OSI has become one of those
things that made ideal fodder for students taking MIS classes in
business schools:

“Final Exam, Essay Question: Finance has OSI,
Accounting has SNA, and Personnel uses
Novell. FIAM over LAN Manager has been
proposed for an enterprise-wide file system.
Discuss. Extra Credit: Write a justification for
the proposition that NetWare is actually compli-
ant with OSI.”

While OSI has not achieved the goals the committees have set
for it, neither was the effort totally wasted. Certainly the careers of
hundreds and hundreds of standards potatoes could have been bet-
ter spent, but there are a few pieces that can be salvaged.

The only way to salvage the work, however, is to get the process
out of the hands of those who would build machines in the abstract
and back to those who are engaged in the hard, dirty work of build-
ing networks. To salvage anything about OSI, the standards have to
be known or the work will simply be ignored.

362

Boulder

Just because the approach to building networks by committee
didn’t work doesn’t mean that lone wolves can do any better.
Rather, a special type of person is needed, one who understands
people well enough to mobilize resources and understands the tech-
nology well enough to know which resources to mobilize.

In my journey, I met many people who effectively built national
networks. Geoff Huston in Australia, Juha Heinanen, in Finland,
Jun Murai in Japan, Glenn Kowack from EUnet, and Rick Adams
from UUnet are only a few examples of people who successfully
built large networks, providing the leadership necessary to turn
technology into infrastructure.

While there are many people who succeeded, there are many,
many more who spent their time building castles out of air. Efforts
like the COSINE project in Europe, for example, kept many people
busy making presentations instead of stringing cable.

The ratio of goers to doers in the world of research is just as
lopsided. For every Marshall T. Rose, Steve Hardcastle-Kille, or
Christian Huitema, there are a hundred people arguing about what
to call the TCP/IP protocol suite instead of writing code.

The trick appears to be to provide an environment which allows
people to do useful work, and then letting them get on with it. Part
of that means providing appropriate resources, but resources alone
don’t guarantee anything. There are too many multimillion dollar
mega-projects that went nowhere to think that throwing money at a
problem has any relationship to the final outcome.

In fact, the contrary appears to be the case. There are many pro-
jects that started on a shoestring, with a few researchers borrowing
a modem and appropriating a phone line. The common denomina-
tor in most of these projects was a desire to get work done and a
fascination with the technology.

An ad hoc approach to building networks convinced the stand-
ards potatoes that the Internet was some kind of academic toy, un-
suitable for doing real things. They were wrong and the toy turned
out to be the piles of OSI paper.

The Internet is a machine, built by thousands and thousands of
engineers to solve real problems. So how can this machine be
turned into a truly global infrastructure? How can the spirit of in-

363

Exploring the Internet

ventiveness that had built the machine in the first place be pre-
served?

One can think of infrastructure in layers, and it seems that most
of the failed efforts can be traced to a tendency to concentrate on
one layer to the exclusion of others. As they would say in Boulder
(if they thought about it), infrastructure is holistic.

At one layer, you have the real things, the technology that makes
networks. Protocols, hardware, and all the technical details that are
studied in engineering schools are all part of the real layer.

Technology alone doesn’t make a network, though. The next
layer is the people layer where technology is applied, deployed, and
networks start being used. The leadership of the NSF, DARPA, and
the IAB are all functions at the people layer.

Don’t think that you can do one layer without the other. NSF
provided money and leadership, but the NSFNET also required the
solution of many technical problems before it became a reality.
Likewise, there are many great technologies that remain research
prototypes because nobody takes the time to move them into the
field.

One can think of the people layer as the process of governance,
but that does not imply government. EUnet and the USENET, for
example, are strong evidence that resources can be organized and
deployed without the involvement of official bureaucracies.

Eventually, we reach the paper layer, where things are docu-
mented. One hopes, of course, that the paper is actually some form
of data online. Documentation is the key to allowing the activities
in one part to spread to other communities.

Too often, technology remains a black box. By writing down
What is happening, others can learn what goes on inside of that
black box and improve it. This is the crucial mistake that OSI made
when they invented paper, but forgot to show it to anybody.

Paper can mean standards, but it can also mean procedures and
policies. Paper has a big impact on how the network is used and
who can use it. Standards and the other forms of paper, virtual or
real, are the laws and regulations of networks.

These laws have a real effect. Is your routing protocol complex?
You've raised the cost of entry. Do you have an acceptable use pol-

364

Boulder

icy? You've limited your population. Have you invented an anony-
mous FIP mechanism and an RFC series? You’ve encouraged the
spread of the network.

Paper thus reflects something even deeper, the fundamental hu-
man values that say what we can do and how. The technology,
standards, and networks are all reflections of those values.

Infrastructure, at all levels, reflects how we apply these funda-
mental human values. Privacy, for example, can be protected or de-
stroyed by a network. There is no inevitable loss of privacy on
computer networks, but that is certainly one possible outcome.
Likewise, free speech can be encouraged, property can be protected,
and we can mold the technology to reflect what we want.

It is the coupling of an awareness of these fundamental values
with an understanding of the technology that allows us to build a
truly global infrastructure. Technically unsophisticated policy mak-
ers make bad laws, just as politically unsophisticated engineers
make bad networks.

Educating policy makers and engineers is certainly a nice, long-term
goal, but it doesn’t provide any concrete steps to be taken today.
Putting standards online, though, is a concrete step. The task is
technically simple and the benefits were clear.

The Bruno project demonstrated how trivial it is to make infor-
mation available on computer networks. That this information
should, indeed must, be available to all is, to borrow the words of
Geoff Huston in Australia, “bloody obvious.”

Yet, in pursuing this goal, I ran into a solid wall of resistance.
To the people running the process, posting standards leads to the
demise of international standards and takes money away from de-
veloping countries.

In visiting twenty countries, though, it was obvious to me that
the developing countries are the ones most hurt by the policies of
the international standards bodies. They are the ones that need
most to train engineers, to develop products for export, and to ap-
ply cheap, easily-available technology to pressing problems.

365

Exploring the Internet

Even the developed countries are not helped by policies that re-
stricted participation to a bureaucratic elite. The real networks that
I visited were built from the ground up. The most effective net-
works were built by people with problems to solve. If we want
those networks to be interoperable, certainly a goal of the standards
process, people building the networks have to be able to get the
standards.

The concept of the standards haven is one way around the bu-
reaucracies, but it is not a very productive way to solve the prob-
lem. While a few countries probably will post standards for their
engineers, nobody had confirmed their intention to do so officially.
Indeed, while Korea and India liked the standards haven concept,
the idea was caught in the mire of their own bureaucracies. Ulti-
mately, the standards haven concept is a temporary stopgap and a
more permanent solution must come directly from the standards
cartel. This is a problem crying for global leadership.

Availability of information is one of those fundamental human
values that needs to be established as a conscious decision, a funda-
mental part of the infrastructure. Bruno was a stopgap, and even if
a few people working on their own could come up with a new stop-
gap, what we need is a real solution.

Returning from the coffee shop, it was time to tackle a three-foot-tall
pile of mail that had come while I was gone. Having the SnailMAIL
to process in a batch let me have a little fun.

As I sorted my mail, I pulled out all the postage-paid cards and
put them into a stack, throwing everything else into the recycle bin.
I then performed triage on the stack, dividing them up into things I
wanted information on (a total of two cards), things I didn’t care
about, and companies and interest groups I didn’t like.

The cards for the bad organizations were left blank and depos-
ited in the nearest post box. Voting with postage-paid cards was a
form of economic democracy I learned from my politically-active
mother. Over time, individual action can snowball and even the
bulkiest bureaucracies will give way.

366

A
Abe, George, 318-19
Academic Computing Services, 134
Academy One project, 289-90
Acampora, Anthony, 141
Acnet (Accelerator Control network),
295-96
Acorn project, 141-43
applications, 142-43
architecture, 142
basis of, 142
ACSnet, 193
Adams, Rick, 363
Address space resolution, 336
Adelaide, Australia, 198-201
Advanced Network and Services
(ANS), 103, 145, 282
AFGnet, 258
Aiso, Dean, 45
Akihabara, 49-50, 55, 324-25
“Electric Town,” 324
AlterNet, 195
Amsterdam, 90-94, 233-44
Bulgarian dial-up UUCP link, 271
CWI (Centrum voor Wiskunde en
Informatica), 233-34, 240
NIKHEF (Nationaal Instituut voor
Kernfysica en Hoge-Energie-
fysica), 233, 235, 238, 292
RIPE (Réseaux IP Européens), 233-
39, 242-44, 250
Andrew File System (AFS), 37
Andrews, W. Clayton, 141-42
Appropriate Use Policy (AUP), 195
Ardis public data network, 62
ARPANET, 9, 88, 98, 252, 277-78, 280,
315, 352, 354-56, 359 ASN.1, 121-
22
Association for Computing Machinery
(ACM), 279
AT&T, 318
ATOMIC project, 313-15

INDEX

Auckland, 182-84
Wellington, 175-77
Auerbach, Karl, 31, 151
Australia:
Adelaide, 198-201
Australian AARNET, 193, 195, 245,
336
Canberra, 193-97
Melbourne, 184-87
Sydney, 188-92
Australian Academic Research Net-
work (AARNet), 193, 195, 245,
336
AUTOCAD, 23, 56

B
Bangkok, Thailand, 219-32
general ledger system, 220-21
Pantip Plaza, 230-31
Stock Exchange of Thailand (SET),
220, 222-23
Value Group, 226-28
Baobab, 284-85
Barad, Bob, 283-85
Bay Area Regional Research Network
(BARRnet), 30, 86, 173-74 BBN,
88, 315, 357
Becker, Jordan, 152
Becker, Paul, 141
Behemoth, 161-64
audio bus, 163
chord keyboard, 162
computers on, 162
cost of, 164
Global Positioning System (GPS),
164
head mouse, 162-63
helmet, 163
power bus, 163
security for, 164
Belgium, 260-68

Index

Bell, Gordon, 303
Bellcore, 304, 307
Berkeley, California, 157-60
Birkenbihl, Klaus, 255-56, 258
BITNET, 15, 38, 59, 68, 84, 86, 358
Bjelke-Petersen, Joh, 201
Blokzijl, Rob, 233-35, 237
Blue Book, The, 5, 11, 22-23, 124, 128,
130, 272
conversion of, 22-25
Bombay, India, 340-51
Bureau of Indian Standards (BIS),
351
ERNET, 341-43
NCST (National Center for Soft-
ware Training), 340-45, 348, 351
PTI (Press Trust of India), 349-51
railway reservation system, 345-47
Bond, Alan, 201
Bonn, Germany, 254-59
Border Gateway Protocol (BGP), 88, 137
Bottorff, Norman, 225
Boulder, Colorado, 21-25, 153-54, 157,
301-7, 361-66
Bratislava, Czechoslovakia, 15
Brim, Scott, 136-37
British Telecom, 71
Broadband ISDN (B-ISDN), 303, 315,
321
Bruno, Giordano, 35
Bruno project, 35, 44, 116-17, 124-33,
170, 302, 366
applying to ISO world, 130
ITU printing operation vs., 127-28
termination of, 125, 153-54
Brussels, Belgium, 260-68
Bulletin board services, Minitel and, 20
Butterfield, Julie, 6
Butterfly, The, 174

(ee

California:
Berkeley, 157-60
Marina del Rey, 311-17
Moffet Field, 171-74

Mountain View, 161-65
San Francisco, 166-70, 318-20
San Jose, 22
CAMAC crates, 294-97
Canberra, Australia, 193-97
Carpenter, Brian, 18, 93
Casner, Stephen, 315-16
CCIR, 125
CCITT, 69, 125, 330
recommendation D.1, 68
CDCNET, 186
Cedok, 13
Center for Telecommunications Re-
search (CTR), 141-44
Acorn project, 141-43
operational nature of, 143
Centrum voor Wiskunde en Infor-
matica (CWI), 233-34, 240
CERFnet, 105, 312, 329-30
Cerf, Vinton, 34, 84-85, 123, 151, 277-
79, 343, 354
CERN, 18, 91, 172, 234, 236, 250, 257,
271, 283, 292-93, 358
Chang, Michael, 60, 337
Chapin, Lyman, 123
Chatchi, General, 229
Chaundry, Chris, 185-86
Chen, Tommi, 67
CHEOPS satellite, 18
Cheung, Eric Lo Hing, 337
Chicago, 291-300
Chinese University (New Territories),
60
Cho, Mr., 65-66
Chon, Kilnam, 329-32, 359
Cisco routers, 17, 102, 105, 194, 235,
2, 275
CITI (Columbia Institute for Tele-Infor-
mation), 143-44
Private Networks and Public Objec-
tives project, 144
City Lights bookstore, 159
Clark, David, 304-5
Clasper, Harry, 93-94
Cleveland, Ohio, 286-90
Cleveland Free-Net, 286-89, 348
CLNS, 107

CNET, 120

CNRS (Centre National de la Recher-
che Scientifique), 273

Cody’s bookstore, 158

Cohen, Danny, 313

Colorado, Boulder, 21-25, 153-54

Colorado SuperNet, 270

Columbia University, New York:

CITI (Columbia Institute for Tele-In-

formation), 143-44
CTR (Center for Telecommunica-
tions Research), 141-44
Comer, Douglas, 34, 357
Commercial Internet Exchange (CIX),
Wisp aha
Communications Week International, 108,
110, 131, 154, 302
Computer Literacy bookstore, 159-60
Computer Sciences Corporation (CSC),
318
Connection Oriented Network Service
(CONS), 90
ConneXions journal, 3, 33, 165, 319
Cooper, Ann, 317
Copyright protection:
of ITU documents, 129-31
Value Group, 227-28
Cornell University, 136-40
CUINFO, 138-39
Gateway Daemon (gated), 137-38
routing protocols, work on, 136
Corporation for National Research In-

itiatives (CNRI), 81, 277, 279, 303-

4
COSINE project, 248-50, 261-64, 363
budget, 249
IXI (International X.25 Infrastruc-
ture), 249-50, 263
origin of, 248
Project Management Unit (PMU),
249
Crocker, Dave, 355
Crow, Trammel, 301
Crowcroft, Jon, 99
CTR, See Center for Telecommunica-
tions Research (CTR)
CSNET, 84, 354, 355-59

Index

CUINFO, 138-39
Curtis, Kent, 354
CWI, See Centrum voor Wiskunde en
Informatica (CWI)
Cypress network, 357
Czechoslovakia:
Bratislava, 15
Prague, 13-16
Czech Technical University, 13-14

DARPA, 268, 278, 280, 303, 311, 312,
315, 354, 364

Data Communications, 3

Datanet, 104-6

Datapak, 63

DEC LAT protocols, 213

DECnet, 17, 39, 40, 85, 92, 104, 105,
186, 257, 271, 293-94

DECnet Phase IV, 172

DECnet Phase V, 55, 110, 172

Deering, Stephen, 315

Delp, Gary, 305

Denning, Peter, 355

desJardins, Richard, 34-35

Develnet, 62

DEN (Deutsches Forschungsnetz) pro-
ject, 257-58, 263, 274, 358

DG XIII, 89, 261-64, 267

Dial-up UUCP, 271, 343

Digital Equipment Corporation (DEC),
73, 150, 151, 295, 304, 358

Western Research Laboratory, 357

Digital Resource Institute, 116

Directory assistance, Minitel and, 19

Directory User Agent, X.500, 96-97

DISH, 96

$Link, Singapore Network Services
(SNS), 75

Domain Name System (DNS), 97-98,
147, 317

DSA, 95-98

Dublin, 83-89

Dunedin, New Zealand, 178-81

D Zero, 297-99

Index

EASInet, 93, 250, 256-57
EBCDIC character set, 24
EBONE (European Backbone), 89, 90-
94, 240, 250, 257, 283
contributions to, 92-93
DECnet Phase IV traffic and, 92
EBONE Action Team, 93, 250
management committee, 93-94
Memorandum of Understanding,
250-51
EDI, 74, 204
Eicher, Larry, 117, 130-31
Elamvazuthi, Chandron, 210, 212
Electronic mail, in outlying areas, 115
Elz, Robert, 186, 193
England, 95-99
ENIAC, 14
Epilogue Technology, 31, 198, 200
Eppenberger, Urs, 17
ERNET, 341-43
floppy-based e-mail system, 342
infrastructure, 342-43
ESnet (Energy Sciences network), 174
HEPnet and, 293
Esprit, 262-64, 268
Estrada, Susan, 312
Ethernets, 17, 18, 54, 62, 73, 120, 126,
306, 324, 348
EUnet, 15, 45, 233, 239-42, 251, 257, 364
EBONE consortium and, 242
Eureka group, COSINE project and,
248-49
European Academic Research Net-
work (EARN), 15, 84, 247, 252,
256, 258, 271
European Academic Supercomputer In-
itiative (EASI), 256
European Commission, 268, 272, 276
DG XIII, 261-64, 267
RARE contributions, 247-48
European Council of Telecommunica-
tions User Associations (EC-
TUA), 252
European Internet, 268
European Nervous System, 261

European Telecommunications Stand-
ards Institute (ETSI), 119, 252

European UNIX Users Group (EUUG),
240-42

European Workshop on Open Systems
(EWOS), 252

EurOpen, 241-42

Exterior Gateway Protocol (EGP), 88,
137

Fair, Erik, 317

Farber, David, 84, 302-4, 306, 355

Faul, Stephanie, 280-81

FDDI, 18, 30, 40, 69, 73, 122, 197, 294,

333
Federal Emergency Management
Agency (FEMA), 244

Federal Express, 62, 67, 231

Feist, Jerry, 138

Felderman, Robert, 313-14

Fermilab, 18, 291-97
CAMAC crates, 294-97
Linear Accelerator (LINAC), 294
network architecture, 294
networking at, 296

Fidonet, 284-85

Finland, 100-107

Finn, Gregory, 313-14

Firefighters, 4

FIX-East, 173

FIX-West, 38, 173, 195

Flaming, 4

Flat-rate pricing, 63, 70

Fluckiger, Francois, 18

Forrester, Ian, 178-81

Frame Relay, 32, 105-6

France:
France Telecom, 19, 273-75
Nice, 119-23
Paris, 19-20, 108-15, 269-76

FRED, 96

Free-Net, 286-89, 348

Fry’s, 160

FTAM, 17, 107, 264, 332

370

FTP Software, 31, 34

Fuchs, Ira, 84, 303

Fujisawa, 38, 43-48

Fujitsu, Value Added Group, 41

Full duplex, definition of, 189

FUNET (Finnish University Network),
104-5

Fuzzball, 88

Garg, Anil, 343

gated software, 137-38

Gates, Bill, 100

Gateway Daemon, 137

Geneva, 3-12, 18, 116-18, 124-33

Germany, Bonn, 254-59

Gigabit testbeds, rearch conducted in,

307

GIPSI, 120

GMD, 255-58
DEN (Deutsches Forschungsnetz)

project, 257-58, 263, 274, 358

EASInet, 93, 250, 256-57

GobNet, 306

Golden Shopping Center, 55-58, 231
Macrosoft stall, 56-57

Goldstein, Steve, 281-82

GOSIP Institute, 35

Goto, Shigeki, 323

Graphics, standards and, 130

Gross, Phill, 149

Grundner, Tom, 286-89

Gruntorad, Jan, 13-15

Hackett, Simon, 198-201, 316
Haleiwa, 37

Halliday, Bob, 224-25, 231, 232
Hanafi, Waleed, 60-63

Happy Valley, 51-52, 63
Hardcastle-Kille, Steve, 95, 363
Harju, Raimo, 100

Hawaii, Honolulu, 36-37
HEAnet, 84

Index

Hearn, Tony, 355
Heinanen, Juha, 103-6, 318, 363
Helsinki Free-Net, 289
HEPnet, 39, 172, 271, 292-93
ESnet (Energy Sciences network)
and, 293
growth of, 293
Hermes project, 288
High technology, in outlying areas, 115
Hill, Robin, 200-201
Hitachi notebook, 49-50
Hoat, Ang Chong, 204
Hong, Yeo Ning, 74
Hong Kong, 51-64, 219, 337-39
Golden Shopping Center, 55-58, 231
Hong Kong University, 59-60
Hutchinson Mobile Data, 60-62
Jockey Club, 51-55, 62
Honig, Jeff, 136-37
Honolulu, 36-37
Huh, Moon-Haeng, 333
Huitema, Christian, 119-23, 278, 363
Huston, Geoff, 193-97, 363, 365
Hutchinson Mobile Data (Hong Kong),
60-62

I
IBM 9000, 74-75
IBM, 93-94, 145, 256, 283, 304, 318, 358
PARIS project, 304
IDNet, 79
IDPR/IDRP, 137
IETF, 128, 147-52, 196, 235, 278-79, 336
certified protocol engineers board,
proposal for, 148-49
MIB (management information
base), 150-51
standards, 148
Steering Group (IESG), 150, 196
working groups, 148
India, Bombay, 340-51
Indian Institute of Technology (IIT), 343
INET, 278-79, 359
Infolan, 105-6
Infonet, 319

Index

Information Communication Institute
of Singapore (ICIS), 69-70
Information Sciences Institute (ISI), 311-
16
ATOMIC Project, 313-15
CERFnet, 312
DARTnet, 312-13
MOSIS service, 311
multimedia conferencing, 315-16
Terrestrial Wideband (TWB) net-
work, 312-13, 315
Infrastructure, 365
Ingres, 56, 243
INMARSAT, 71-72
INRIA, 119-23, 255, 269-71, 273, 282
ASN.1 benchmark, 122
computer network project, 120-21
MAVROS compiler, 121-22
robotics projects, 119-20
The Obviously Required Name
Server (THORN), 121
X.400 project, 121
Institut National de la Recherche en In-
formatique et Automatique, See
INRIA
Intercontinental Research Networking
(CCIRN), 252
International Direct Dial (IDD) links, 72
International standards documents, 3-
12
availability of, 3-5
copyright protection, lack of, 4
International X.25 Infrastructure (IXI),
249-50, 263
Internet, 3-5, 8, 18, 97, 99, 109, 125- 26,
128, 169, 177, 195-96
Finland, 103
France, 112
installing at San Jose Convention
Center, 29-35
Internet Activities Board (IAB), 11,
97, 149-51, 196, 277-78, 315
Internet Assigned Numbers Author-
ity (IANA), 315
Internet Monthly Report, 317
Japan, 39
Managers Guide, 211

372

in outlying areas, 115
policy routing protocols, 106
Request for Comments (RFCs), 3
subnetwork support, 106
Internet Society, 196-97, 252, 278-79
Internode Systems, 198
INTEROP, 22, 25, 29-36, 130, 165, 198,
352
Ireland, Dublin, 83-89
Irish Universities Network, 83-84
ISDN, 47, 77, 109, 235, 264-65, 303, 322,
324
Singapore, 71
ISDN Karaoke Machine, 323
Ishida, Haruhisa, 38-40
ISI, See Information Sciences Institute
ISODE (ISO Development Environ-
ment), 95-96, 148, 248
ISO (International Organization for
Standardization), 8-12, 130, 319-
21
standards, 332
IT2000, 77-78, 80-81
Ithaca, 134-40
ITU (International Telecommunication
- Union), 3-12, 20-25, 44, 116-18,
124-25, 153-54, 170, 272, 330-31
Blue Book, The, 5, 11, 22-25, 124,
128, 130, 272
Consultative Committee on Interna-
tional Telephone and Telegraph,
5
copyright protection of documents,
129-31
Information Systems Steering
Group, 124-25, 129
lack of computer literacy at, 6-7
mail system, 131-32
printing department, 127-28

Jacobsen, Ole, 33, 165, 319-20

JAIN, 39

JANET Coloured Book protocols, 84, 96
Jansen, Peter, 225

Index

Japan:
Akihabara, 49-50
Fujisawa, 38, 43-48
ISDN and, 47
KEK, 292
networks in, 38-39
Tokyo, 38-42, 321-28
JARING network, 211
Jennings, Dennis, 83-90, 92-94, 122, 145
Jensen, Mike, 284
Jipguep, Mr., 117
Jockey Club (Hong Kong), 51-55, 62
Customer Information Terminal
(CIT), 53-54
Telebet system, 53-54, 62
vendor/ technology policy, 54
Johnston, Peter, 261
Joint Network Team (JNT), 96, 282
JUNET, 39, 45-46
Junsec, 44, 48
JVNCnet, 335

Kahle, Brewster, 166-70

Kahn, Bob, 277, 280, 303-4, 315

Karrenberg, Daniel, 234

Keio University (Fujisawa campus), 39,
43-48

Kent, Stephen, 123

Kerberos, 33

Kern, Bill, 355

Kim II Sung University (Pyongyang),
66

King, Ken, 139-40

Kirstein, Peter, 98-99, 358

Knowbots, 81

Knowles, Stev, 31, 148

Korea, 38, 329-34

Korea Telecom, 331-34

Kowack, Glenn, 239, 241-42, 363

Kowloon, 60

KREN (Korea Research Network), 330

KREOnet (Korea Research Environ-
ment Open Network), 329-30

Kuala Lumpur, Malaysia, 209-18

MIMOS (Malaysian Institute of Mi-
croelectronic Systems), 209-12,
215
Plus project, 212-13
Tenaga Nasional Berhad, 215-17
Kummerfeld, Bob, 188-92, 197
Kundapur, Gourang, 349
Kuo, Frank, 85

Lambert, Carol, 134, 136, 139

Lambert, H. Dave, 134-36, 139

Landweber, Jean, 359-60

Landweber, Larry, 45, 84-85, 122, 278,
352-54, 355, 358-60

Landweber Seminars, 358-59

Lanlink, 104-5

LANs (Local Area Networks), 305-6

Large Electron-Positron (LEP), 18

LAT, 73, 294

Lauck, Tony, 85, 150

Lauder, Piers, 188-90

Leased lines, 15, 17, 90, 101-2, 193

Le Guigner, Jean-Paul, 273-75

Leiner, Barry, 319-20

Linear Accelerator (LINAC), 294

Linear lightwave networks (LLNs), 142

Local Area VAX Clusters (LAVCs), 294

London, 95-99

Los Nettos network, 312

Lucas, Peter, 294, 296

Lynch, Cliff, 349

Lynch, Dan, 4, 123, 160

Macau, 65-66

McKenzie, 9-12

Madison, Wisconsin, 352-60

Malamud.COM, 270

Malaysia, 209-18

Management Information Base (MIB),
34

Marina del Rey, California, 311-17

Index

Information Sciences Institute (ISI),

311

Markwell, John, 54

MBA Experiment, 135-36

MCI, 145

MCI Mail, 109, 270, 280

Medin, Milo, 171-74

Medinet, Singapore Network Services
(SNS), 75

Melbourne, Australia, 184-87, 193

Memnet, 307

Message Handling Systems, 190

MFEnet (MuffyNet), 85

MHSnet, 190-91

Michau, Christian, 273, 275

Michigan state network (Merit), 145

MicroGrafix Designer, 23

Microsoft Word for Windows, 23-24

Mikrokonsultit Oy, 100-102

Mills, David, 88

MILNET, 252, 335-36

MIMOS (Malaysian Institute of Mi-
croelectronic Systems), 209-15

Minitel, 19-20

Mobitex system, 62

Moccia, Mary Jo, 221-22

Mockapetris, Paul, 147, 305

Moffet Field, California, 171-74

Mohamed, B. Awang-Lah, 210, 212

Mosaic, 313-14

MOSIS service, 311

Motif, 56

Mountain View, California, 161-65, 171

Multichannel, definition of, 189

MultiMedia Data Switch (MMDS), 199

Mun Chun, Chew, 71

Murai, Jun, 39-48, 324-25, 328, 363

Murakami, Ken-Ichiro, 323

N-1net, 38

NACSIS, See National Center for Sci-
ence Information Systems (NAC-
SIS)

Naemura, Kenji, 321, 323

Narong, Khun, 226-28
NASA-Ames Research Center, 30, 171,
195
NASA Science Internet (NSI), 171-74,
244, 252
applications, 171-72
as Internet backbone, 173
links, ownership of, 173
Network Control Center, 172
SNMP and, 172
NASA Telephone Company, 173-74
Nationaal Instituut voor Kernfysica en
Hoge-Energiefysica (NIKHEF),
2334230, 2087292
National Center for Atmospheric Re-
search (NCAR), 87
National Center for Science Informa-
tion Systems (NACSIS), 38
National Center for Software Training
(NCST), 340-45, 348, 351
National Computer Board (NCB), 78-80
National Science and Technology
Board (NSTB), 67-68
National University of Singapore
(NUS), 67-69
Naylor, Chris, 175, 177
Naylor, Jeremy, 175
Naylor, Richard, 175-77, 289
Neggers, Kees, 90-94, 245-47, 251-53
NESTE system, 100
Netherlands:
Amsterdam, 90-94, 233-44
Utrecht, 245-53
NetPhone, 199-200
Network Information Center (NIC),
243-44
Network Interface Unit (NIU), 142
Networld, 29
Netzzentrum fuer die Wissenschaft, 255
Neuman, Clifford, 312
New Mexico, Santa Fe, 147-52
New Territories, 60
New York City, 141-44
Acorn project, 141-43
CITI (Columbia Institute for Tele-In-
formation), 143-44
New Zealand, 178-81

Index

World Communications Laboratory,

181

Ng, Nam, 59

Nice, France, 119-23

Nielsen, Torben, 36-37, 45

NiftyServe, 42

NIKHEF, See Nationaal Instituut voor
Kernfysica en Hoge-Energie-
fysica (NIKHEF)

Nippon Telegraph and Telephone
(NTT):

Kasumigaseki showroom, 323
vision, 321-23
Yokosuka research laboratories, 321
Ni Shtilleabhain, Aine M., 144
NJE, 15
Noam, Eli, 144
Nojima, Hisao, 323
Nomadic Research Labs, 164
NORDUnet, 93, 104-5, 234, 236, 240,
250, 282
Norwegian Telecom, 71
Novell, 12, 59, 105, 280
Nowlan, Michael, 242
NREN, 103
NSF, 38, 85-89, 122, 197, 281-83, 303,
354, 364
connections program, 146
NSFNET, 11, 30, 82, 83, 87-91, 137, 145-
46, 174, 195-197, 236, 243, 251-52,
256-57, 273, 281-83, 302, 312, 335,
364
NSTB, See National Science and Tech-
nology Board (NSTB)
NTT, 45
NUS, See National University of Singa-
pore (NUS)
Nynex, Minitel and, 20

OCR software, ISO standards and, 332

ODA (Office Document Architecture),
37, 98

Office of Technology Assessment, 146

Okado, Tomoo, 41-42, 326

Open Desktop (SCO), 56, 348
OrderLink, Singapore Network Serv-
ices (SNS), 75
ORSTOM, 111-14, 119, 273
computer support, 112-13
functions of, 111-12
research centers, location of, 112
RIO network, 113-14
OSI IS-IS, 137
OSI (Open Systems Interconnection), 3,
12, 17, 45, 86-98, 191-93, 197, 235,
237, 246-48, 258, 276, 331, 361-62
OSPF, 137

PACCOM, 180, 330

PACE, 261

Packet Assembler/Disassembler, 356

Palalikit, Surat, 220-22

PARADISE ESPRIT project, 96

Paris, France, 19-20, 108-15, 269-76
CNRS (Centre National de la Re-

cherche Scientifique), 273

INRIA, 119-23, 255, 269-71, 273
Renater, 273-76

Parkkari, Vesa, 100-107

Partridge, Craig, 33

PBX, 33

PDP-11, 88

PDP, 53, 110, 292, 295-97

PeaceNet, 284

PERL code, 25

PhoneNet, 356, 358-59

Phone Shell Project, 46

Poireau, Gerard, 19-20

Port of Singapore Authority, 204-7

Postel, Jon, 3, 149-50, 311-12, 316

Prague, Czechoslovakia, 13-16

Prem, General, 229

Prestandardization research, 248

Privacy, 365

Programmers, memory and, 305

Protocols, basis of, 304-5

PSInet, 105

PTTs, 89, 91, 247, 256, 265-66, 318, 319

Index

Q
Quarterman, John, 170
Quipu, 95-96
interfaces interacting with, 96

RACE (Research and Development in
Advanced Communications
Technolgies for Europe), 261, 264-
66

Rahman Bin Shafi, Abdul, 215

Ramani, S., 340-45, 348-51

RARE, See Réseaux Associés pour la
Recherche Européenne (RARE)

Rate arbitrage, 265-66

Recommendation D.1, CCITT, 68

Relevantum Oy, 100

Renang Group, 213-14

Renater (Réseau National de Télécom-
munications de la Recherche),
273-76

Renaud, Pascal, 111-13

Request for Comments (RFCs), 3, 311

RFC 1181, 234

Réseaux Associés pour la Recherche
Européenne (RARE), 91, 233,
234, 237, 242, 247-48, 250-51, 272

contributions, sources of, 247

Council of Administration, 238, 247,
252

Operational Unit, proposed setup
of, 251-52

Réseaux IP Européens (RIPE), 233-39,
242-44, 250

Network Control Center (NCC),
237-38

Reynolds, Joyce, 316-17

Richter, Walter, 125-26

RIPE, See Réseaux IP Européens (RIPE)

Roberts, Steve, 161-65

Robotics projects, INRIA, 119-20

Rolnick, Harry, 65-66, 337-38

Romkey, John, 33-34

Rose, Marshall T., 34, 95, 111, 148, 196,
317, 363

Rosebud interface, 168

Rosenbaum, Jurgen, 268

ROSE project (Research Open Systems
for Europe) , 264

Routing domains, reachability be-
tween, 137

RTF, 24, 25

Rutkowski, Tony, 3-8, 13, 22, 35, 44,
116-17, 124-25, 128-33, 154, 277-78

S
St. Silicon bulletin board, 286
San Diego Supercomputer Center, 312
SDSCnet, 86-87
San Francisco, 166-70, 318-20
SF Net, 166
Wide Area Information Servers
(WAIS), 167-69
San Jose, 22, 29-35
Santa Fe, 147-52
Schaumberger, Dean, 298-99
Schiller, Jeffrey, 33
Schooler, Eve, 315-16
Schrader, Bill, 87
Schwartz, Michael, 170, 302
Schwartz, Mischa, 141
SciTech, 299-300
SDSCnet (San Diego Supercomputer
Center), 86-87
SEEDNET, 335-36
Seoul, Korea, 329-34
Korea Telecom, 331-34
KREN (Korea Research Network),
330
KREOnet (Korea Research Environ-
ment Open Network), 329-30
SGML, 37, 332
Sham Shui Po, 55-58
Shaw, Robert, 153
Shein, Barry, 195
ShowNet, 31-32, 34, 166
Simple Book, The (Rose), 34
Singapore, 67-82, 202-8, 219
IDNet, 79

376

Information Communication Insti-
tute of Singapore (ICIS), 69-70
ISDN, cost of, 71
IT2000, 77-78, 80-81
National Computer Board (NCB),
78-80, 204
Port of Singapore Authority (PSA)
systems, 204-7
Restricted Zone, 202, 204
SCATS system, 203-4
TECHnet, 67-69
Tradenet, 74-75, 81, 204
Singapore Network Services (SNS), 74-
76
$Link, 75
Medinet, 75
OrderLink, 75
Tradenet EDI system, 74-75, 81
Singapore Telecom, 68, 70-72
Skyphone, 71-72
Teleview, 72-74
Singapore—The Next Lap (Times Edi-
tions), 78
Skilling, Peter, 225
Skyphone, 71-72
SMDS, 32, 67, 106
Smith, Mike, 117, 130
SMTP, 191
SNA, 92, 102
SNS, See Singapore Network Services
(SNS)
SnailMAIL, 366
SNMP, 32, 33, 148, 198, 316
NASA Science Internet (NSI) and,
172
Software Tool and Die, 195
Solutions Showcases, 32
Song, Joo-Young, 333
Song, Tay Eng, 80
Sons of Bruno, 35, 117
Sprint, 33
STACKS (Malamud), 45, 65-66, 67, 130,
323
Standards, free distribution of, 131
Standards haven concept, 330-31
Stanwyck, Don, 69-70
Sterba, Milan, 271-72, 302

index

Stern, Thomas, 142
Stock Exchange of Thailand (SET), 220,
222-23
Stockman, Bernhard, 93
Sun Microsystems, 23-25, 151, 161
SUN protocol suite, 188-90
compared to TCP/IP, 189
E-mail, 190
messages/ packets, architecture of,
190
streaming protocol, 189
SURFnet, 90-92, 245-47, 257
restrictions on content of traffic, 91
Swiss UNIX group (CHUUG), 243
SWITCH (Swiss Telecommunication
System for Higher Education
and Research), 17, 121, 242-43
Switzerland:
Geneva, 3-12, 18, 116-18, 124-33
Zurich, 17
Sydney, Australia, 188-92, 193
SUN protocol suite, 188-90

T1 lines, 30, 256-57, 312

Taipei, 335-36

Taiwan, 38, 335-36

Tampere, 100-107

Tan, Rosli Md, 212

Tarjanne, Pekka, 7, 22, 35, 116-18, 132,
153-54, 302

Tarmidi, Ahmad, 214-15

TCP/IP, 15, 17, 34, 40, 45-46, 47, 68, 85-
86, 90, 91, 95-96, 102, 106, 112,
122, 172, 186, 189, 200, 211, 233,
234, 237, 240, 248, 256-58, 268,
271, 274, 277, 294, 315, 329, 354,
363

protocol suite, 9, 11, 12, 47, 85, 88,
114, 196, 311
routing protocols, 137

TECHnet, 67-69

Telebet system, Jockey Club (Hong
Kong), 53-54, 62

Telecom Finland, 104, 106, 117, 125

377

Index

Telekom Malaysia, 211, 213, 214
Telenet, 356
Teleview, Singapore Telecom, 72-74
Telnet, 139, 189, 316, 356
Tenaga Nasional Berhad, 215-17
communications system, 216-17
distributed systems project, 215-16
TeX, 37
Thailand, 219-32
Thinking Machines, WAIS and, 169
Thio Hoe Tong, Dr., 67-69
TIFF files, 23, 24
Time Engineering, 214
Todai International Science Network
(TISN), 39
Tokyo, 38-42, 321-28
Tokyo City Hall, 326-28
Top quarks, 297-98
TPS, 24725
Trademark protection, Thailand, 228-29
Tradenet EDI system, Singapore Net-
work Services (SNS), 74-75, 81,
204
Trade press, technical knowledge of,
109-11
Transpac, 20
Tsim Sha Tsui, 58-59
Tymnet, 356

U
U.K. Coloured Book standards, 193, 358
Ultra networks, 18
United Nations Development Program
(UNDP), 341
United States:
Berkeley, California, 157-60
Boulder, Colorado, 21-25, 153-54,
157, 301-7, 361-66
Chicago, Illinois, 291-300
Cleveland, Ohio, 286-90
Honolulu, Hawaii, 36-37
Ithaca, New York, 134-40
Madison, Wisconsin, 352-60
Marina del Rey, California, 311-17
Mountain View, California, 161-65

New York City, 141-44
San Francisco, California, 166-70
San Jose, California, 22
Washington, D.C., 145-46, 277-85
U.S. Defense Advanced Research Pro-
jects Agency, See DARPA
University College London (UCL), 95-
99

networking at, 98-99
University of Hawaii network, 36
University of Melbourne network, 184-
87
University of Sydney, 190
UNIX, 24, 47, 56, 241, 244, 339
European UNIX Users Group
(EUUG), 240-42
Routing Information Protocol (RIP),
137
Swiss UNIX group (CHUUG), 243
USENET, 46, 239-40, 348, 364
UTP-based Ethernets, 40
Utrecht, 245-53
UUCP, 15, 39, 103, 122, 177, 189-90,
211, 239, 355
Dial-up UUCP, 271
ORSTOM and, 113
UUnet, 45, 193, 211, 288, 341

Value Added Group, Fujitsu, 41

Value-Added Network (VAN), 266

VAX, 24, 53, 54, 61-62, 292-93, 303, 347,
348

VAXen, 54, 73, 295-96, 298, 346

VME crates, 298-99

VTP, 17

Wah, Chew Keng, 204, 207

WAIS, See Wide Area Information Serv-
ers (WAIS)

Waikiki (Hawaii), 37

Wanchai, 60

378

WAN links, 18
WANs (Wide Area Networks), 306
Washington, D.C., 145-46, 277-85
Weird Stuff Warehouse, 160
Wellington, Auckland, 175-77
City Net, 176-77
WHOIS database, 97, 237, 243
Wholebrain Technology, 153
Wide Area Information Servers
(WAIS), 167-69, 186, 281, 348
client interfaces, 168-69
clients, 168
document indexing, 167
document storage, 168
NeXT and, 169
pieces of, 167
servers on Internet, 169
simple WAIS (SWAIS), 168
Thinking Machines and, 169
user requirements, 168
WAIStation, 168-69
XWAIS, 169
WIDE (Widely Integrated Distributed
Environment), 39, 44-46
ISDN module, 47
Williamson, Scott, 243
Wilson, Robert, 291
Wolff, Steve, 88-89, 145-46
Word Perfect, 24-25
World Communications Laboratory
(New Zealand), 181
Worona, Steve, 13
Wurzel, 21-22

X
X.3, 356

X.25, 17, 41, 54, 62, 73, 83-84, 90, 101-2,

Ps ol207 219023;

211, 246, 274-76, 318, 343, 356-57

X.28, 84, 211, 264, 356
X.29, 84, 211, 264, 356
X.75, 113

X.400, 7, 17, 32, 95, 102, 121-22, 129, 153

X.500, 186, 317
alternative hierarchies, 97-98

Index

common schema for directory, de-
velopment of, 97

directory structure, 95-96
Directory User Agent (DUA), 96-98
integration into ISODE, 95
name structuring, 97-98
performance, 98
relaying mechanism, 97
replication, 97
rigid hierarchy, 97-98

X.800, 23

Xerox Corporation, 304, 315

X Window System, 96, 198, 296
kanji support for, 46

Yap, Michael, 76-77, 80, 82
Yeung, Edwin, 51

Yii Der, Lew, 202-3

Yusoff, Rafee, 209, 212, 214-15

Fé

Zakharov, Dr., 9-12
Zentec character set, 24
Zurich, 17

ens, oe
Tekry jew, ‘ra A ame eee ey
TRS PO aA aol Dolesid: |
Tenay : RIOR oy hiteipeint
Marte CD aetna Reon

sMintrebaited ay chor li? sapere weg
re “ 3) * Jueinetem qaiveler

Mhall wi, TNE 2 MORE
fare. é “ARR emai high ~

ite Ele <ag es ‘2, Li Fuso

FY! fies, 7et ARG He eeio JF Rotax
Mere 2 Sas, mm hye Rete ieee } wa bertW x
paloth pebtiegguedineg

jaa 14

in% 2 B y
r% de BOR STAN Fears gsi

he 2 rider ay
wh. oh Siw 5 44 iy
alien * BP erat of 2 peck ——

&? ‘
2% {@& . ‘3 Or bars +

; £).938 Ved onine welts
‘ i *
: Cas std
Tre Yok ty r
it s : ‘ aah —
7 rl a MAMAN Ss, 343, S04). -

Estes petwrarks, LS
Lipeltet Mg Gorm Dew horpnerit :
{UN DP), 342
rtiond Stains’. ’
fete eg, CAV ford TAP Gi
eo: Colegace, 25-25 15554, :
7, BS) th
«Tht ah Padois, # aoy She: | aa
Chaveiand, (thio Zbl. .
Honaltthy, Seaangeesp e te
{thaca, Mew York, age me
| nin dled ey © iaorale 3

a 19; TALTH

Ef QU -

| SBS,

Way oe - e.

pepe vere
ARTS *

Ueiverdnrer 3

niverniiv of Gf
e> .

Linivers ty OB T antec dn

UNIX, 4, SOC

muro LINX gees

= LCi, 240-82 cor Joe

a béepowell

é 68! yah ZAM
aT SORE a

yah Oe |

ob-bet 08.8

yoo modal aandesintiy

¥ ist (teks we a
Valu Aided € sit SD, aie’ Fa i /

Vege Adda} Meteork Te 4 ri ater

aa, So Bd, alae sae MS, S47,

Jae .

peas ae ate.

ote ae hit? Th i
i 4 at }
Pils hy Fat 7 Ae Pre
ee “en ‘ : | a |
Le | 4 ’
Tae A, he fi 4
ET ee ee aa tit OV; ni ‘*
ies Gerke emt Byki :
: eel ine ’
; ray iy ; Aa
Vala) bye 4h )
ty a i
Peal a i j
7 4 at
re | EN '
y Ar} Tye aa
« * rug
rs i » ” A vs
H
Cat a i 4
r if @ é Py
Pe : i, . 4) f
4 ee SP
A
1
i re
ip NN
e ¥ ,
ik 4
i) j a (
= f ‘ f i et 3
| | ok ,
pot 17 '
it ; a \f
2
xy ed
4 ‘ .
|
ra % '
AS
uy
j
‘
! '
< j i mh)
¢ net aN : 7 '
i
ri
pest | '
i
‘
—
:
4 yt j
:
4
‘
|
{ N
Pd
h ; )
- s : 5 ;
a7 i
i “7
Peet van
a4 i ‘ angel ae
; j , nt i has Hl
. an 7

ye AY we
/ q

(Continued from front flap)

MK
INTEROP

What is INTEROP?

INTEROP—The Interoperability Conference
and Exhibition — was developed to help you
understand the rapid and far-reaching
changes in the computer and communica-
tions industries. INTEROP Fall (San
Francisco Bay Area) and INTEROP Spring
(Washington, D.C.) are the premier network-
ing computing educational forums, and
Interop Company is pleased to include a
copy of this Prentice Hall professional refer-
ence book in the course materials of all
INTEROP conference attendees.

INTEROP provides a leading edge technical
program consisting of several dozen in-
depth tutorials, 50 conference sessions, and
informal Birds of a Feather sessions. The
unique INTEROP exhibition provides a “lab-
oratory of learning” where vendors demon-
strate the interoperability of their products
and technologies through a connection to
the show network. INTEROP’s exclusive
Solutions Showcase™ Demonstrations fea-
ture competing vendors cooperatively
demonstrating the interoperability of their
products and illustrating the benefits of new
and emerging technologies.

To request more information about
INTEROP:

Interop Company
480 San Antonio Road, Suite 100
Mountain View, CA, USA 94040-1219

1-415-941-3399, Fax: 1-415-949-1779
1-800-INTEROP, Info@Interop.COM

PRENTICE HALL
Englewood Cliffs, NJ 07632
Printed in the U.S.A.

COMPUTER NETWORKING

—= A Technical Travelogue —= ee
CARL MALAM U D

After traveling around the world 3 times in 6 months and visiting 56 cities! 21 countries,
Carl Malamud chronicles the results of his voyage of discovery i in Exploring the Internet: A

_ Technical Travelogue. \n an easy-to-understand narrative style, he provides you with techni-
cally sophisticated descriptions of everything from new networking research to efforts to
build national and international infrastructures to the many fine lunches and dinners he |
enjoyed. This book is a technical travelogue that looks at people, laboratories, and institu-
tions that illustrate the incredible scope and diversity of the Intermet. :

Here are just a few of the places he visited:

e In Hong Kong, Malamud explored the Royal Hong Kong Jockey Club nines a massive :
network handles millions of bets each race day.

¢ In Amsterdam, he witnessed the birth of EBONE, a long- cawated pan- Eucbern pee
backbone, and visited EUnet, an organization on nS front lines of the pes into places ae
Tunisia, Algeria, and Bulgaria. 2

e In Cleveland, Malamud visited with Tom Grundner, a network evangelist wna. puts com-
munity networks together and spreads the faith in places like Peoria. .

STACKS, also by Carl Malamud, presents a technically sophisticated overview of interoper- Se
ability in today’s computer networks. Bringing key technical trends into perspective,
STACKS combines descriptions with case studies on subjects ranging from HIPPI to ATM,
SONET to Frame Relay, and SMDS to Broadband ISDN. The book shows how technology i is
used and includes real-world studies of gigabit testbeds, terabit networks, any agi :

libraries. |
“Malamud is one Crazy Gaijin! Exploring : "Exploring the Internet is a real surprise! —
the Internet will be a big hit in Japan.” Malamud has figured out how to do
Professor Jun Murai, Director research and have fun too!”
WIDE Network, Japan Glenn Kowack, CEO.

| - EUnet, The Netherlands —
“STACKS succeeds marvelously ... a great book."

2 nu NTL?

oe es eet i ail Travelogue ||

"ANI I

PRENTICE HALL 89

Englewood Cliffs, NJ 07632

Internet Archive Audio

Featured

Top

Images

Featured

Top

Software

Featured

Top

Texts

Featured

Top

Video

Featured

Top

Mobile Apps

Browser Extensions

Archive-It Subscription

Save Page Now

Full text of "Exploring the Internet, A Technical Travelogue"

See other formats