another way, you've made him into a political dissident, someone who might have been otherwise aye political. if you tell boy scout troops they're not allowed to be boy scouts anymore and they're to be young piners -- young pioneers and one group decided they don't like that underground scouts were very important. you would just create another group of political opponents from a political teenagers. >> read more of this conversation and other featured interviews. this is now available for a father's day gift at your favorite bookstore. >> a look at the fcc's open internet policy which they voted to move forward with later this month. if only implemented the proposal that with provide better service providers to move through their

whilek uninhibited providing a fast lane of service for a fee. this is 1.5 hours. >> thank you very much. i would like to ask the and wie willcome up get you mic'd up. i am not going to give long and deductions. you have the full bio with you. i will briefly introduce the panelists was they have a chance to sit down. this thing out here for a second. just do this for second. some of you may be very familiar with this.

we have asked our good friend from the carnegie mellon in university to give us a 10 or 15 minutes sort of technical background on all of these terms and how they fit in and what the architecture is. it is a daunting task to do. turn it over to john. thank you. it over to john. thank you. >> thank you. i will tell you everything you need to know. the internet feels like one big network, when we use it. but, if that were the case there would be no such thing as interconnection. the actually the internet is a network of networks. there are in fact over 66,000

independent, autonomous networks somehow work as one and each network in there is connected to one or more neighboring networks and that means, information that i send may travel from network to network to network before it finally gets to its intended recipient. so, for example, i have a student right now in uganda and i sent her a message this morning. amazingly from whatever network i am at, it somehow figures out how to get my message to her in uganda. there are real challenges here. one of them, how does the network i'm connected to, know which of its neighbors can move that file toward her? that's technical problem. that is a routing problem. also, not only does there have to be a path all the way to uganda, but every network along the ways that to be willing to carry the information. brings me another question. why should it? there has to be a cost to this, has to be some intent tiff. the solution to both of these challenges is buried in the

magic of interconnection agreements. an interconnection agreement where two networks come together agree on technical and business issues of exchanging internet traffic, including, will i carry any of your traffic, if so, which traffic will i carry. talking about internet traffic, which described, telephone world, even though technology converge something quite different. this is internet. don't, and these agreements are unregulated and typically highly confidential. we don't really know what the agreements look like between most networks. we do know they fit two basic categories. they are peering and transit. let me talk a little bit about each of those possibilities. pair something where networks recipro alley, provide connectivity to each other's customers. in this figure and i should thank bill norton, i'm using his great figure here, in this figure, we have a green network and a blue network that are

peering with each other. that means announcement flows from the green network to the blue network says, hey, if you want to talk to any of the customers of the green network, they're over here. i know how to reach them. send me your packets. and a similar announcement goes the other way, from the blue network to the green network. that solves our technical problem. now all the customers of the green network can talk to all the customers of the blue network. similarly in this figure, the blue network is peering with the red network. so all of their customers can talk to each other. note peer something not transsieve. green and blue are peering, blue and red are peering, but customers of the green network have no way to communicate with customers of the red network, not through peering. we need something else for that. so that technically what is going on. in terms of the business arrangements they vary. traditionally, historically, peering was settlement free. we'll search eve other's customers and no money will

exchange hands. that worked among peers and equals, if a small network approach as large network today, that small network will be asked to pay money. that is paid peering. technically the staple but there is money involved. the other arrangement is transit. through transit one network is able to provide access to the entire rest of the internet. some subset of it they agreed to. so, in this case, we have a customer and a transit provider, the light blue network in my slide is the customer and it is going to the orange transit provider, saying let me communicate with all of the internet. and the transit provider says, here, i will send us announcements, here is every network in the world that i know how to reach. if you want to communicate with any of them, just send me your packets. i know how to reach them. and similarly all the light blue one sends announcement, if you want to talk to any of my customers, here is where you

reach those. once those are exchanged, anybody in the light blue network can communicate to anyone anywhere in the world, assuming the transit provider knows everybody. everybody in this case, includes both transit providers direct customers. since this transit provider in my figure is peering with the yellow transit provider, it includes the yellow transit provider's customers too. technically that is what is going on. business arrangements, well the light blue network here is going to pay, and they are going to pay depending how much traffic they send to the transit network and how much traffic they get back from the transit network. once you put those two types of agreements together you can put, create a little order in the chaos that is the internet. you can create something called a tier one network. a network is tier one if it is able to send traffic to the entire internet without paying anybody any peering or transit fees. is not easy to be a transit network or a tier one network. you have to know about all those

66,000 networks in the world. you have to run a big infrastructure. you have to be a highly reliable. it's a lot of work. once you do this you're in a new business. you can now offer transit services to any isp who wants it for a fee, which is exactly what happens. tier one networks, charge, tier 2 networks to provide transit service. tier 1 networks compete with each other to do this there is a lot of competition. tier 1 networks peer, not generally, peer with each other settlement free. you see the hierarchy where the red tier 1 networks are connected to each other and tear 2 networks get their transit service from tear 1 and some cases smaller networks get their service from a tier 2. that is the internet at least as we understood it about a decade ago but it's been changing. there are a few new things we need to understand. one of them is tear 2s a decade ago generally didn't

connect with each other, that's been changing. more and more tear 2 networks are peering with each other to bypass the tear 1. in my top figure here in the slide, you see, ispa and ispb, both connected to a tier 1 transit provider and all the traffic flowing between them goes through that transit provider and they pay for it. you say what if they were to direct connectly as if in the lower figure? that means both of them would no longer have to pay transit fees for that traffic. they save money. on other hand there is a cost. they have to pay for peering connection which costs real money. basically if there is enough traffic flowing between them this deal can be worthwhile and they can save some money. increasingly we're seeing tier 2 networks peering with each other. another change over the last decade is the rise of content distribution networks or cdns. traditional way of thinking about this is the left side, the left figure in my graph and my

slide where content provider says, all right, here is my content, everybody in the world come get it from the servers but they may have to get it from all the way across the internet. that is slow and inefficient and actually much better if you can have them get that content from someplace close by. so they rely on cdns which are a collection of geographically distributed servers which copies of content can be stored. there are companies like akamai and limelight and others that offer this service. some build their own networks which can be quite big. of course if you're really big, you don't have to go to a cdn at all. really big might in this case look like google which run as huge, global network, carrying an enormous amount of traffic that they just run themselves and they can interconnect directly with their very large networks because they're very large themselves. so when you put all of that together, if you are a content provider and you want to get your content to say, customers

of a given isp, how do you do it? you have three main choices. one of those choices you can build your own big network and you wan peer directly with that isp. second choice, you can contract with a transit provide, a bunch of them out there and the transit provider will carry your traffic from your system over to that isp and reach those customers. or, a third option you can go to commercial cdn and say, you host my content near those customers. any of those are possible. well, actually whether a cdn makes it depends on your application. for some application this is is great idea and for some it doesn't work as well which is another question. let me bring to the most controversial discussion in this world which is the comcast and netflix discussion. start with a caveat. i have no idea what is going on with comcast and netflix. i read lots and lots of accounts, but everything i read i consider basically from

unreliable sources. so i will be a little careful here but it seems that netflix is, for a while, served comcast customers via transit network such as cogent and level 3 and probably a few others and netflix and also cogent were complaining that there was congestion on the transit connections. and after a while we know that netflix and comcast announced that they are in a peering arrangement. we assume it is paid peering. that netflix is paying. so how do we understand this? how do we think about this? let me first talk about how i think we should not think about this but a lot of the press articles i read seem to. they annoy me and matter of therapy i i want to spread my annoyance on to all of you. there is a narrative that says, if content-heavy network like a network with lots of netflix content peers with an provider like comcast, one group says this inevitably transit traffic a burden on the content provider

so the content heavy network should pay the cost of this burden. there is another work that says this kind of peering arrangement inevitably generates traffic that benefits both networks so the peering arrangement should be settlement free. both of these arguments are wrong. you start to see that they're wrong, peering doesn't generate traffic. end systems generate traffic. if i am a customer of comcast and i want to watch a netflix video and netflix is willing to show it to me, the netflix video goes through the comcast network, peering, no peering, doesn't matter. is the end systems that generate. now there are cases where peering matters but we need to peer a little more closely in terms of how much traffic comcast would have to carry. it matsers if end systems react to congestion in the network. they end the peering path and transit path, one of them is much more congested than the other and that actually seems to be what is going on here, right?

what seems to be happening, we think, is that there was congestion in the transit links but the direct peer did not have that congestion. why does that matter? some application it is wouldn't. so happens for netflix streaming if you're watching a netflix video and you're watching high-definition and congestion comes along, that application will automatically revert to low definition. so it will back off because of that, congestion reduces the load. so, seems by both accounts there was congestion. the interest question is why. first of all, stress, part of the point of story is you can't look at peering relationship in isolation. you have to look at the broader context to make sense of this. the interest question is why. there are some people who say, well, comcast likes the idea of congestion on their transit links. it got companies like netflix to enter into this paid peering arrangement and have to pay comcast. there are others who say that is

ridiculous. comcast wouldn't like congestion. note that wouldn't just affect netflix. that would affect a lot of traffic and comcast says this was about cutting out middleman. this was about saving transit fees. if it was about congestion those transit providers were being unreasonable and weren't doing their part to relief it. which of these story is right? i don't have a clue. i have only got unreliable sources here. let me wrap that up. interconnection agreements are a central part of making a network of network like the internet work. they take lots of forms. there is peering, there is transit, there is paid, there is unpaid. all the result today of private, unregulated negotiations and those agreements are changing. there are new rules for cdn and content provider networks. an increasing number of peering relationships and changing financial terms and the question is why? maybe we'll address that in the next panel. >> thank you.

[applause] >> [inaudible]. we have discussion on number of fronts about this before i introduce my panelists, i want to tell you if you're following this discussion on twitter, the hashtag for our event today, is hashtag fc cnet rules. fccentrules. if you have questions and comments as we're going feel free to do so. i have extremely distinguished panel to talk about these issues from variety of standpoints. i will introduce them real quickly and i will tell you the layout of all of this. i will ask everybody a opening question and ask everybody a closing question at the end. i have about three hours of questions clearly we'll not get to. we'll save time at very end, last 20 minutes or so, for your questions. if you have them, prepare them and we'll get to you when we can. let me introduce the panelists from my left down to the end.

first is kevin warbacmh associate professor the legal studies and business ethics at university of pennsylvania of are worth ton school. hal singer, progressive policy institute and principal with economists incorporated. professor jer fahlaber, professor emeritus business ethics an public poll i wharton business school and chief economist of fcc. anna maria coke evacuations. georgetown, policy center and. -- cove evacuations. kovacs. start with a quick question for all of you. john just explained that interconnection with refers to the back end of internet. currently it is regulated it is by thousand thousands of these private and confidential agreements between network operators, most of which are so far, at least we understand are done even on a handshake basis.

not a lot of them are even written down. i will ask each of you very quickly to give me your top line here. what role, if any do you think the fcc should be playing in regulating those agreements? i will start with kevin and go down the line. >> sure. i think he started us off exactly in the right place which interconnection is essential to all communication networks and essential to internet. i think it's a false dichotomy to suggest there is this internet world and which is this purely private nirvana and there is this communications world and there is heavily regulated area where the fcc decide on every agreement. that was a false dichotomy in the past. it is more of a false dichotomy now as we go into the future. those worlds are converging. we're going into a transition where the public switch telephone network is converging into an ip network. the fundamental question in that environment, is it appropriate to say, there is no role for a public policy backstop? i think that would be the wrong approach. i think for all the reasons that we've heard, interconnection is

essential and, interconnection is a key opportunity for anticompetitive behavior. and for results that would constrain the information ecosystem. so i think it is absolutely right to say the fcc should not micromanage the process. for the most part it shouldn't be setting prices in the same way that it has for regulated incumbent telecommunications carriers. but i think it is important to have an fcc public policy backstop. one of the reasons was the reason that john gave us. we don't know. we're all speculating about this because all of these agreements are totally outside of the discussions that we have in public policy circles. but, all along, the fcc has been there in the background. the fcc never said, we're never going to have anything to do with internet interconnection. what they said was, we belief traditionally this is

competitive market we don't need to get involved in. that was a factual determination. the world is changing. >> okay. hal? >> thanks. i look at this from a economic perspective, simply, what are the benefits from imposing mandatory interconnection, what are the costs and i tried to lay those out in a short policy brief that was left with you and very quickly on the benefit side of the ledger, if you force these networks to interconnect you might be able to minimize the amount of disruptions. certainly disruptions of anyone's service impose as serious cost but in my review of the episodes looking back in history of the internet, we don't see a lot of disputes. he mentioned a few but in fact what struck me with how few there were relatively could have occurred. i counted to six major disputes and only half led to service outage for consumers and other half it led to disruptions of three, four, five days. that is on the benefit side of

the ledger. it is something and it is not huge. if the probability of these disruptions occur something close to zero, then the expected benefits of imposing mandatory interconnection is small as well. on the cost side of the ledger, what i'm worried about you start telling networks they must connect, you could up sate the maker decision. could upset goals of section 706 which is to encourage deployment. some people point to sprint and t-mobile's reluctance to deploy their spectrum into rural areas. they point to the fact that mandatory interconnection in data roaming agreements causes them to make the bayh decision. wonder what it would do to incentives for isps and middle mile folks and content providers are getting a little taste what it is like to be last mile access. i lay out the costs and benefits and i'm worried the costs may exceed the benefits. that would cause me to look for

something a bit less interventionist or less invasive than outright mandate or obligation to interconnect. >> jerry. >> i also an economist. i take an even simpler view than hal does here. my question is what works? okay? and to note also, interconnection is not just a communications issue. it occurs in virtually any business which the producer of something, that could be canned peas or they could be movies has to distribute something to customers through distributors like supermarkets or comcast, okay? they all do this, via, voluntary private agreements. all over the economy. and what do we see? we see competitive markets, we almost never have any trouble. so, my question is, do we want to regulate this? for lord's sakes why? this is distributing. distributing peas through such

supermarkets works. nobody is calling for regulation of supermarkets because of this why are we calling for regulation of internet access? because it has changed? look, you have 30 years of success of where the problems are at most diminimus. why are we going to come in here and say, oh, well, the fcc has to look at this. they have to look at these agreements. we have to see whether they are correct. the fcc has a pretty terrible reputation as an adjudicator. why is that in any sense going to improve matters? seems to me this is really clear. we have a civil that works. stay away from it. >> anna maria? >> my background is as a financial analyst following telecom for a very long time and so i have a very long memory of all of the disputes over access charge, settlement rates and international arena, all of that

relative to that, i think what we've seen in the way of disputes around the internet in terms of disputes on transit are truly minimal and so i also come down on the side of saying, it is essential that everyone be able to interconnect but we have now several decades of history in the internet saying that private agreements can in fact, commercial agreements can get us there. and barring a breakdown, we really should not be intervening because the rigidity that regulations would bring to the system would probably create far more chaos than the occasional disputes that you have between parties. >> john. >> so, if i were to create an internet from scratch i could imagine cases where you would

have problems, you would have market failure, particularly if some big providers had enough market power. i could imagine cases where you wouldn't have any problems. i look at the internet today and i can't tell. i can see something is changing, whether that is indication of a problem or not, i can't tell. seems to me if, for us to say, the fcc shoulding involved, we have to first be able to say there is problem. second be able to say they can do something to make it better. in order to even ask those questions, i would like to see more information gathered. i think that is where the fcc can do something, constructive, is to try and shed a little light on all of those private agreements and see if there is anything we have to be concerned about. >> great. so that's a fairly wide range of opinions. let's dive into this a little bit more. so, kevin and hal, i wouldn't say dueling papers but you have papers that come at this from a different perspective so let me

come back to you guys next. kevin, you've written in all could of papers that interposing the fcc, let as use the word interpose, we won't say how or how detailed, some fcc involvement in interconnection both for voice as we make the transition but i think you said for all traffic, aside from voice, would have a lot of benefits. can you kind of briefly list some of those benefits that you've described? >> sure. and, again, i take issue somewhat with the notion that this is a purely unregulated space where where should fcc been interdisposed. fcc has been there all along even though it hasn't mandated carrier rules for telephone companies and nor should it. it is not right to say the issue is do we have private agreements and fcc. we could absolutely have lots of room for private agreements but still have a sense there are

certain practices which are anti-competitive and regulator there in part to get the data. jon is absolutely right. we need to know more. the data doesn't just appear. the data appears if there is a regime to say that data has to be made available in some way. so there's a lot of ways to look at your question. one of them is, what ruth said is absolutely true. formally this set of issues is distinct from the set of issues we're fighting about network neutrality. this is about what happens to the edge of the access provider's network. it is not what happens on networks of comcast, verizon or at&t when the traffic goes all the way to their end user. that is what the net neutrality fight is about. many ways the same issue. the question is, let's say that what happened in a dispute like comcast netflix is that comcast, the access provider was deliberately degrading netflix's traffic in order to extort netflix paying it some anticompetitive fee. i don't think that is actually

what happened. but by all indication i read the same unreliable sources jon does. seems like this was resolved in amicable and reasonable commercial negotiation but let's say that's what's happening. fundamentally that is not so different from the net neutrality story that comcast or a similar company would do exactly the same thing, degrade traffic and differentiate traffic on its network. so the extent we care about the nature of the information environment that we live in. to the extent that we care about openness to innovation, that we care that any provider can come on the internet with new services and reach customers or reach everyone else without going through gatekeepers that try to keep it out, to the extent we care about i think we need to think about having a regulatory backstop on private connection agreements. it is not just about the disputes. it is do we believe somehow magically this will work itself

out given the way the network is changing, given we have some providers who have some market power? and they have market power certainly once a customer chooses them. even if there are multiple choices for isps, when you have large broadband isps they control the traffic to their customers. so i think in that environment, to say there's no role for public policy, to say some practices are beyond the pale, i think that is troublesome. >> so, this came up at the last conversation but you think when you talk about a regulatory backstop you talk about the fcc, you don't think there's a role or alternative of the ftc in this process? >> i think there is a possibility for the role of the fcc. there are different kinds of agencies. the fcc has expertise in communications and has more of an ability to adjudicate things in real time. given the nature of the way things have been traditionally done, i think saying this is purely a consumer protection issue which is what the ftc

knows how to do, i think that would be too limited but i think that is an interesting debate to have. i think the first principle is to say we need to understand that there is a possibility of real harm here, to consumers and to innovation and to economic growth in the marketplace. and then, yeah, we can have a great discussion about the exact methods. in some of my law review papers i argued for specific adjudication regimes. happy to talk about that but i think the first principle we can't just start with the notion that anything goes. >> and hal, to your paper, the one you have today in fact for us you look at more of the cost side of the equation here. can you list off 4some of the costs endeared to the system if we had more role of fcc in the near future. >> upsetting incentives of is purchases to continue to deploy and expand their networks. i also mentioned quickly what happens if now google or content

provider or even a transit provider knows it has the fcc's backing when it goes into these negotiations. it might decide at the margin not to make certain last mile investments in its network. i guess another cost i add, i think i offered eight of them through the paper, i won't take you through eight now, but another one i'm worried about unraveling content agreements between content providers and i is sps. netflix and comcast voluntarily enenterred into agreement. if netflix looks out the window, that cogent get as better deal, that would upset voluntary agreement. why would that upset an economist? there is something efficient by them doing a deal. puts costs where it should be and put downward pressure on price of broadband access f you're telling isp is the only way to raise money on backs of its customers. can't make money off transit and cdn service you're basically

putting upward pressure on access prices for consumers. >> okay. jared, let me come back to you. say a little bit more, you're pretty clear and firm about your position on this. what other elements go into your thinking that a case against fcc involvement interconnection is so strong, based on your economics, based on your experience, based on both? >> two things. let me mention first that for the decade from 1990 to 2005, the fcc said everything it could in every possible language that said, we're never going to regulate the internet. that's a really stupid idea. . .

let me tell you what it shows. power.ilroad to truck of these things happen. commission interests itself in a particular area and put a sign up that says "open for business" which is what we did with the open internet order. what happens? to make money looking at customers and making investments. i make money by going to the regulators. getting them to favor of me and disfavor others. the phrase they use is [indiscernible] seeking."t

for 30 years we never had any complaints about interconnection. since 2010 we have had a number of complaints. open for business. the second thing that happens, and this is, i think, the most dangerous part of it, is even though the commissioners may say, well, we want to limit how much we regulate, okay? that won't happen. they will be under constant pressure to expand the regulatory writ. and we've seen this happen now. level iii, for example, in 2010 said let's try to leverage network neutrality into regulating interconnection. now, genachowski said at the time -- quite wisely -- no, no, no, no, that's not network neutrality. we're not going to touch that, okay? well, even that appears to be changing. so it looks like we're now leveraging the interest of the fcc and network neutrality into

in interconnection. into interest in interconnection. how much more do we expect? it will encompass the whole internet eventually. camel's nose under the tent, okay? there's a backstop, light touch regulation. doesn't happen that way, guys. just look at what regulation has done in the past. it's a disaster. if you love the bell system and title ii regulation, you'll love this with the internet. was that clear enough? [laughter] >> well, we'll let the audience audience -- anne marie, a lot of your work, as you said, is how investors think about the broadband infrastructure. what could you tell us about how investors might respond to a more active role for the fcc this regulating interaction? >> one of the things investors do is look at data, and whether you're looking at data from

cisco or telegeography or any number of other sources, what you're going to see is tremendous growth not only in the amount of traffic that runs over the internet, but as john pointed earlier, a tremendous amount of flexibility in the way the internet has evolved to respond to that. and what we have evolved out of the network of networks that john talked about is really a network of innovators, millions of them around the globe. and what has made all of that possible is the tremendous flexibility of the commercial, very highly informal as someone said earlier, handshake agreements in many cases to change rapidly. and one thing, you know, some regulation has some advantages, but one thing that regulation does not offer is flexibility.

once you encode something in law, you're kind of stuck with it for a very long time, for an awful lot of litigation. and so i think a major concern for investors and why i think investors would be very concerned to see government regulation of the internet is that that flexibility which enables the whole ecosystem -- not just the network layer, but all of the application layers, service and layers above that -- would suddenly become rigidified. anywhere sitting in that ecosystem -- anyone sitting in that ecosystem is going to have to start thinking about how do i make sure that my application sitting up here is not suddenly going to be subject to regulation. what keeps me from being a telecom service when so many of the things out there not being regulated could qualify?

so i think what you would find is a tremendous concern about that lack of flexibility, and you would find investment drying up. >> okay. >> john, let me ask you a technical question. so part of this discussion, and i alluded to this earlier when we were talking to kevin, part of this discussion is about voice traffic. so as we're completing the ip transition and moving off of the tdm network and on to all ip networks, one of the questions is whether or not interconnection requirements on the voice network, how and if those translate to the, to an ip network and whether or not it's kind of just expanding that to all traffic and treating them all under different kinds of interconnection rules. if we were just talking about voice traffic, from a technical stand point is it possible for isps to distinguish voice packets from other packets and subject, you know, just those to interconnection rules? and if so, is there a way of

doing it without introducing other kinds of technologies that maybe have some downsides that we have seen in other contexts? >> so complicated question. let me divide that into two cases. there is a voiceover ip service that would broke look to you like a normal telephone. they're never mixed together with other internet traffic. in the world today, they may actually be converted back to old circuits and then back to ip again and sort of a leftover of the legacy circuit switch days. i can imagine the world where, hopefully, where that sort of art official technology change -- artificial technology change will go away and they'll all be ip packets, but they'll all be voice. and it won't be any different from what it is today. contrast that with something like sciech where the -- skype where the voice packets are all

mixed together with all the internet traffic, and now i can't tell it apart unless i use deep packet inspection. interesting question. the first kind of voiceover service is probably going to live on for the foreseeable future even as we go to ip. we will have internet connection that technically is ip but in terms of business arrangements looks a lot like telephone. but the more we mix these together i'm -- technically, i don't know how we're going to separate them. and from a policy perspective, i'm actually getting a little worried as we start to treat these different types of traffic differently as to why and how we're treating it differently, and if it's because of exerting market power or just normal business. >> yeah, well, go ahead, anna maria and then jerry. >> well, to me, looking at it again from the investment perspective, think about someone who's designing a game, right? i want to invest in someone

who's designing a game. i want that kid this his garage to think about the -- that kid in his garage to think about the best possible game, not to think about how to design it in a way that will keep the voice part of the game out of regulation, you know? look at skype. you figure out how to divide voice from the video packets, what's skype going to look like? how are they going to sort of arbitrage the rules to stay out from under the regulation? that's what i meant when i said earlier that for investors, the huge concern is going to be eliminating a lot of innovation and a lot of flexibility because the energy that now goes into those is going to go into regulatory arbitrage. how do i design my product, my service to avoid regulation? >> jerry? go ahead. >> this was a very interesting point which was the transition from the old public switch telephone network, the copper

wires, to eventually to ip. and i want to commend two folks. one is kevin, he has a great paper on this. t i disagree with all his conclusions, but it's a great paper. [laughter] the other thing i want to mention that chairman wheeler has teed this issue up, and it's an ugly issue. i'm glad he did it. somebody's got to look at it because, you know, telephony, wire telephony is in freefall. and we've got a lot of companies that have all kinds of obligations to maintain copper wire in a situation where that's a dying, dying market. i mean, okay? and we have to figure out how to do that. now, the only thing i don't want to have happen is that the death of the pstm become the tail that wags the interconnection dog. and i think we have to keep our view on what do we want interconnection to look like? do we want it to be title ii regulated, or do we want to continue as a free market thing? and i don't want the death of

pstn to really determine that outcome. i think that would be a mistake. >> i don't want to mischaracterize your position on this, kevin, so let he just clarify this. is your view that however the fcc gets involved or more involved or stays involved with interconnection, do you see a distinction between voice traffic and all other traffic in terms of what that regime would look like, or do you see a similar set of rules, similar set of regulations applying to all of them, and if so, is it 251 and 252 of the existing act? >> there's a set of legacies. the paper that jerry's talking about is called "no dial tone," and the issue there is we have a set of industries that is subject to certain regulation, and the problem is the way things have evolved, that regulation is all or nothing, and once traffic goes to ip as a technical matter, the argument is suddenly all those rules go

away. while the industry stays the same. while you still have the existing structure of the communications industry where we have no longer legally mandated monopolies but still some very powerful, very dominant players, and players that understand how to play that regulatory arbitrage game. you can't just say, guess what, ip is magic pixie dust, now you're in the internet, and now it's a get out of jail free card. that's the flip be side about the concern that you're hearing about regulatory arbitrage is an arbitrage that goes the other way. so as that transition happens -- and i agree with jerry, a difficult set of issues, but one that appropriately the fcc is taking on, and the industry is taking on as well -- we need to work through what the transition looks like. at the end of the day, no, it doesn't make sense to say there's some magic about voice packets that's somehow different than any other kind of packets. at the end of the day, i think we need one interconnection regime which is, again, why i've

been trying to argue these distinctions we've historically made between what happens on the pstn and what happens to voice traffic that somehow is not on the pstn, what happens to net flew central -- net neutrality on the network versus off the network, all these distinctions ultimately don't make sense n. the middle there's a lot we have to do to maintain relationships and to manage a transition. i think that's appropriate. but the ender of the day, the question is what we want to have. and what i'm worried about is what jerry said at the beginning. i would hate to see the internet turned into a supermarket that's just selling us peas. that's not what the internet is today. that's not the kind of open platform that's generated so much extraordinary innovation. that's a traditional market where you've got a distributer that controls what's on the shelf space, and they have the power of life and death over every purveyorover products by -- of products by allocating

that shelf space, and it's a linear market. nothing comes back from consumer back the other way. that's not the internet that we've had over the past few decades, and that's not the internet we should have in the future. >> so, i mean, obviously, the fcc can only work with the tools the fcc has available to it by law. we've mentioned, i mentioned section 251, section 252, that deals with interconnection now on the switch telephone network. interesting enough, if that was sort of within the tool set that the fcc was going to use to get more involved here, that would also kind of ask an interesting question about the role of state regulators in that process. so i want to ask all of you or any who have an opinion about this, is there a role -- leave the legal aside for a moment -- should there be an appropriate role for state regulators in managing or regulating interconnection, and if so, what should that role be as distinct from the federal role? who wants to take that on? jerry probably has a very short

answer. [laughter] >> good god. [laughter] let he mention one thing, however, because we haven't talked about this at all. but generally, if we have problems with terminating access to monopolies as kevin mentioned, if we have some problem in what is otherwise looks like a competitive industry where we might get into some market power, the answer to that is not regulation, it's antitrust. and i think the same answer applies at the state level. getting the state regulators involved if i don't even want the feds involved, i certainly don't want the states' guys, but can they do something in the antitrust arena? can state attorney generals take action if they think it's necessary? because i think that's the appropriate place to have action if we have whatever problems might arise with sewer connection -- interconnection, that's what we ought to be doing about it, not regulating it. >> anna maria? >> i don't believe that something at the level of the

states has a role in what is global traffic effec

discussion that momentum h that kind of -- that doesn't have those kinds of resources and would also find itself fighting in 50 different state arenas and a whole bunch of courts. that just doesn't make a lot of sense. >> i think either side of this event wants the states involved -- >> i want to make sure kevin gets, goes on the record one way or the other. >> that would be a complete nightmare. i wanted to just elaborate on something gerry said. he said we do have antitrust as a backstop if a transit provider thought it was being excluded from a deal. we have something else, too, that's important, and i mentioned this in the policy brief, and that is we're about to get a new set of rules that are designed to protect edge providers, namely content and app providers and device makers, and these are going to be a no blocking rule -- >> you're talking about the open internet. >> yeah. it isn't just entirely antitrust we'd be leaning on, we'd be

leaning on antitrust plus whatever new productions are going to come out for content providers including a nondiscrimination rule, a no blocking rule. so to me, the real way to kind of tee up the policy debate is once you have antitrust and the rules that are aimed to protect content providers, you know, what extra protection are you getting by overlaying an interconnection obligation? for me, it's a pretty small amount of protection that you're buying. i think that the only folks you're really helping out are going to be these stand-alone, transit providers or content delivery networks, and it's just not obvious to me what additional protections are needed, what kind of social purpose is served by breathing life into intermediaries, and i mean that in a loving sense -- [laughter] well, i upset someone from a level iii when i called them intermediaries. i didn't mean anything pejorative about that. but, you know, what are you getting with this at the margin? that's, to me, the key policy question that should be addressed. >> so, kevin, let's be clear, do

you see a role for states in terms of interconnection regulation, and if not, how do you keep them out? >> well, turn the question around, what could states contribute to anything in this regime or in this world? i think there's a good argument that states will continue to have a strong role in universal service because they're closer to the ground, and they are familiar with local variations which still exist even in a global internet interconnected world. and states, perhaps, have a role in consumer protection, again, because they're closer to the end users and heavy a better understanding and a better capacity to resolve some of those issues. i think in terms of an overall interconnection regime for the internet or for ip-based services no reason to say that that has to go through 50 state commissions, and i would

certainly agree that that just, you know, magnifies a regulatory problem. but, you know, there are some of these issues that come into play where, for example, if you have a rural provider that's being blocked or traffic that's not being delivered, i think a state could have some role at least as a fact-finding matter in that dispute so we could talk about how to do that. but i don't think that's really core to any of these issues that we're talking about here. >> yeah, john, go ahead. >> so i've already made a comment on whether we regulate and transparency versus regulation, but assuming you were going to regulate, i'd just remind you that we're looking at interconnection between two big networks. each of those networks may operate in dozens of states, and they probably have interconnection points in at least half a dozen states, maybe a dozen states. if you have a dozen different sets of state regulations, what does that a do to interconnection agreements, and how can you even find where the problem is if you're only looking at what's going on in

one state? i think you need federal agencies involved. >> yeah. >> i mean, i guess to me one reason i was happy to see section 706 is sort of the preferred mode in the open internet order that came out -- even though title ii is also raised -- because it seems to me that that leaves the fcc a lot more flexibility in all of these areas in terms of what it would draw into the net of regulation and what it would exclude. i mean, it's lovely for kevin to say states have this role but not that role, but if you actually -- as i know he has done -- extensively look at 251 and 252, states initiate steps, you know? states come back and say, wait a minute, i think voice is actually under my jurisdiction. it's kind of hard to prevent that if you've brought in that regime. and what that then mean is the a

very lengthy process during which investors have no idea what the outcome is going to be, during which the attention of the companies that are involved is focused on the legal issues as opposed to focused on their business. so that, to me, is the concern. >> did you want to follow up? >> sure. >> one is this nose under the camel's tent, look, i was at the fcc in the 1990s. they didn't want me to have the word "internet" in my title, so i got to be called council for new technology policy. we were dealing with all these issues. we have dealing with regulation of voip in 1998 when the fcc put out the stevens report which if you look closely says, essentially, phone-to-phone, carrier, voiceover ip is a communications service, we're just not going say that right now because we want the market to grow.

if you had open internet rules from 2010 to 2014, and all investment didn't dry up, and all the internet didn't just become this regulated thing that everyone's afraid of. so it doesn't necessarily go all one way or the other. the one thing i'll say about the states, the one thing that i worry about with is pstn transition, and this gets fairly technical, but the efficient way to do ip interconnection is with a smaller number of interconnection points. so the legacy interconnection of what's called tdm, the kind of interconnection we have on the phone system, you basically have interconnection all around the country. so there are lots of companies that are interconnected into the phone network in lots of different places. ip, so you look at companies like comcast, they're the third large phone company in the u.s., they switch all their traffic in something like four or five places across the country. that's the efficient solution, that's where we're going to go to, and i don't think regulation should stop that. but i think, you know, if you have a jump pod from a world

where there are hundreds of providers around the country that have interconnection into the network and you say, guess what, it's all gone, if you want to carry your traffic all the way to cleveland and interconnect with us here in this way, have at it, i think we these a transition, and maybe states have a role there. but i don't think that that inevitably means we're going to have 50 sets of rules and states are going to be monkeying around just because there's a transition period. >> all right. we'll hold you to that. [laughter] let me, i want to ask two more questions, and then i'm going to turn it over to the audience. again, to remind you the hashtag is @fccnetrules. let me ask two last questions before i turn it over. a couple of you have talked about netflix, and i want to sort of explore that a little bit further. so beyond the actual relationship they have with comcast, we do know that they've been developing their own cdn technology now around the world, in fact, for some years, and they've recently announced new transit agreements with a number of isps, not just comcast.

there has been a lot of discussion about this that described that as a sort of first of its kind arrangement and a dangerous one, but i don't know how many of you saw there was an article last week from dan rayburn that actually gave a very nice chart showing of all the main content providers including google, apple, amazon, facebook, microsoft and all of the direct interconnection relationships they had with kind of -- it wasn't complete, but with many of the leading isps and some of the leading intermediaries, is there a better word that they want now? [laughter] >> no. i just know they don't like that one. >> so let me just ask all of you, and take it whatever order you're most inspired by, are these kind of arrangements which are apparently more widespread than suggested, are these kinds of arrangements dangerous, and is this something that the fcc should encourage, discourage or do nothing about? anna maria? do you want to go ahead? >> i would be on the side of encouraging commercial

agreements possibly by simply doing nothing about them which is the way they've evolved for a couple of decades. i mean, another piece of news recently was google saying they're going to be pairing with netflix settlement free. and i think it's lovely. one set of companies reached an arrangement that works for them. netflix itself was another company reached a different -- >> you're talking about google tyber in this case. >> yes, google fiber that works for them. i, you know, that's fine. i don't know -- i don't see any reason to intervene in that. and i think just one piece of data that is at least worth thinking about in this context which comes from netflix's own web site is the fact that they have the speed tests that they're showing around the globe, and there is no place on any isp that they run higher than four megabits per second. that includes google's gigabit

fiber and fiber all over the scandinavia, etc. so i think before we make assumptions about what's going on in terms of congestion with individual isps, etc., we need to start to understand that perhaps netflix is making choices that work for it with a whole range of different isps. >> yeah. >> so i think these sorts of engagements should be encouraged. they're going to happen. and i think as an economist if we see content providers, large content providers basically vertically integrating into cdn networks into transit, they're doing it because they must be getting better terms than if they were trying to go through an intermediary. and so you start to get concerned for the intermediary who gets squeezed out of the relationship whether it's cogent or level iii. but i just want to point out there's still a lot of traffic

from small and medium-sized content providers that's going to continue to use these intermediaries. level iii announced in its earnings call that netflix was not even the top 30 of its customers. so in that sense, you know, you think these guys are going to do just fine without the regulator breathing life into their business models. >> yeah. this is a red herring. the rules are not to protect one set of companies. the rules are about protecting the openness of the internet ecosystem. so here's a data, packet clearing which is an organization that studies things like peering did a study in 2011. they looked at 144,000 peering agreements. they had data from 86% of all of the internet providers this the world. do you want to guess what percentage of them were settlement-free? 99.73%. all right? very, very few had any positive charge of the sort that exists in this netflix/comcast

arrangement. well, you might say, all right, that's because traffic would equal netflix is different because it generates so much traffic. what percentage of netflix are settlement free? over 99%. now, i think absolutely that deal with comcast, the deal they have with other parties may very well be spirally legitimate -- entirely legitimate. again, every indication we have suggests that this was hashed out between two parties. netflix got some benefits, and that's fine. and i think we should allow for that differentiation to exist among companies. one reason, you know, it's funny, for you in the audience, i'm the one farther to the left in this debate, but in the net neutrality debate -- >> not far. >> well, okay, to the left. this terms of i'm the one here that's advocating for some role in the interconnection, and the net neutrality i often find myself on the other side where i'm suggesting the approach that chairman wheeler suggested in the net neutrality or the open internet notice of proposed

rulemaking which would allow for companies to do some level of private agreements as long as there was not a slow lane and someone was degraded, i think that's a reasonable approach if it can be done in a way that has a backstop. but it's the same issue. just because two parties sign a contract doesn't necessarily mean that there isn't anticompetitive behavior, that it's not, essentially, an extortion relationship. we need some he can nhl to get that -- mechanism to get that data to have some sort of backstop. within that, absolutely, go at it. and i think it's entirely reasonable to think there will be some of these deals like the ones that we're seeing that have a price, have a charge, and that's fine. but if we go so far as to say that anything is allowed because we just assume that parties are doing a competitive deal, that's when we have the danger of closing off the openness of the internet and that freedom for innovation that we've all benefited from.

and once that's gone, it's going to be hard to ever get it back. >> john and gerry. there are agreements that bring tremendous peering and interconnection agreements that bring tremendous efficiencies. we know that. to pick an extreme case, there are isps in africa that not all that long ago used to have to send a transit provider on a different continent and then back to africa to get to an isp nearby. that they can peer directly is a huge win. we want to encourage lots of those arrangements. but there is something new going on here which is interesting. a dozen years ago i wrote a paper which i called "the benefits and risks of net neutrality," which i described if you had a broadband provider that had market power, they're always going to be able to exert that in the broadband ma market. but at that time -- market. but there were new technologies that said, well, they could

apply that separately in the video market or the web market or the voice market, and if they can openly discriminate, there might be some interesting new problems we've got to keep our eyes on. and i thought that had nothing to do with interconnection. at the time, it did have nothing to do with interconnection. now it's a little less clear. now that we have interconnection agreements just with a video provider, you start to raise those questions again, you know? is this treated differently because this is a video provider that, by the way, competes with the capable service? i don't know. with the cable service. i don't know. i don't see any evidence of problem. one deal isn't going to make me worry that there's a problem, but it seems to me this is something we ought to be watching and we need a little, again, transparency in peering agreements. >> gerry, go ahead. >> the issue of do these things happen, could they possibly happen, yes, indeed. could we be signing contracts with guns to heads? possible. but there's certainly legal

recourse to things like that. we don't necessarily need a regulator to stop that. you know, contract law is sufficiently strong to cover that. so to me, people sign voluntary contracts, end of story, okay? if not, bring a suit. one of the things i want to point out, this should be very clear here. we're talking about disputes about interconnection and what have you, and we sort of think about, well, ken says we're talking about the -- kevin says we're talking about the open internet or customers, or we're talking about small, innovative providers. we're not. these disputes are between big companies, okay? and many regulatory disputes are between big companies, because they're fighting over the rents. that's what this is about. it's about comcast and verizon and cogent and netflix fighting over the rent. and to pretend it's anything else, you know, you don't understand regulation or business.

>> all right. well, let me -- i'm going to stop there. i'm going to ask one last quickie question, and then i'm going to turn it over. so we've been hostly well behaved here, we were talking about interconnection -- mostly well behaved here. >> red herring. [laughter] >> we were talking about interconnection which is the back end and, obviously, you know, the elephant in the room here is discussions about the front end. and we've mentioned a couple times that the nprm that was issued a couple weeks ago on new open internet rules, one of the discussions that came out of that and that's certainly part of the data collection the fcc is doing is about the applicability of title ii which is the old, essentially, the old pstn set of rules as a basis for the fcc enforcing or, in fact, getting involved in this. i just want to go down the line here, and we'll go the opposite order. this time i'll start with you, john, and come back this way. do you think that this is a wise

road to go down, or do you think otherwise? >> title ii for the interconnection or title ii for the open internet? >> title ii for the open internet. for the cus end -- customer end. >> i think i am not yet sure we can achieve what e need to achieve without title ii, so at in this point i'm going to abstain from that. >> all right. ann that maria? >> i think it would be a disaster which is why i said earlier i was glad to see 706 is the more viable option. i think all layers of the broadband ecosystem get sucked into regulation. whether that's intended or not, it would be done through state commission suits, it would be done through suits of individuals who feel that they can gain something in that process. it's really not avoidable. so i think it is far wiser to start out in a limited way and

make, see if that works rather than start out with something that is potentially 20 years of litigation pretty much stopping innovation. >> gerry? >> it's hard to improve on their statement. it really is. >> okay. [laughter] >> i would say this, i don't -- if we have to have something, i think where chairman wheeler is pointing us to nondiscrimination, and it's hard to argue about that. the only problem i have with that -- again, who'd be against nondiscrimination? but i think it's the nose in the tent argument, and i think that's where you're going which is to say once it's in, it's going to expand, as you mentioned. that's my concern. >> hal? >> so i think using 706 authority is the right way to go. i think chairman wheeler's instinct, initial instinct was the correct one. i think that you have to tolerate some priority deals but police them on a case-by-case

basis after the fact for any, you know, if discrimination rears its ugly head. but bringing title ii in is like installing a fire hydrant in the kitchen for the chance of putting out a kitchen fire. it's aesthetically not pleasing, and your kids could flood the kitchen. >> could be fun on a hot day. >> i'm on record, i think 706 can get the fcc to where the it needs to go, and it's the simplest way forward, and i think it's a reasonable approach. just a couple of days ago there was a decision in the tenth circuit about the universal service intercarrier compensation rules which added further legal sanction, judicial sanction saying that 706 is a viable path forward from the fcc. i think the whole legal debate about network neutrality was really a debate about does the fcc have a leg to stand on in jurisdiction once you're out of the legacy traditional public

switch telephone network world, and the answer is, yes, it does. and so i think that's a reasonable way to go forward. but we're going to have this fight tight, as we should, about the difference between that and the title ii regime, what this means to be a public utility, and i hope that the preceding will hopefully get all that on the table. >> well, that was -- i'm not going to argue with a good answer. let he turn it over and see if there are any questions in the audience. and if you'll wait until a microphone shows up, i've got one on the back, on this side of the room. just briefly introduce yourself and keep your question brief, thank you. >> [inaudible] fcc. we've dope and gave -- done and gave a quick history lesson of all the interconnection disputes that we've had in probably regard time. one that hasn't been mentioned is the retrans issue, namely

where you have two entities and a content provider and a distribution entity hashing out the appropriate interconnection fee if you like. it's not quite the same, but it has some of the same hallmarks. and i wonder if experience or success or than success depending on how you see that where there's a largely unregulated negotiation, if that could be seen as a harbinger of things to come, or if is that sufficiently different that it is not a precedent or an indicator? >> right. so you guys understand. all right. go ahead, kevin. >> thank you. so i was going to -- so that ooh's the answer to gerry's notion that the it's a private contract, we should just let it happen. what's been the result when you have this classic holdout problem with retransmission for video, you have companies doing this game of chicken where they cut off service to customers.

they're starting now to block traffic on their internet services for customers, and the ultimate result is programming costs spiral up and up. you wonder why your cable bill keeps going up? one big reason is that the resolution of these disputes over retransmission where the fcc is really hamstrung by the rules and the ways interpreted, the congressional mandate to get involved here. the easy result is these parties eventually agree to deals which raise prices and give consumers lots of channels they don't actually want. so that's what i'm afraid of. i think that to say that interconnection happens in a private way is great, and i think there definitely should be room for private deals. but if we get to that point for interconnection, i think it would be a real tragic outcome. >> hal and then anna maria. >> i do think there are some lessons from retransmission consent. i don't know if i'd draw the same ones as kevin, but it seems to me you have two behemoths

that are trying to cut up the rents. and in that kind of bilateral negotiation, there's very little room for regulation to do something good. we're just kind of putting our thumb on the sale and pushes the pents to the party who has better lob bests. i would -- lobbyists. if you told the content provider in those situations that it had to license its content or else, you know, there was no way it could walk away from the deal, you know, that, of course, would affect the terms and, number two, it might even create more disputes than you have currently. >> anna maria? >> i think one major difference which is we're still going to have no blocking on the internet at the access level, it's not going to be possible for the access provider to block the content provider. and the -- layers, at levels above that there is a choice of networks for, say, netflix to reach me, right?

so if my access provider can't block and netflix can get to me through a whole web of connections all over the internet, how do you get to that problem? >> john? >> of course, i'm happy to say we do not yet see evidence of those kind of failed negotiations on a massive scale. they happened a few times, and that's not good, but they're not happening as often as in this scenario. if they started to happen more often i, frankly, would first hope that we would see third party organizations -- i happen to be part of btag, for example, or other extra government organizations perhaps playing a mediation role. but we'll have to see if we get there. i hope we don't. >> gerry, did you want -- >> yeah. something that hasn't come up and i think is relevant here particularly when we're talking about the fcc kind of jumping in and regulating here is when we have interconnection regulation of the telephone network which

in the old bell system days worked pretty well because it was pretty easy to do, but there was a period when we had things called clix. does everybody remember that? okay. so we introduced a kind of partially competitive/partially regulated network, and everything went to hell in a hand basket. the whole thing about intercarrier compensation, terminating access fees. i don't know if you remember, clex were positioning themselves in front of dial-up isps. remember no ez? yeah -- remember those? they jacked their terminating rates up very high, and the ilx could not respond, okay? it was a disaster which was eventually fixed in 2004 by the fcc. but, of course, they were the ones that created the problem to begin with because it was partially regulated. okay? partially competitive, partially regulated. what i'm trying to do is to avoid making that happen in the

internet, having it partially regulated and partially competitive. because we know what happens this that case. it's a disaster. >> other questions? i have one in the front here. just wait one minute. >> mike nelson with microsoft and georgetown university. is there anything we can learn by looking overseas? the same discussion is starting to happen in other countries, and i'm just curious if they've got a different way of talking about it or another way of approaching it, when we're seeing a growing number of problems there as well. >> two answers to that question. one is in response to what gerry just said, the disaster of dial-up internet access was when the carriers, the bell operating companies came to the fcc and said the internet traffic is congesting our switches, we need to impose permanent access charges on dial-up internet, the

fcc said, no. the result of which was that the internet market in the did not have permanent charges for most subscribers unlike most of the rest of the world. the result of that was the internet exploded in the u.s. far more quickly and developed far more sophisticated ways than anywhere else in the world. so, absolutely, there was arbitrage, and there was rent seeking by some of those clecs, i but there was also a benefit. and that's one lesson historically. the lesson today on this issue is if you look at oecd papers on this, what they talk about is most of the rest of the world has settlement-free interconnection. and part of that is they have different market structures where they've required open access to broadband providers which almost every other developed country in the world, we haven't done in the u.s. so they have different sores of issues -- sorts of issues. the only place we've seen things

similar are, for example, france. there was an issue where free, a competitive provider of broadband, waséñésqñ allegedly degrading youtube traffic. and it was a bunch of different things going on. the french regulator looked into it and eventually concluded that there didn't appear to be any deliberate keg degradation, buzt was within this regulatory -- but it was within this regulatory regime, they looked into the data, made public the information, and things seemed to be generally resolved. that's the closest case i can find. but most of the rest of the world has a different structure of their broadband market. it's worse in some ways than ours, but it doesn't necessarily have the kind of concentration and market power that we have here. >> of course, a lot of them also, too, have direct government censorship -- >> oh, absolutely. >> -- of content as well, which is different. anyone else want to weigh in on the international question? okay. other questions in -- other questions?

all right. well, i don't see any more audience questions. let me ask you guys since we have a couple of minutes, do you have any closing thoughts? have you changed your mind, kevin -- [laughter] have you changed your mind, hal? gerry, i know you haven't, that seems pretty clear. but any additional thoughts from what you've heard this afternoon? >> well, john referred to bill norton's work, and one of the charts that i have in the pack of my mind -- in the back of my mind on the same web site, i think, is a chart of what's happened to transit pricing. and it is basically, you know, started here, whatever, 15 years ago, and we are down here and still defending. so even where we're not doing settlement-free peering, the pricing of transit per megabit has declined exponentially and done so without any intervention. >> john?

>> i very much agree, but with a couple of caveats worth mentioning. it is worthwhile, we talked about competing approaches of direct peering, transit and cdns, and they are checks on each other in terms of price to some extent. that is great. the caveats being, one, peering is falling incredibly fast. although i note the volume of traffic is going way up, and so far peering steams to be falling faster -- seems to be falling faster. the other being that all assumes quality of service is not an issue, that it's just dollars, and that may be the case as well, but our last dispute raises the possibility that in some cases if congestion and quality is an issue, it's important. >> gerry? >> i'm going to question myself because we hear about. oh, this is all small innovative

firms, garage firms, that sort of thing, and i never hear from them. has anybody heard from somebody in a garage saying, oh, yes, i'm in favor of regulating interconnection? i kind of suspect they weren't, but that's just a guess. i've never heard from that constituency, and i'd love to. >> well, since i lived it, i can tell you -- >> you live in a garage. >> i live in a garage. [laughter] >> well, you're the guy. >> silicon valley. as a silicon valley person, i'll tell you -- and i'll remind you if you have forgotten this -- but we basically don't want to know that you are here. the basic attitude of silicon valley for years has been, there's been these sort of hurricane antitrust actions that every now and then wake us up. but the bottom line is even among the venture capitalists is if we ignore washington, it's not there. which i've argued for years is extremely dangerous and a foolhardy approach to things. but that is still, unfortunately, i think the prevalent view of entrepreneurs as well as their investors at least out in the valley.

>> both encouraging and depressing. >> exactly, yeah. >> the reality is you can't get away from these issues. i mean, you can't get away from strategic behavior on all sides. i think when people say, well, you can't trust the incumbents because they're going to behave strategically, i think that's fair. that's what businesses do, they try and take advantage of opportunities. and when people say, well, google and netflix behave strategically, that's right too. they're not just in there to serve some altruistic goal. investors are in this to make money. they're not in it just to create good innovations just for the progress of science. the question is what's the minimal regime that we need to channel this in a positive direction, you know, do we need anything, or do we need to realize that these issues just aren't going to go away? there's no set of circumstances, i think, where we're going to turn around in 10 or 20 years and say, oh, it all worked itself out, you know?

we can all shut down the fcc and go home. this is the nature started at the beginning of the day, the nature of any network industry that there are these potential for these flashpoints. >> hal, last thought? >> i'm actually worried that we're learning about the interconnection debate through the prism of the netflix, really the netflix wars. whether it's netflix/comcast, netflix/whomever. and netflix may not be all that average representative of your content provider. so i'm concerned that regulation is going to be spawned through a prism that might be a false one. i also want to point out that netflix has certain counterstrategies in its pocket. if it wanted to, you know, for example if an isp gave it a hard time, they could impose an isp surcharge for customers of that isp to discoarnlg people to switch to a rival isp. they're in the driver's seat,

and that's just another reason why we should take, i think, a go-slow approach to adding on another layer of regulation. >> i'm going to give anna maria the last word. >> just to answer gerry's question, there is a blog post on the fcc's blog that was put out friday in response to the chairman's blog by brett glass who runs a wireless isp in laramie, wyoming, and it is an impassioned plea to not regulate the internet explaining from a very small isp's perspective what that would mean to his ability to gain investment. so that is, at least, worth a read. >> all right, thank you for that. and thank you to my exceptional panelists here. and thank you to ppi for organizing this event, and thank you to all of you for m can cooing. i've learned a lot, so i -- thank you all for coming. i've learned a lot, and so i hope you have as well.

and help me in thanking the panelists. [applause] [inaudible conversations] >> a look at a recent article, chuck hagel attending a security conferences in asia. the u.s. and china squared off. they were in singapore, speaking, when secretary hagel accused china of destabilizing me region.