homeaudiencesearchbrowseglossarypast issuesspecial reportshow to subscribe

listsome abstracts

  Contact Us  
special services
  Annual Report
IP insurgency
January 1999
  Seminars to order  
  gateway to Russia  
  Some Russian History  
  Some Russian Friends  
  Some Russian Pictures  
  other news  
  Cluetrain Manifesto Interview  
  Grey Ribbon Campaign
Board Meetings
Open: Support the
Grey Ribbon Campain

  webdesign @ Arcadia, Inc.
COOK Report Web Pages designed by Arcadia, Inc.
FASTNET Internet Services

IP Insurgency: Internet Infrastructure and the Transformation of Telecomm

Can Revolutionary Promise of Stupid Network Be Retarded by ICANN Model of Central Office Control?

Volume Four of an Annual Handbook on the Commercial Internet's Business, Technology and Management Issues
An Anthology of Recent Articles from The COOK Report

The COOK Report Annual Handbooks provide vital reference material in the form of an indexed handbook that can assist those who are trying to evaluate the complex and fast changing phenomenon known as the Internet. For Volume 4, IP Insurgency, all the articles from the March 1998 COOK Report through the February 1999 issue have been reorganized into topics that unfold and overlap over time. A detailed six page index makes it easier to track these complex developments.. The handbook's six sections cover the following six critical issues:

  • Critical Technology Changes of 1998 pp. 6 - 69

  • Internet Telephony: Protocols, Design and Business Model Issues pp. 70 - 129

  • Internet Architecture: Peering Models. Who sets cost of connection to the Internet? pp. 130 -184

  • Quality of Service: Still Elusive pp. 185 - 225

  • Varied Infrastructure Expanding Technologies and Case Studies pp. 226 - 267

  • Internet Governance or Management: a Still Unresolved Crisis pp. 268 - 348

For existing subscribers, the annual handbook has become a means of organizing all the material from past newsletters for reference and review. For many others, it serves as a valuable introduction to the COOK Report. The cost is $495 a PDF (electronic) copy. Double sided GBC bound copies cost $575.

The rest of this message contains:

Rapid Technology Development Pushes IP Everywhere

While in the popular imagination the Internet is the World Wide Web, email, mail lists, and e-commerce, for a much smaller group of experts, the Internet took on new dimensions in 1998. First during 1998 it became evident to everybody that one could not be considered a player in the telecommunications arena without providing access to the Internet. Less widely recognized, but in the end far more significant, was a second development. In the eyes of a small number of cognoscenti, the Internet, considered in the broader sense of the technologies that support TCP/IP, began to play the central role in developments that are now transforming telecommunications on a global level.

In 1998, currents set in motion by the growth of the commercial Internet began to impact all of telecommunications. Many of these currents were identified by the analysis in David Isenberg's Rise of the Stupid Network. At the most basic level, the Internet's on-going expansion helped to fuel a very rapid growth in the overall amount of network data traffic. It created a momentum that helped to ensure the deployment of optical network technologies previously not expected to be commercially viable before the end of 2000.

In this context, some of the people who were watching the build-out of the fiber networks of the next generation telcos were gripped by some gnawing doubts. Might a world-wide, build-out of TCP/IP connectionless networks become an extremely cost effective undertaking? Might it turn into a tsunami that could endanger the economic viability of the multi-trillion dollar PSTN worldwide? This may seem like a bold claim. However, its substance is brought starkly home by the reality of the laying of more than 50,000 miles of new fiber by companies like Qwest, Level 3, IXC, Williams and Enron -- companies whose names are generally unknown to most of the American public.

Advances in Optics Enable Building of New Nets for a Penny or Two on the Dollar

From the end of1996 through the end of 1998 was a period of time when ATT's spin-off of Western Electric, renamed Lucent, came to have a market value exceeding that of its parent company. During this same two years, the maturation of wave division (WDM) multiplexing and then dense wave division multiplexing (DWDM) increased the carrying capacity of fiber from 20 to 400 gigabits per second. The speed of this growth far out stripped the pace of change dictated by Moore's law which, by making mainframe processing power affordable for desktop computers, was turning the legacy telco view of the world on its head. Cheap intelligence sent potential control of the public switched telephone network into the hands of every user at its periphery. The speed of these changes sent tremors through telco business models by rendering obsolete the 20 to 30 year depreciation cycle for the telco's central office switching and network equipment.

The next-gen upstarts would use the commonality of IP packets as envelopes into which the entire range of communications protocols - voice and video, as well as data - could be stuffed. TCP/IP made it possible for these new companies, each of which possessed in its new dark fiber capacity in excess of that of ATT or MCI, to implement networks in which the cost of delivering services would be a small fraction of those incurred by their legacy predecessors.

But while the legacy networks were saddled by the burdens of regulation, they also had the customers. Their existence rested on the installed base of more than 600 million telephones world wide known as the public switched telephone network (PSTN). While next-gen telcos would gain some businesses and early adapter individuals as customers, they knew that their ultimate test would be to take on the huge PSTN with streamlined systems that cost a fraction as much to operate. To do this they would need the development of protocols to enable new services at a fraction of the cost of the old. They would also need protocols to cheaply and transparently merge critical portions of the PSTN to their new IP services. One force impelling the triumph of the TCP/IP insurgency which, seen another way, is merely the ubiquitous application of the protocol, is that legacy companies are generally finding it necessary to offer their own versions of IP telephony. The next-gen telcos are surrounding and enveloping the PSTN like amoebas. They are doing this as the stronger and more forward looking legacy telcos are rolling out system upgrades that, in high user density areas, are essentially copies of the DWDM fiber nets of Qwest and Level 3.

NSP Movers, Shakers and Dinosaurs-- Assessing the Ranks

While the next-gen build-out was underway, consolidation became the battle cry among ISPs. Verio continued its buyout binge spurred along by a $100 million investment from NTT. But Verio loomed not much larger in the scheme of things at the end of the year than at the beginning. Except for the changes in the prices of their common stock, (some prices much more absurdly inflated than others) business continued as usual for DIGEX, PSI, GTEI (BBN), ATT Worldnet, Sprint, AOL, and MSN. Among year's the smartest moves was the Sprint Earth Link alliance which made Earth Link into an excellent national service.

IBM Global Net was purchased for more than $4 billion by ATT. Known in some circles as the "American PTT," ATT thereby gained the biggest legacy TCP/IP network in the world. ATT continued to show its lack of understanding of the new times with its purchase of TCI. ATT's leadership, not grasping the full potential of either satellite or wireless as a local loop solution, seemed fixated on competing with the rapidly merging ILECs for a second wire into American homes. ATT's combination of these two huge legacy dinosaurs, we predict, will go nowhere fast. These behemoths are unable to understand that, in their ability to keep up with the speed of technology changes, their huge size is a liability. ATT's January 99 win of the new backbone for @Home is certainly interesting -- especially when one considers that @home rejected bids by all the next-gen telcos. Certainly ATT is looking to challenge the LEC strangle hold on the local loop.

The one hopeful glimmer to come from ATT all year long was its hiring of Sally Floyd and Vern Paxon for its Silicon Valley based Internet research group. (Cisco however completed the plucking of Lawrence Berkley Labs and increased the size of its very considerable brain trust by hiring the legendary Van Jacobson.) Meanwhile IBM as the other major global dinosaur of American industry, rather than compete internationally against upstart Teleglobe and UUNET, sold Global Net and cast its lot with e- commerce by building up its Washington lobbying organization with Mike Nelson, Roger Cochetti, and John Patrick. Arm in arm with Vint Cerf, these IBM folk rather than push new technology development, got ever more firmly behind their Global Internet Project which collected big industry money on behalf of the newly emerged and morally bankrupt ICANN.

Bernie Ebbers' combined MCI-Worldcom made major waves as an exercise in debt leveraging. It took one major backbone (ANS) and turned it into a corporate firewall security arm of UUNET. Its move to take the largest backbone of all (MCI) and combine that with UUNET was properly attacked as an effort to build monopoly share. John Curran's analysis of the implications of this move was accurate. He offered a very powerful explanation of why it would lead to damaging industry consolidation. The upshot was that MCI was forced to sell its backbone business to Cable and Wireless, an international legacy telco still in search of its first Internet clues. While the successful merger gained MCI executives a pool in excess of $150 million in golden parachutes first promised them by BT, it is likely to spell the end for MCI's very considerable supply of Internet technical talent which is leaving for players who appreciate it. Doug van Howelling's Abeline alliance with Qwest while leaving a smiling Richard Leibhaber has stripped MCI's vBNS research partnership with the NSF of the depth of R&E support that it otherwise would have enjoyed. MCI does have a strong Internet telephony R&D effort, but given Ebber's efforts to strip costs in order to pay off the money borrowed to finance his acquisitions, we wonder whether it has any chance of surviving. Four months after Ebber successfully swallowed MCI, the over riding priorities of the new mega-company appear to be cost cutting not technology advancement.

Peering discussions continued apace and served as a mirror for changing models of the network. John Curran (GTEI CTO) showed how peering, even among the top five or six backbones, depends on market place share. His analysis played a significant role in forcing the divestiture of Internet MCI. Later in the year GTEI (BBN) was in the center of a peering dispute with Exodus. In the resulting squabble those who thought content providers could call the tune found out otherwise. Finally, the model of regional multi-homing to many backbones, pioneered by Fast Net, was used successfully on a national level by Savvis.

The Local Loop Challenge

@Home remains a savvy player. We know many delighted customers. Nevertheless implementation in some areas has been slipshod. From what we can tell problems almost always have been the fault of the Cable provider. However, cable modem internet connections, now approaching a million, are growing in a serious way. The Canadian cable industry seems more intelligent than its US counterpart. Videotron's 1999 experiment with Cisco for the delivery of telephony via cable internet connection shapes up to be quite exciting.

The FCC is revisiting Part 15 no-license radio rules that include spread spectrum wireless for the first time in a generation. It's December 98 Notice of Inquiry Proceeding (98-153) has generated substantial support for its suggested ultra wide band use of the spectrum by a range of devices that can cut across television and other previously prohibited frequencies. One notable comment filing is from Paul Allen's Interval Research Corporation that recommends changes to the rules that will enable a wide variety of devices and services hitherto restricted. IRC conducted a study validating Tim Shepard's MIT dissertation that asserted that millions of radios could operate without interference in a very small area, such as a city. IRC challenged, with impressing technical calculations, fears that such radios and devices operating at very low noise levels would interfere with traditional services. It did a study that showed that a Pentium class PC makes about the same noise as a proposed Part 15 device will - and nobody is complaining that PCs interfere with television reception. The rethinking of FCC policy in this area could help break the LEC monopoly by permitting manufacturers to make high performance, low cost, no-license digital radios that can bypass traditional wire based data services, and turn this technology into a major means of local loop access.

Satellite has the potential to become an alternative to local loop connectivity. Unfortunately it is currently too expensive. Sky Station isn't off the ground yet. Nor is Teledesic. We recently published a cover article on Alohanet and innovative effort to increase the efficiency and there for lower the price of satellite by use of spread spectrum. Alohanet is currently having difficulty at the very critical stage of putting their prototypes into production.

Craig McCaw's NextLink Communications has done a very interesting deal with Level 3 that could have interesting synergies for Teledesic (another McCaw owned company) when its system goes live. Give the amount of fiber involved, McCaw could sell access to that at market rates and throw in satellite local loop for free, if he ever wanted to mount a LEC killer. Readers will find it well worth taking a look at the details of the arrangement found in section 11 of the Level 3 10q filed on November 13, 1998 at http://www.edgar- online.com. Basically Internext, a subsidiary of NextLink will gain the right to use 24 of Level 3's dark fibers by paying a total of $700 million in cost sharing as the network is built.

Technology Development: Fiber, DWDM, Routers, IPV6 & Telephony Protocols

The impact of Bernie Ebbers' speculative excess on MCI is a shame since it is MCI that has developed the very clean, lean and mean, telephony model outlined in our December issue. This model has only five levels beginning with WDM over fiber, framed by gigabit ethernet, on top of which ride IP and MPLS. Applications are handled in two layers, one for web IP applications and one for IP servers which packetize voice video and data. The five layers are tied together by a single hierarchy of web based information systems and network management. It is likely that six systems will make up the management organizational structure of the next generation telco's compared to the 12 systems that must be maintained and coordinated so that the legacy telco can provide separate voice and internet services. If MCI doesn't bring this new architecture to market, it is likely to be realized by Qwest or Level 3 whose backbone is being built largely by ex-MCI senior engineers.

Plans to send IP over glass became reality. CANet3 became the first operational IP over fiber network. By year's end it was generally understood that many DWDM light paths (lambdas) running with gigabit Ethernet or SONET framing over fiber and doing without ATM and without complete SONET gear could cut the transmission cost of data by a factor of as much as a 100. Faster routers were also coming on-line. These included boxes from Lucent, Nortel, and Juniper. Implementations of IPv6 stand in the wings. While quality of service at the IP level remained a seemingly intractable problem, we find it intriguing to begin to hear talk from Sycamore Networks and elsewhere that multiple lambdas at level two may be used to provide a kind of on-demand QoS. (Details on this may be found in this issue.)

Having an 18 month head start over Level 3, Qwest acquired EuNet and Colorado Supernet and started offering IP services and various transport services priced very aggressively in an effort to win market share from ATT and other old line telcos. But Level 3 made a mark for itself with efforts to hurry into development a protocol that would allow providers of IP telephony to seamlessly interconnect their systems with the PSTN. Williams, meanwhile, got read to test Sycamore Networks' first optical filters. Mediatrix, a small French Canadian company, developed a persuasive argument which pointed out that for Internet telephony to take off most quickly, the tools for its implementation should be located at the user's intelligent desktop on the periphery of the network. As far as telephony protocols themselves go, we learned why H.323 is generally to be avoided and why SIP is both elegant and simple and therefore good. In a December 1998 follow-up conversation with Mediatrix we also learned that the industry, instead of giving birth to a dominant IP telephony protocol, is nurturing whole families of protocols. Such families will cover a wide range of Internet telephony applications. Thanks to this widespread protocol development we see few, if any, roadblocks to the development of a considerable variety of IP telephony applications created by a variety of vendors ranging from ILECS to next generation telcos.

No Real Boundaries Separate Commercial Internet from Rest of Telecommunications

In general, two 'rivers' are moving forward. On the one hand, the commercial Internet is growing rapidly. On the other hand it becomes very evident that the innovations designed to propel the commercial Internet need not stop there. If IP Sec is useful in the commercial Internet, it is equally useful in private internets. If IP telephony is useful in the commercial Internet, it is probably even more useful in using corporate Virtual Private internets to take telephone traffic off the PSTN.

In short, by the end of 1998 the technology supporting packet data networks had become so productive that entrepreneurs began to find compelling economic arguments to begin to use it to carry the payloads of older, legacy, connection-oriented voice and video networks. Fueled by the opportunity to entirely eliminate the expensive SONET and ATM infrastructures of the PSTN, companies have been formed to take advantage of these technology blank slates and are beginning to force the older networks to think about emulating them or die. The commercial Internet has unleashed currents that for the first time are really pushing convergence between voice and data networks and between the commercial Internet and the PSTN. These currents are moving cost-effective unregulated TCP/IP nets into direct competition with the regulated PSTN. The grounds on which the unregulated networks compete with the telcos could hardly be less equal. The telecom technology development engine nurtured by the open, non-proprietary intellectual meritocracy of the IETF has created substantial new wealth and entire new industries. It is beginning to dawn on some of the legacy telcos and legacy computer companies that they cannot compete. They will have to cannibalize their installed technology base and do it quickly or die. A lot of powerful people in a lot of big companies will soon have to face the reality that they should be very afraid of the TCP/IP Insurgency bearing down upon them.

The Mysterious ICANN Coup

In this context the Internet now finds itself caught on the horns of a serious dilemma. Three years of domain name wars, culminating in 1998, finally brought home to the American government that the dynamic Internet engine of wealth creation was politically immature. Some of the most critical functions of the network were done by Jon Postel on the basis of nothing more than informally agreed upon custom. What Postel did had no authority in law except as function supported by a DoD contract for a small homogeneous (and largely U.S.) research community. His Internet Assigned Numbers Authority (IANA) functions had no institutional basis. While the Internet was an obscure research network, it did not seem worthwhile to have anyone argue over who set the rules for domain names, IP numbers or protocol port assignments. Suddenly in 1998, with the impact of the TCP/IP insurgency about to change the face of a multi- trillion dollar world wide telecommunications industry, the stakes were very real.

The year 1998 marked the culmination of what some considered to be nothing more than a process designed to provide a formalized mechanism for the execution of the IANA functions. For others, we suggest that the creation of the Internet Corporation for Assigned Names and Numbers (ICANN) was a much more serious matter. These people saw ICANN not just as the means for the administration of critical technical functions but as a vantage point from which they could determine how the Internet should be governed by using it to make the rules under which the Internet would operate.

We have documented extensively our complaints about ICANN in previous issues of the COOK Report and will not repeat the details of our indictment here. Instead, what we wish to call attention to is the question of the limits of ICANN's power. Defined narrowly, ICANN should be nothing more than a body that sets policy for the development and use of domain name space, the assignment of IP numbers, and the assignment of port numbers to new protocols. These are considerable powers, especially when we recall that the first allocations of IPv6 numbers are expected this year.

In reality, the importance of ICANN's role is increased well beyond these functions by the fact that virtually all players see it as the first legally-constituted, international governing body for the Internet. While acknowledging that ICANN's initial role is limited, Esther Dyson has indicated that there many other issues, including e-commerce and privacy, with which she would find it attractive to become involved. She has also left open the door to ICANN's adding a new 'supporting organization' at some as-yet-to-be-determined point in the future. By definition, if ICANN added a new supporting organization, it would be extending its authority.

Indeed, the dilemma created for the IETF by ICANN's tendency to increase the bounds of its authority became starkly apparent as the fall out from the ICANN BOF at the Orlando IETF was debated on the POISED list in late December. The one IANA function that was free of controversy was the assignment of port numbers for new IETF protocols. Nevertheless, the ICANN bylaws elevated what some claimed was largely a clerical function into a Protocols Supporting Organization (PSO) that would give the Internet Architecture Board the right to name three members to the 19 member ICANN Board of Directors.

At the ICANN BOF at the Orlando IETF, Mike Roberts and Esther Dyson explained the ICANN position to a thoughtful audience which has become aware that ICANN is no longer a continuation of the old IANA functions. The wait-and-see attitude of the audience was a marked difference on the part of the unthinking endorsement of ICANN demonstrated at the IETF plenary in Chicago in August. It was pointed out that ICANN would require the IAB to sign a contract for IANA services with it (ICANN) if it wanted ICANN to recognize the IETF as the Protocols Supporting Organization and to permit the IAB to fill three ICANN Board seats. ICANN was told that ISOC was the organization that would have to sign any contracts on behalf of the IETF and the IAB. The leaders of the IETF have always resisted incorporation, fearing quite correctly, that letting the lawyers in the door would be giving them carte blanche to destroy IETF culture.

We asked Fred Baker, IETF Chair, to assess our understanding of what happened. Fred replied "I think you missed some points here. ICANN requires the PSO to be a corporation, and the IETF is not and probably never will be, for a list of good reasons. There was no argument that the IETF should incorporate, during the plenary, the POISSON meeting, or any other time. There was merely the observation that the IETF could not itself be the PSO - and the observation that European Computer Manufacturers Association (ECMA) and others want status in the PSO when it is formed (which is also different from being the PSO themselves). This has been a settled issue for quite a while; what was discussed in both forums what form the PSO might take."

Fred added: "The PSO is an independent entity which the ITU, ECMA, ETSI, and IETF might be members of - and as currently constituted, the IETF would be the only class 1 member of. I think there is only one proposal, and it is to be found at http://www.ietf.org/drafts/draft-ietf-poisson-pso-bl-01.txt. Basically, it describes a shell corporation that has members which are themselves organizations, and describes what their various classes of membership are. The PSO has the right to appoint ICANN Board Members. But the IETF would have a pretty strong voice in the PSO. The IETF cannot sign contracts, as it is not a legal entity. ISOC signs contracts for the IETF when appropriate. But nobody is planning on the IETF signing any contracts here."

Judging by what the lawyers who gave birth to ICANN as a corporation have done to the IANA functions, the IETF has been very well-advised in its refusal to incorporate. As we pointed out in our January issue, the establishment and early operation of ICANN has been done in a way that is totally antithetical to the time honored open and democratic processes of IETF working groups. The ICAN and IETF cultures could hardly be farther apart in their approach to problem solving. Those who might have felt comfortable working with an ICANN with Jon Postel at the head had to ask themselves at the ICANN BOF if they trusted Esther Dyson and Mike Roberts in the same way that they trusted Postel. The answer was surely "no".

We suggest that our readers ponder these issues very carefully. The intellectual meritocracy of the IETF working group is the fundamental cultural foundation from which springs the TCP/IP insurgency that is overturning the telecommunications world. Nothing would suit the agenda of the huge legacy telecom empires better than a world in which their lawyers are able to tell the engineers of the Internet what they can and cannot do. It seems likely to be a few weeks before we will see the IETF/IAB final response.

The Golden Egg - Will ICANN Kill the Goose or Just Steal It?

Right now we know only that ICANN was established in the most mysterious of ways. We know that its leaders give lip service to openness while pursuing an unclear agenda behind the scenes, justifying their bylaws with specious arguments at best, and practicing a style of operation that renders them accountable to no one but themselves. Borrowing the epithet of "adult supervision," we must ask: is ICANN nothing more than a benevolent dictatorship imposed on the Internet by the mighty five for the Internet's own good? Or, is it something else? Time will tell. We find it hopeful that the NTIA's November 25, 1998 MoU with ICANN made it clear that ICANN is not ready to provide adult supervision to anyone. Indeed, ICANN is now receiving parental supervision from NTIA, pending a show of its ability to muster strong enough consensus support from the Internet Community.

Even so, knowing that the IP Insurgency is now so close to total triumph in undermining the old telecom order, we would be naive not to consider the possibility that more serious intentions lay behind the conservative, old line computer companies, and ITU- oriented telco interests saw an opportunity. In the process of institutionalizing the IANA functions they could try to form ICANN as an international regulatory governing body for the Internet - one that they could indeed use to protect their own interests.

The only problem is that they still fail to recognize that the Internet, as the Stupid Network, where the processing intelligence is at the periphery and not inside central office switches, has unlike the telephone system no Central Office that can be used as a point of control. They would like to make ICANN into a Central Office control point for the Internet. What they fail to realize is that, unless they can convince the government of the United States and the governments of the all the other major internet using nations of the world to pass laws giving ICANN legal enforcement powers over its supporting organizations in their respective countries, they are likely to fail. Why? Because there is no Central Office point of location for the Internet. To the Stupid Network, the telco Central Office model is simple irrelevant. Attempts to create one, barring acts of unthinkable international coordination, will fail.

Put simply they have failed to understand that they have no authority beyond the consent of those whom they would pretend to govern. So far Interim president Mike Roberts, has led the secretive and Internet ignorant ICANN board in ways that will likely ensure that when ICANN speaks, no one will listen.

There can be no proof that ICANN will do evil because ICANN is not yet fully constituted, and has not yet been given NTIAs' final blessing. But make no mistake about the legal and operational framework that Mike Roberts, Larry Landweber, Joe Sims and their corporate supporters have built for ICANN. They are trying to structure it with the hope that it will be possible for ICANN to get control of telecom rule making globally including the Internet for whatever purposes they may have in mind.

While the world's legacy telcos and computer companies may be slow to innovate, they are run by people intelligent enough to realize that if they can't win on technical merit, ICANN may be the vehicle for their self-preseveration. In the minds of these people, ICANN looks destined to become the first Internet international regulatory body. Therefore, these companies have ample reason to want to gain direct or indirect control of ICANN. Such control may be their only way to see that the IP insurgency doesn't overwhelm and absorb them.

But something even more profoundly important is at stake. The technologies of the industrial age raised the economic barriers for anyone wanting to start a business beyond the reach of most ordinary Americans. Because our culture and history has long preached the virtues of self-reliance and economic independence, this was hard for most of us to swallow. For the first time in a century, the small cost and enormous power of the personal computer hooked to a modem and to the Internet re-opened the door to individual freedom and economic self-reliance. But this re-opened door is a profound threat to both those business interests that seek monopoly market power and those whose livelihood depends on social and political control of the masses. These people fear the Internet and are looking for a way to control it. ICANN, as constituted, represents the last best hope of achieving their misguided goal.

January 27, 1999


IP Insurgency contains the full text of 63 articles appearing in the COOK Report on Internet between March 1998 and February 1999. They are organized into six sections and presented in chronological order within each section. They are also indexed (about 600 subjects) on pages 349 - 354 of the report. Each section is introduced by an executive summary (written especially for this volume) that explains where and how the material contained in that section fits into the "bigger picture" . The articles in each section are also preceded by their individual executive summaries. The report is structured to guide the reader, through a series of gradual steps starting with the summary and leading to increased levels of detailed description of the way in which the complex systems that make up the Internet fit together and interoperate.

IP Insurgency: Internet Infrastructure and the Transformation of Telecomm

Can Revolutionary Promise of Stupid Network Be Retarded by ICANN Model of Central Office Control?

Volume Four of an Annual Handbook on the Commercial Internet's Business, Technology and Management Issues


Part One: Key Technology Changes in Optics & Next Generation or Greenfield Telcos

    Executive Summary p. 7
    New Infrastructure Costs Are Plunging --Sycamore Networks to Enable SONET Elimination - Terabit Methodology May Enable Inexpensive, Very Reliable, & Open Architecture Router Fabrics p. 11 New Companies' Technology Forces Attention to Cost and Scaling Impacts of DWDM Applications 18 Months From Now p. 17 Optical IP Backbone Revolution Emerges -- Canarie Runs IP over WDM Eliminating SONET & ATM, Expects Cost Savings of Greater than 95% -- Avici Terabit Packet Switch Capable of OC- 1536 -- Plummeting Costs Mean Creation of "Inexpensive" New Carriers Able to Out Perform Old Giants p. 33 Qwest Plants Fiber and Leverages Result -- Exploits New Technology to Leapfrog Older Telcos -- Nayel Shafei Explains Bandwidth Strategy: Will Nurture Demand But Will Not Start Price War - Rumored Bandwidth Glut Meets Reality of Paying for Non Congested Service p. 40 Level 3 Building 16,000 Mile Fiber Network to Link Urban Businesses as Did MFS Upgradable State of the Art Network to Bring Moore's Law to Impact Telecommunications Cost p. 48 Frontier Global Center OC48 WDM Build Out via Qwest Fiber Detailed With an IXC Business Customer Orientation But Without WorldCom Debt, Company Well Positioned for Profitable Next-Gen Telco Role p. 53 Architectural and Engineering Issues for Building an Optical Internet p. 57 BBN's Blumenthal Describes 1995 - 96 Regional Network Integration ProcessExplains Details of Current OC-192 Sonet Buildout p. 58 Carrier Consortia Alleged Cartel Policies Keep Internet Bandwidth Pricing High -- Widely Known But Little Talked about Tactics Ration Bandwidth Availability of Trans Oceanic Cables - Practice Keeps Gilder's Predictions of Plentiful Bandwidth at Bay p. 64 Asian ISP's Cross Pacific Bandwidth of 45 Mbps or More p. 67 In Search of a Meta-View of Internet Change -- A Conversation with Tony Rutkowski on Issues Facing Internet Industry that Cut Across Technology & Jurisdictional Bounds p. 68

Part Two: Internet Telephony -- Protocols, Design and Business Models p. 70

    Executive Summary p. 71

    Internet Telephony for the Stupid Network -- Small Canadian Company Produces Device to Connect Phones to IP Nets & Explains an Architectural Mindset & Standards Framework by Means of Which IP Telephony Can Replace PSTN p. 75

    New Protocols Integrate IP Nets & PSTN Public Net is No Longer Impregnable Bastion -- L3 - TAC IP Device Control Specification Completed Result of Multiple Efforts Will Permit Packet Networks to Envelop PSTN & Empower End User Driven Applications p. 84

    State of the Art of IP Telephony Protocol Design -- MCI's Henry Sinnreich Assess IPDC, Shows Why H.323 Is Badly Broken, and Describes How Hybrid IP and PSTN Networks Will Rapidly Become All IP Nets p. 93

    Cisco's David Oran Discusses Forces Propelling Development of Voice over IP -- Surveys Device Control Protocol Issues from SGCP to IPDC p. 99

    ITXC Corp. Builds Internet Telephony Service Provider Wholesale Business Tom Evslin on Voice over IP Technical Issues & Protocol Developments and Mary Evslin on ITXC Business Model p. 105

    Internet Fax To Challenge PSTN Fax -- Joint ITU - IETF Standard Puts $25 Billion North American Annual PSTN Fax Income At Risk Machines Using New Standards Available Soon - Advocate Suggests IFAX Is Trojan Horse for Adoption of IP Telephony p. 114

    Effort Ramps Up to Map Phone Numbers to DNS -- Viable Protocol Mapping Sought to Enable IP Telephony to Transparently Cross Provider Boundaries p. 124

Part Three: Internet Architecture -- Peering Models Who Sets the Cost of Connection to the Internet?

    Executive Summary p. 131

    WorldCom/MCI Merger: Internet at Risk? -- Merger Would Lessen MCI/WorldCom Incentive to Upgrade Its Interconnections to Remaining Backbones -- John Curran's March 13 CWA Speech - Merger Would Yield Single Dominant Player p. 134

    Vint Cerf Replies to John Curran: We Can't Measure Size of Network Share - Does not Directly Respond to Curran's Private Interconnection Hypothesis p. 142

    Further Thoughts on MCI WorldCom Merger -- Who Pays for Inter Provider Connectivity? What Are Symptoms of Inadequate Connectivity? Other Risks? p. 146

    WorldCom's Proposed Divestiture of MCI Backbone Found Wanting in List Discussion p. 147

    A Content Versus Carrier Peering Battle in High Stakes BBN vs. Exodus Dipute Web Farm Based Business Model Is Weak & May Lack Incentive to Engineer TCP Congestion -- John Curran Discusses Best Exit & Paid Peering p. 149

    Peering Continues to Perplex -- No New Peering Among Privately Interconnected in More than Year - Adjudication of Complex Issues Tempting but Unlikely to Produce Positive Results for the Industry p. 160

    Problems in Exchange of Internet Traffic and Network Performance Issues Debated -- How to Keep Local Traffic Local? Is Metering an Answer? p. 169

    Peering: Backbone X Gives Most Candid Look Yet Private Peering Unobtainable -- Public Exchange Full Peering Rendered Worthless by Bandwith Bottleneck Between Exchange & Big Backbones p. 172

    Changing Economics of Interconnection: Largest Backbones Find Diminished Reason to Continue No Cost Peering at Public Exchanges -- Claims Made That Best Effort Business Model Encourages Network Under Provisioning p. 175

    Who Sets Operational Routing Policy? p. 179

    Discussion of Bogus Route Injection and Other Issues from NanogImportant but Difficulty of Cross Provider Cooperation Emphasized p. 180

    Priori Networks Fails -- Discussion of Lessons Learned p. 183

Part Four: Quality of Service

    Executive Summary p. 187

    Quality of Service Described by Ferguson -- Not Just a "Knob" to Be Installed - QoS Is Complex Mesh of Engineering Decisions - Tag Switching's Role - Network Engineers to Acquire Toolset to Tailor Performance p. 188

    Lucent Uses IP Routing In Silicon to Solve Speed and Quality of Service Demands Building Routers That Can Power The DWDM Internet Design of Lucent's Packet Star IP Switch Explained p. 192

    Diff-Serv Working Group Debates QoS Model Incentives for Most Effective Use Bandwidth in the Face of Network Congestion Debated - Preferential Treatment p. 199

    Two Must Read Papers from the Web on Peering and QoS p. 205

    Mini-Book Review: Ferguson and Huston on QoS p. 205

    Internet Performance Monitoring Services Can Show Details of Private Interconnects Matrix Information and Directory Services Announces Expansion of Internet Weather Report to Serve ISPs p. 206

    Web Proxy Caching by Digex Slammed Downstream Customers Not Informed -- Action Causes Cybercash Users Problems p. 210

    Sean Doran Supports RED as Internet's Most Important Bandwidth Managment Tool p. 215

    What Does Business Want from Internet QoS and from diff serv? p. 216

    Evaluating SONET in High Speed Mission Critical Private Data Network NANOG & Conscientious List-User Yield Unusually Good Technical Summary: Criteria, Strategy, and Choice of Vendors and Equipment p.217

    Technical Decisions from Inet-access -- Making LECs More Efficient - Routing Decisions Based on Correct AS information - Routers vs layer 3 Switches p. 223

Part Five: Survey of Infrastructure Expanding Technologies - Case Studies

    Executive Summary p. 227

    ALOHA Networks Applies Spread Spectrum Increase Satellite Bandwidth Efficiency Will Offer Inexpensive Satellite Local Loop Bypass Internet Connections to Small Business in Americas &;amp Soon Globally p. 229

    IPv6 to Begin Deployment in 1999 Limitations of NAT Technology in IPsec, IP Telephony and Mobile Computing to Force Corporate Moves -- Number Allocation Is Hierarchical With Operational Improvements p. 236

    Issues of Internet Security: Cryptography, Buggy Software and Denial of Service Attack -- Steve Bellovin Discusses Issues of IPSec Deployment, Denial of Service Attacks and DNS Sec p. 242

    Understanding Scalability of DNS and BIND -- Jerry Scharf Examines Viability of Current DNS Tools in Context of Directory Services and Need for DNS Security p. 247

    Internet Research Institute Pursues Innovative Program to Nurture Growth of Internet IN Japan -- Hiroshi Fujiwara Explains NTT's Approach and Japanese Interest in IPv6 p. 253

    Telecom Finland: the Finnish PTT's Strategy -- Be Creative in Business Use of Internet Telephony to Challenge Entrenched Urban Competitors as a CLEC p. 256

    Sprint's Integrated On Demand Network Assessed p. 260

    Texas Internet Service Providers Association Plays Varied Role in Assisting its Members p. 262

    Dave Hughes Hits Universal Service Fund Refusal to Pay Wireless K- 12 Connectivity SLC Action Shows Bankruptcy of Hidden 2.25 Billion Tax on Phone Users Implemented for Benefit of LECS p. 265

Part Six: Internet Governance or Management? A Still Unresolved Crisis

    Executive Summary p .269

    Jan. 99: ICANN Constitutes Itself in Order to Apply "Adult Supervision to Internet -- Esther Dyson Acts as Mouthpiece for Shadow Cabinet -- Mike Roberts Transmission Belt for ISOC & IBM Wishes -- Backing Interests Threatened by Ubiquitous TCP/IP, ICANN Prepares to Overturn the Values that Made Possible Currant American Dominance of Telecommunications Technology p. 273

    Dec. 98:New IANA Formation Ending Amidst Uncertainty, Congressional Inquiry of Magaziner & the Death of Jon Postel -- ICANN Supporters Use Closed Process to Create and Sell Corporation That Now Must Operate Without Jon - Modification of ICANN Bylaws is likely Outcome p. 282

    Nov. 98: Magaziner tells COOK Report: at Least One More Draft Needed -- An Overview of Recent Events as Creation of New IANA Enters Home Stretch p. 302

    Oct. 98: Contemplated Names Council Putsch: Is This Postel and ISOC'S Understanding of Democracy? Does IANA Do Open Processes? p. 306

    Sept. 1998: Magaziner Explains U.S. vs. NSI Position -- Negotiations' Intent Not to Dismemeber NSI But to Protect Interests of Registrants and Competition NSI Presents Appearance of Arrogance & Communicates Poorly p. 309

    Aug 98: Behind Closed Doors & With IANA Backing USPS Floats Plan to Take Over .us TLD -- Kahin and IANA Support Attempt to Offer All Citizens Electronic Mailbox in Context of Continued Non Disclosure by Burr's Task Force and ITAG p. 313

    Jun 98: ARIN Critics Need to Realize There Is Much More at Stake p. 318

    May 98: New pgMedia Attorneys Appear More Savvy than Bode - Yet Their Technical Analysis is Weak - Difficult to Take Garrin's Case Seriously p. 319

    May 98: US Department of Justice Fails NSF in Court -- March 17 Hearing Overview & Survey of Some DNS Court Actions - Magaziner Sells House: Pledges to Remain on Job p. 321

    Apil 98: Internet Facing Governance Gridlock? -- We Trace Two Years of Win/Lose Power Plays NTIA Rule Making Procedure Reinforces Lobbyist's Zero Sum Mentality Endangering Magaziner's Efforts to Find Consensus -- Positions Taken on "Public Trust" Aspects of DNS Not Helpful p. 326

    April 98: Analysis & Critique of Magaziner Direction Laudable Framework and Stated Goals Conflict with Self Interest of Large Players & Leave Little Policy Flexibility p. 341

    Feb. 98; Magaziner Acknowleges That Issue Is Internet Governance (Management) Not Just DNS -- Public Release of Draft Policy Expected on January 23 p. 345

Index p. 349

Contents p. 355


Within the national Internet service provider community, "The IP Insurgency" is intended to enable strategists understand the complexities facing their engineering and operations staff. Among smaller ISPs it should serve as a tool to bring owner-operators, who are busy 18 hours a day ordering lines, installing them and servicing their customers, up to speed on the changes going on in the environment in which they must operate. LECs and other phone companies will also find it useful. Familiarity with the issues discussed within the Handbook will provide their corporate MIS people and strategists with a valuable knowledge base of the issues facing the survival of their companies

However, since these infrastructure issues are also critical to the continued growth and success of the Internet industry, this "Handbook" is expected to be a tool for use by those in the banking and investment community. If those in the financial community understand the changing technical and power relationships in the industry, they will be able to improve the quality of their investment decision making. Finally the Handbook should also be useful to corporate strategic planners who will be advising their companies' decision making in Internet applications for vertical industry markets.


"The IP Insurgency" is priced at $495.00 per copy shipped electronically as an Adobe Acrobat PDF file. It is also available double sided xeroxed, GBC bound for $575.00. A site license hard copy is shipped single sided laser printed.

Current site license subscribers to the COOK Report may have site license privileges for the desk top published hard copy at a cost of $600. Those who are not subscribers to the COOK Report may save 5% on the cost of a subscription if they order both at the same time. The PDF format of the report will be available for posting on an employee only intranet for $1500. Those with web based site license subscriptions may order the report for adding to their webs at $1250.

This Annual Handbook is NOT provided as part of a regular subscription.

Volume Three in this series is still available. Building Internet Infrastructure, an Anthology of Articles from the COOK Report on Internet for $200.

Inquiries by email welcomed. But orders should be placed by phone and confirmed by fax.

Please Contact:


(609) 882-2572