Replacing the PSTN: How to Get There from Here

Conference debates proposals for a ‘Next Generation Network’ For 19 years, the evolution of networking has been debated in an authoritative yearly conference called "Next Generation Networks". But at this year’s gathering, held September 26-30 in Washington, DC, the future took tangible form in a concept advanced by the carrier-based Interna-tional Telecommunication Union. The ITU has proposed the successor to the Public Switched Telephone Network—a plan also dubbed Next Generation Network (NGN). Conference attendees’ response to the ITU concept was decidedly mixed. Mick Reeve of U.K.-based BT Group, the new network’s most forceful proponent, presents NGN as a "radical plan to close down legacy networks" (PSTN, ATM, Frame Relay, etc.). Their replacement is to be a federated multi-carrier IP-based voice-data network endowed with end-to-end Quality of Service. It is to utilize MPLS and Ethernet for transport and the much-heralded Internet Multimedia Subsystems (IMS) for signaling and "intelligence." NGN aims to unify today’s "multiple service-specific networks," each of which has a different interconnect model, and within which "arbitrage or unfair use abounds," Reeve said. An early goal is Fixed Mobile Convergence—the unification of landline and cellular networks. The new framework is to incorporate PSTN-mandated features like emergency services, fraud prevention, assistance for the disabled, plus billing and inter-provider settlement, and new IP-based applications like multimedia, while keeping out Internet afflictions like spam and cybercrime. "It will have a limited set of services in a protected area, but include access to the Internet where all other services are available," Siemens VP Helmut Schink told conference attendees. IP Telephony- Who Needs It?IP Telephony was little discussed at the Next Generation Networks conference, except for Rohan Mahy’s day-long seminar, "VoIP Update and Outlook." Mahy’s detailed presentation included some harsh words about the IP-PBX: "Users have been underserved by VoIP," Mahy said, "and vendors have not delivered on its promise of more choices. Most enterprise systems are still single-vendor. Lower costs have not materialized. New applications haven’t panned out and neither has interoperability. We end up with VoIP deployments that are basically replacing a piece of wire. Most deployments are not using the efficiency of the Internet, because they create only islands of VoIP, and the calls are still usually going back to the PSTN...."Why do folks buy an IP-PBX? For the new applications? Name one application shipping on a VoIP phone that was not available first on a TDM (circuit switched) cellphone. I asked 250 vendors, and they couldn’t come up with one. Cost savings? There are some savings on costs, but they are minor after including training costs and upgrades needed on the data network. TDM systems also offer Web-based MACs and login on a moved phone. And VoIP moves with 911 are not simpler than TDM MACs. The cost savings are there but not compelling. "Because it’s fashionable? Bingo! Buying TDM now seems quaint. VoIP is hip. That’s what vendors are selling and what the press talks about. No one gets fired for buying VoIP today-although it’s a miracle they don’t."Telemanagement will present its own views on these questions in our January issue. Contribute your two cents’ worth by writing jriddell@decima.comMoney-Routing Protocol The ITU’s NGN is planned as a technological infrastructure for a new business framework called IPSphere (www.ipsphereforum.org). "The Internet has been technically and culturally terrific, but it doesn’t provide a profitable business model," Tom Nolle, president of CIMI Corp., told the conference. "Service providers and Internet players have the same problem: lack of money. It’s as if the ‘money-routing protocol’ got left out of the Internet." By establishing a premium-service networking environment, IPSphere aims to "ensure that any attempt to set up a signal involving significant cost can be mapped to a business order, so the service provider knows it will be paid," Nolle said. It was to be expected that these carrier-centric proposals would run into some resistance at the Washington conference. The names may be identical, but the Next Generation Networks conference (sister to the celebrated VoiceCon show on business telephony) represents a different tradition from that of the ITU plan’s proponents, a tradition rooted in data networking and the Internet. Conference discussion reflected the cultural divide between "Net-heads" and "Bell-heads." Walled Garden?The NGN-IMS-IPSphere projections were repeatedly assailed as representing a "walled garden"—that is, an array of premium services whose pricing is sustained by a barrier against the surrounding Internet. This model, once attempted by AOL and CompuServe, failed in the nineties when the World Wide Web took off. It flourishes today in the form of cellular data services, whose closed environment enables cellcos to rake in $5 billion a year for humble ringtones. Siemens’ Schink said that NGN did indeed aim to create an environment superior to that of the public Internet, but without erecting barriers: "a protected garden with an open door." The IMS framework was also criticized for needless complexity. "I am told that an IMS call from one domain to another goes through 14 interfaces," said Scott Brimm, Senior Consulting Engineer for Cisco Systems. In addition, improving Internet performance may make premium services unnecessary, said conference chair David Passmore of Burton Group. "Packet loss rates are improving by 40%-50% a year; round trip times are improving by 10%-20% a year." If video conferencing works well over the public Internet, who needs assured MPLS? Nonetheless, as Passmore himself noted in the conference windup, "every service provider of all varieties says, ‘Yes, it’s going to happen.’ As a SIP-based overlay over the public network or, more likely, a private IP network, it will be there in a couple of years."Meanwhile, the technological landscape is shifting in ways that may both sustain and confound the carriers’ projections. Consumer as King There was much agreement with keynoter Roger McNamee, author of The New Normal, for whom the consumer has replaced the enterprise as the focus of technological innovation. PCs are being bypassed, he said. "Consumers try to avoid them." They are buying new single-purpose devices, "none of them made by Microsoft." Consumer-driven audio and video "now account for a huge proportion of Internet bit-flow, most of it outside the traditional economy." An increasing proportion is consumer-created, like family photos, videos and podcasts—"we’re going over to an active culture." We can add that major new innovations, which used to diffuse from the enterprise to the consumer realm, now flow the other way, as we have seen with Instant Messaging and peer-to-peer telephony (e.g. Skype). For the corporation, this reverse osmosis can be disruptive. The Future of Voice What is the role of voice services in this new world? According to Steve Kamman, analyst with CIBC World Markets, the trend is toward an overall decline in voice minutes—in Denmark, total wireless plus wireline minutes have fallen 6% in three years. "The quality of the voice experience gets worse and worse" as we descend into "voicemail hell," Kamman said. That sounds a bit one-sided. To this observer, voice seems headed toward incorporation into a multimedia framework that offers practical ways around the voicemail barrier. From BlackBerry phones to the marriage of Skype and eBay, we see more and more combinations of voice with other services. Scot Cleland, CEO of the Precursor Group, sized up this shift as the replacement of telecom by "techcom." "Everyone in the tech sector is incorporating voice," Cleland said. The bond of voice, phone and PSTN is broken; instead, voice is integrated as a subordinate application over devices of every description running over half a dozen different types of physical connections. The carriers are "disintermediated," and the value of legacy telecom facilities falls. When eBay buys Skype, Cleland said, "this is not about telecom; it’s about eBay—more revenue and higher margins in its core business. Similarly, Intel got into Wi-Fi to sell more chips; Microsoft has LCS to strengthen Outlook and Exchange." In addition, some conference participants called into question the future of landline voice. Investment in the PSTN has plummeted, said Kamman; only wireless investment is up substantially. The number of wired PSTN connections is everywhere in decline. "By the time anyone wins the landline Voice over IP battle," said Kamman, "there won’t be any landline business left." Peer-to-Peer One of the NGN conference keynoters was Bram Cohen, the 30-year-old programming whiz who created BitTorrent, an open source peer-to-peer file transfer program now used for about a third of traffic on the Internet. Which Next Generation Network?In his opening comments, conference chair David Passmore proposed five issues that, he believed, would shape the character of a "next generation network." Here they are, slightly reworded. 1. Can the IMS-based Next Generation Network overcome inherent challenges such as: System complexity. Violation of the layered Internet model. Scalability. Unviability of "walled garden" business models. 2. With wireless becoming the dominant form of telephony, with more than two billion mobile phones, will Fixed Mobile Convergence-the merger of wireline and cellular networks-show the way forward? Or is the whole project a last-ditch effort by wireline service providers to prove they are still relevant?3. Video is widely viewed as the next killer application for broadband. Will video- on-demand and customer peer-to-peer video exchange make present-day broadcasting obsolete?4. The rapid improvement of Internet performance is leading to enterprise substitution of access to the public Internet in place of traditional Frame Relay, ATM, or MPLS. How big a market is there for "assured" network services like these, embracing one or many carriers?5. More intelligence is moving into the network, as application processing is embedded into switches and routers. Does this cut against the needs of users and, if so, how do we balance users’ needs against those of their suppliers? Peer-to-peer software has grown both popular and notorious for enabling users to swap music, movies, or software, or to converse, without an intermediary—a capacity often used for illegal downloads of copyrighted material. BitTorrent enables material to be downloaded not merely from a single site but from many computers simultaneously. "A few thousand simultaneous downloads is routine," said Cohen. Launched in 2001, BitTorrent has become an Internet sensation with no funding and no sales revenue, sustained only by user donations. Cohen believes peer-to-peer is reshaping communication. "Does anyone still watch TV any more? I’m not sure how important real-time is any more," he said, conceding it was still useful for "sporting events and disasters that aren’t being handled very well." Certainly, the success of agile and impudent peer-to-peer start-ups contrasts strikingly with the massive investment plans of carrier giants. Peer-to-peer now accounts for about 60% of Internet traffic. Its impact is not limited to the consumer realm. BitTorrent has enabled software suppliers to publish massively without high-volume, high-cost access connections. Skype now runs on many corporate laptops. And Avaya has just broadened its enterprise communications portfolio by buying Ottawa’s NimCat, which makes peer-to-peer call processing software that removes the need for a server. Grid ComputingThen there is the growth of grid computing, which is architecturally a peer-to-peer service. Grids are distributed collections of servers that can be assigned to work on different parts of the same task. Initially used for super-complex computation, such as compiling flight plans for interplanetary rockets, grids are finding their way into mainstream enterprises—for example, to speed drug discovery, or to run large financial models. As Rob Strechay points out in the September 2005 issue of Business Communications Review, grids are now coming into greater use because of storage requirements (increasing 50% to 100% a year) and adoption of Service Oriented Architecture, a framework that unifies Web services and grid computing. The implications for enterprise IT could be sweeping. "Client-server may go obsolete," U.S. analyst Frank Dzubeck told the Washington conference. "Within 10 years almost everything may be peer-to-peer." In the meantime, grids will increase bandwidth requirements. "Corporate bandwidth usage grows 10% to 15% a year; with grid computing, this will change to 30% yearly growth," Robert Cohen of Cohen Communications Group told the conference. Dzubeck said peer-to-peer could triple bandwidth requirements. Management ChallengeWhen enterprises use peer-to-peer, said Andrew Parker of CacheLogic, it is "dramatically different from Internet peer-to-peer," where systems like KaZaA and eMule are designed to be "aggressive and difficult to control." The corporate network seeks central control over applications, balances risk and productivity against cost, and requires assured metrics of service. An application like Skype can bring significant benefits, but it is a "nightmare for IT administrators," Parker says. "IT systems will need to be modified to cope, and in some cases the culture will need to evolve." And what about the vastly increased torrents of critical data surging over the WAN? They will require more attention to Quality of Service and transmission metrics. And here our discussion comes full circle. Disruptive peer-to-peer initiatives might seem to be undercutting the carriers’ model of an organized, regulated networking garden. But at least when transposed into the enterprise environment, they may increase the need for this assured-services model. Back to RealityThe entire discussion at the Next Generation Networks conference concerned the future. We did not hear about a single real-life business user experience. And talk of the future, while stimulating, is hardly conclusive. But two general trends were clear:Voice is becoming more integrated into a complex environment of multimedia and data services. Both the PSTN and the Internet as we know it are headed into a period of sweeping change. These shifts will create new challenges for business communications decision makers in Canada.

Replacing the PSTN: How to Get There from Here

News | 10/26/2005 4:00 am EDT

To keep reading, get a free trial.

Already a Subscriber?

Reuse & Permissions

Inside The Wire Report

Information

Hill Times Publishing

Get in Touch With Us