Soft talkers unite: Life at the forefront of the softswitching revolution

Several service providers and equipment vendors banded together about six months ago with the idea of getting something done as quickly as possible. Specifically, they wanted to speed up the development of softswitch-based applications that ride on the IP voice and data networks now spreading across the landscape.

Last May, upon launching the International Softswitch Consortium, member companies said they intended to focus on "developing applications for a distributed set of hardware and software platforms that can seamlessly interconnect the traditional telephone network with information and applications currently available only over the Internet."

These software-driven platforms, generally referred to as softswitches, broadly distribute functions currently performed by digital circuit switches. Softswitch architecture includes open, standardized protocols connecting distributed hardware and software; seamless interconnection, at both the signaling and transport levels, of the public network and IP-based networks; and open, standardized application development environments.

Faster innovation for everything

Ike Elliott, who is chairman of the consortium and vice president of softswitch-enabled services for Level 3 Communications, explains the group's mission a bit more succinctly than the consortium's formal announcement did. The basic idea, he says, is "to enable and accelerate Internet innovation of all communications services," with the emphasis on "all."

"The Internet exists, of course, for the Web and e-mail and those kinds of things," Elliott says. "What we're talking about is extending that so Internet innovation encompasses, envelops and subsumes telecommunications in its entirety - telephony and multimedia communications included."

The International Softswitch Consortium is an outgrowth of another loosely knit group, service providers and equipment vendors brought together by Level 3 Communications in the summer of 1998 to hammer out an IP device control specification. Within 60 days, the IPDC Technical Advisory Council produced an interface "that anyone could use to start creating solutions," Elliott says. Indeed, the IPDC Council put the spec out for free distribution on August 1 of that same year, and by September and October, several vendors began to ship products based on it.

Impressed by such swift and tangible results, several members of that group pressed Level 3 to set up a permanent group that would tackle other issues, including softswitch-to-softswitch protocols and application layer protocols that would make the IP device even more useful. The company declined, Elliott recalls, but in the face of continuing pressure from several companies, CEO Jim Crowe eventually agreed that Level 3 would establish a permanent consortium.

With member companies committed to producing tangible results quickly, the consortium within one month merged with the Open Multimedia Protocol Alliance, which had been established by Vovida Networks. The two groups combined their efforts, under the International Softswitch Consortium name, to promote support for five IP protocols: H.323, real-time transport protocol (RTP), session initiation protocol (SIP), media gateway control protocol (MGCP) and real-time streaming protocol (RTSP).

Two months later, in August 1999, the consortium conducted its first big event at Cisco Systems facilities in San Jose, where several hardware and software companies convened to test the interoperability of their IP-based voice products.

Elliott points out that the Softswitch Consortium does not duplicate the efforts of the iNOW! initiative, an industry effort to ensure interoperability among H.323-based products. Although H.323 is one of the consortium's five "foundation" protocols, the consortium does not focus much on H.323 interoperability testing, he says.

"We think that space is pretty well covered by iNOW!, and they're doing a good job of that. Why would we want to compete with someone who's doing a good job? We want to add value where it's not already being added," Elliott says. "The SIP and MGCP protocols needed another kind of outlet."

Nor does the consortium act as another standards-setting body. The problem with standards, as consortium members see it, is that for any given standard, various implementations of it emerge among various vendors.

"Without some sort of structure for allowing people to bring those implementations in and see if they interoperate and to fix them, the standards are of limited usefulness," Elliott says. In other words, consortium members are not trying to set standards but "to use them to make money."

Cooperating to grow a market

The 10 founding members of the International Softswitch Consortium include: Cisco Systems, Enron Communications, Hewlett-Packard Co., Level 3 Communications, Lucent Technologies, Nortel Networks, NorthPoint Communications, pulver.com, Rhythms NetConnections Inc. and Telcordia Technologies Inc. (formerly Bellcore). Within four months of the consortium's inception, the number of member companies grew to 64 vendors and service providers, representing seven nations.

One of the first to join the consortium was Salix Technologies Inc., a Gaithersburg, Md., vendor that produces "class-independent" telephony switches, service layer devices that integrate all the service functions of both voice and data networks. Such integration is necessary to allow service providers to roll out new offerings as fast as possible.

Lew Bobbitt, vice president of marketing, says participating in the consortium enables Salix to track evolving standards and, during the evolutionary phase, build products that not only comply with those standards but also work with other vendors' products to produce a solution.

"By working cooperatively to ensure that products can function together, we can move solutions into the service providers' hands more quickly than [we could by] working strictly from the standards," he says. "By getting solutions to the service providers more quickly, we can accelerate the growth of the market."

In fact, the industry is evolving so much that Salix never considered not joining the consortium, Bobbitt says. Next generation networking architectures comprise multivendor solutions, ensuring that their products are interoperable is in the best interests of vendors.

"These architectures are becoming distributed in such a way that service providers no longer can rely on a single vendor to provide a solution that covers all elements of a network architecture," he says. "Therefore, until we have standards - if it's ever possible to have standards that ensure total interoperability - vendors have a responsibility to work cooperatively to deliver that interoperability."

Unisphere Solutions Inc. in Chelmsford, Mass., also was among the first members of the consortium. Siemens formed Unisphere earlier this year to develop converged voice/data and Internet networking solutions. Pedro Colaio, director of IP telephony, emphasizes the importance of softswitches - and the Softswitch Consortium - to the growth of the IP services market.

Building bridges between the traditional public network and packet-switched networks requires that the call control layer and the actual switching fabric be separated, he says. That requires softswitches or similar products.

The consortium's overall objective meshes perfectly with that of Unisphere: grow the market by promoting interoperability of next generation network components. Thus, Colaio says his company's decision to join the consortium was a no-brainer. "Given that this market is still pretty small, interoperability is going to be a key issue for wide acceptance [of products] by service providers."

There are some additional, more direct benefits that Unisphere derives from its participation in the consortium. For one thing, Colaio says his company "can feed in our views of the network architectures and discuss them with peers - vendors and service providers. Secondly, it gives us the opportunity to get working-level agreements with some of our competitors, to make sure our products work together in the way they're supposed to. And it exposes us to what everybody else is doing, so the consortium is an important mechanism to get feedback from the marketplace."

A revolution with rules

Although individual member companies may have their own reasons for joining the group, Elliott contends that one factor is common to all members - frustration with traditional ways of trying to get things done.

The consortium really is "a revolution," he says, one that intends to tear down a lot of the industry barriers to real competition. Companies are tired of "the fear, uncertainty and doubt that have been sown. As an example, in the H.323 arena, people don't know that they can use it without licensing it from someone."

To end the confusion and "to create and sustain a fair and competitive component market," Elliott says the consortium has laid out some stringent, possibly unique, intellectual property rules governing its member companies. To wit, each one must sign a royalty-free license grant on the five foundation protocols - H.323, SIP, MGCP, RTP and RTSP. That means a member company cannot charge royalties to other players, even if that member company owns intellectual property involving any of the five protocols.

"If we reach critical mass in the Softswitch Consortium, then people will be able to act with confidence on, for example, H.323," Elliott says. "Anyone can implement it, and no one has to worry about paying a royalty. If we get enough members, that's going to happen. It's one of the things that has prevented the acceleration, the adoption, of these Internet-based protocols for telephony communications."

Defining "critical mass" is not a question of how many members join but of which particular companies sign up. It's important, he says, to recruit specific companies that hold intellectual property in one or more of the five foundation protocols. "Until we get a critical mass of those companies, in terms of almost all of them joining, our mission will not be completed."

Some of the "big players" have remained on the sidelines thus far because they're worried about their intellectual property databases. However, consortium leaders are working steadily to address those companies' concerns. Elliott expects those companies that initially disagreed with the consortium's intellectual property rules to be joining the group during the next couple of months.

In addition to the usual person-to-person networking that takes place at industry dinners, conferences and trade shows, the International Softswitch Consortium recruits new members by sticking to its agenda - produce practical, solid work fast. For example, the group held its second interoperability event in September when it co-sponsored the Voice on the Net developers conference in Atlanta. The Softswitch Consortium provided a day's worth of educational programs, each designed, as Elliott says, to show "how you use all these protocols to make a useful thing."

The consortium plans to conduct another interoperability event in a month or two that will focus on SIP and MGCP interoperability. The consortium holds biannual meetings, and the next one is slated for January 18 in a yet-to-be-decided city. To ensure that new leadership comes in every six months, the consortium holds elections at each of the twice-a-year meetings, with staggered terms for the top posts. Membership on the board of directors is for a maximum term of three years.

Yet Elliott points out that consortium members by no means are sitting around doing nothing between these biannual sessions. A lot of "definitive work" is happening within the various working groups, and each of those is charged with delivering some specific results to the consortium's Technical Advisory Council. Currently, those deliverables include everything from protocol usage guides and architecture documentation to test suites that can be used by vendors that attend the interoperability events.

Given its early successes in promoting cooperation, openness and interoperability, can the consortium possibly prevent the "not-invented-here" proprietary mentality from invading next generation IP-based networks? Elliott says only that "for people who are doing business in the softswitch way," rejecting the not-invented-here mentality translates into a competitive advantage.

Well, then, could the consortium turn out to be so successful that eventually there will be no need for it? "We have to explicitly vote in a few years - I think it's five years - as to whether we want to keep going. There are mechanisms in our charter to disband. I don't think five years is going to be enough, but I'm not sure. We could just surprise ourselves and finish the revolution in five years. But the short answer is, I don't know."