In a space of just three weeks, operators AT&T, BellSouth and Cingular Wireless, three companies already financially linked, announced to the world that they would further cement their bonds technologically, inextricably linking their networks with a common IP multimedia subsystem, or IMS, architecture.

Though each says it chose its vendor independently, the fact that each selected Lucent Technologies isn't lost on anyone — particularly those carriers themselves. With a Lucent architecture in each core, the convergence opportunities are enormous. In theory, all IMS solutions are interoperable — the entire foundation is built on the concept of open interfaces and a common session initiation protocol (SIP) — but there is a wide gap between theoretical and actual interoperability. The three carriers will have identical frameworks that, though they will use them for separate purposes, were intended to communicate with one another.

Although AT&T, BellSouth and Cingular acknowledge that convergence potential, they aren't prepared to tear down the boundaries between wireless and wireline just yet. They have their immediate sights set on implementing new next-generation services, application by application. These carriers are changing the conceptual and physical make-up of their systems' architecture and the fundamental business models that accompany them. It will require years of adding new infrastructure — application servers, gateways, softswitches — bringing access gear and handsets up to a level where it can handle the network's new features; and testing, planning and eventually migrating the services and subscriber data for their millions of customers onto the IMS core.

“IMS isn't something that's built and done — it's a framework,” said Bill Smith, chief technology officer for BellSouth. “We have the foundation, but we'll be constantly building on that framework.”

Each of the three operators has its own plans for its first IMS application. Cingular is planning a major launch of peer-to-peer 3G communications application by mid-year. Cingular Executive Director of Technology Strategy Kelly Williams, however, is a bit cagey about what exactly that app will be. The application promises to be a major step beyond mobile IM, though. Williams said Cingular is implementing the service over its new UMTS/high-speed downlink packet access network, issuing new handsets loaded with SIP stacks and IMS clients.

“That's the way our plans are playing out so far,” Williams said. “We're planning to roll out a usable working application to our customers that will have an immediate impact. This is not a bare-bones release.”

AT&T and BellSouth will use IMS for their next-gen wireline networks. Chris Rice, AT&T's executive vice president of network planning and engineering, said AT&T plans to link IMS to its Project Lightspeed fiber-to-the-node buildout. The operator will initially move its residential voice-over-IP (VoIP) customer on Lightspeed to the IMS core and then transition its business VoIP customers and AT&T's CallVantage residential and Call DNA hosted-VoIP services onto the same platform. From there, it's a matter of migrating its legacy service customers onto the IMS core.

“The nice thing about IMS is it's access-agnostic,” Rice said. “All it knows is it has a connection.”

As AT&T is already offering VoIP across its many channels, the impact of IMS may not be readily apparent to its customers, but the operational advantages it attains will be significant. In addition, as AT&T migrates those VoIP customers to IMS, it can launch new SIP-based services easily, adding application servers to the IMS architecture and extending those features to all of its VoIP customers.

Although there will be some service enhancements noticeable immediately, such as caller ID on the TV screen, the biggest impact of IMS won't be felt until AT&T begins its first wireless convergence projects with Cingular next year. Rice said AT&T and Cingular are already mulling over applications that would allow AT&T to extend its Lightspeed TV services from screen to handset, basically allowing a customer to take his or her cable subscription on the road.

“That's the real goal here: IMS is truly the anywhere, any place, any device service,” he said.

BellSouth is looking to offer wireline/wireless integration as early as this year. Starting with basic features like a common cellular/landline voicemail box, BellSouth's plans likely will expand into big-ticket items like hand-off between the Cingular network and the home wireless local area network, Smith said. Eventually, that service could evolve into a monumental SIP stack, effectively erasing the distinction between the home and wireless phone. A customer could have a single number, but the network intelligence could determine not only where that number rings but under what circumstances, Smith said.

Once the basic framework is in place and the initial applications are launched, AT&T, BellSouth and Cingular are expected to ramp up new services quickly. Although IMS requires investments in all parts of the network, building upon an initial application is relatively easy — it's all software.

A push-to-talk (P2T) app becomes a push-to-show app with the addition of a video server. Add presence software and a peer-to-peer link between those two applications and you get video conferencing. Bridge that peer-to-peer link to a set-top box, you have cross-network video conferencing. AT&T, BellSouth and Cingular all have the critical core layer of their architectures in place. It's time for them now to look toward the other layers of their network.

BellSouth, AT&T and Cingular have all agreed to deploy the central elements of an IMS architecture, the home subscriber server (HSS) and call session control function (CSCF) that make up the control plane layer of an IMS architecture — in this case, the Lucent Unified Subscriber Data Center and the Lucent Session Manager. The control plane is the central hub of an IMS-powered network, the CSCF using SIP to manage every connection and transaction and the HSS acting as the central repository for all subscriber data. Basically, HSS lets the network know if any given customer has permission to do something and, if so, how he or she likes to do it. Once that permission is granted, the CSCF directs the network to do it.

Beyond that “brain” plane, however, there are two more layers to IMS: the network plane and applications plane. The network plane is where all the softswitches, gateways and the actual infrastructure itself lay. The applications plane is the domain of every feature, application and service on the network. A fully functioning IMS architecture spans all three — the control plane pulls features and services off the application plane, recombines them to form the desired service and then sends that app across the network plane to the intended customer.

But here's the rub. Although many carriers have announced deployments of specific IMS elements, no carrier has achieved a full IMS architecture. European carriers, which have taken a very application-centered approach, have announced specific IMS services, such as video conferencing or P2T, but haven't revealed their plans for the core control plane elements of the network. Others are feeding SIP into the network or are rolling out VoIP services, critical functions of IMS but not an IMS architecture unto themselves.

A lot of this can't be helped. Some of the key network IMS elements, such as the new mobile softswitches based on Advanced Telecom Computing Architecture (ATCA) standards, are still in development. Many developers are waiting for standards and carriers to commit to IMS before releasing IMS-enabled versions of their applications. Whatever the reason, carriers are putting IMS together piece by piece.

BellSouth's Smith said some might consider that approach cautious, but in reality it's just the natural evolution of the technology. IMS as a whole and its individual components are still in various stages of development and standardization, he said. By starting with a few key elements, he added, BellSouth can start an IMS foundation and build off of it no matter which direction the technology goes.

“There is a lot of stuff out there that is pre-standard, and we don't know what it will look it,” he said. “But as long we get the framework right, we don't have to worry about it. We can just upgrade with software in the future.”

Venture Development Corp.'s telecom practice has developed a rating scale from 0 to 3, tracking the stages of IMS deployment. Zero means no IMS elements, while 0.5 represents the first SIP applications eking into the network. The first real stages of IMS, 1 and 2, represent the first IMS-compliant applications installed and the first migrations of subscriber data to the HSS. Stage 3, “ideal” IMS, signifies a complete transfer of all applications, signaling and subscriber data to the new architecture. VDC has North American carriers firmly in the Stage 1 or pre-IMS phase at the end of 2005, but between 2006 and 2007, they'll have their first IMS-compliant applications in place and in 2008 will have reached stage 2 or “real” IMS.

BellSouth, AT&T and Cingular definitely seem to be adopting this transitory approach — all of them are starting with the control plane and working outwards. Lucent was named as possible vendor for the switching and gateway products as well as its presence server, but the three operators haven't committed. So far, BellSouth has purchased Broadsoft's feature service, and Cingular is getting its mystery application from Lucent, but because it's a peer-to-peer app, it doesn't require an application server.

Although IMS promises transparency, there are still areas of opaqueness the industry hasn't quite worked out. The 3rd Generation Partnership Project standard for IMS requires open interfaces between all network elements and establishes SIP as the underlying communications protocol, but beyond that, it leaves much room for interpretation. The result has been a wide array of different IMS architectures, each with different degrees of transparency. They all have open interfaces, but some of them are mere open interfaces between black boxes — they can communicate with one another, but no vendor has any insight into what's going on inside of another vendor's box, said Bill Gough, telco industry business manager for Sun Microsystems.

That's why carriers are pushing for as much specificity in the IMS standards as possible, Gough said. It's also why some vendors, like Nortel Networks and Alcatel, are combining ATCA standards with those of IMS. On paper at least, they're promoting an architecture that has transparency and interoperability within the individual IMS components, not just between them. The problem with that approach is ATCA components such as the new lines of mobile softswitches are still in the development phase.

As those applications servers become available and carriers start installing them, conventional wisdom would have it that carriers start transitioning customers application by application. But conventional wisdom doesn't quite seem to gel with IMS. AT&T, BellSouth and Cingular are planning to maintain two distinct but separate networks, migrating customers gradually from one to the other. Thus, they'll approach the transition to IMS customer by customer — or rather customer category by customer category. It doesn't make sense to transfer one service like VoIP or IPTV or 3G data over to the IMS framework only to leave a customer's other services on the legacy systems, or subscriber information spread across multiple databases like BellSouth's Smith said.

“If you can cut the link between the next generation and the legacy network, then you have a much easier transition,” Smith said. “Otherwise, you have to make the new systems talk to the old systems. It takes a lot of money linking the old to the new, so instead of trying to bolt IMS to our legacy applications, we're looking for a clean break.”

So as the early adopters sign up for next-generation services, BellSouth and AT&T manage them entirely over the IMS framework. As those carriers begin migrating whole application platforms such as VoIP or IPTV to IMS, they'll move those service subscribers to IMS databases until they have moved all legacy services and systems to IMS.

Most customers won't even realize their service is on a new architecture. IMS isn't like a new network box or application a carrier can implement and market to customers. It's a new way to deliver services that in order to reach its full potential will have to turn fundamental telecom business and billing notions on their heads. As Smith pointed out, everyone's going to “get” IMS, regardless of whether they're aware of it. Yet, AT&T, BellSouth and Cingular are making multi-billion dollar investments to reconfigure the basic structures of their network. They want to get their money's worth.

John Giere, Lucent chief marketing officer, said that's why the transition to IMS will be gradual, not because of technology constraints, but because of cultural ones. IMS, he said, is the technology face of a larger trend: the transition from a service-model based on time to one based on subscription. As IMS blends numerous applications and features together, charging for each individual feature becomes not only a challenge for the carrier but a headache for the user to track. Giere said services will be sold as lifestyle-based models, allowing customers to access pools of related and closely linked services for a flat fee. Though IMS is not only the preferred technology to handle such a model, he said, it may be the only technology that could pull it off.

WEB EXTRA

Read about Lucent Technologies' IMS platform and how Verizon plans to leverage IMS.
www.telephonyonline.com