Featured Story
Deutsche Telekom's 'open RAN' plan slips after Huawei reprieve
Deutsche Telekom had promised 3,000 open RAN sites by the end of 2026, but the date has now been changed to 2027. And Germany's refusal to ban Huawei has implications.
AT&T is prepping a document to define how it will create, manage, and deliver services for its fixed and wireless networks in the future
June 26, 2007
AT&T Inc. (NYSE: T) is working on a monster document to be distributed to IT and telecom vendors that will define the carrier's next-generation services platform, according to Siroos Afshar, a chief architect at AT&T Labs.
But, he stresses, this isn't just another way of talking about IP Multimedia Subsystem (IMS) -- it's much more than that.
Afshar, talking to Light Reading at last week's NXTcomm show in Chicago, said the document will be the second version of one that explains AT&T's CARTS (Common Architecture for Real-Time Services). In basic terms, CARTS describes all the middleware that sits between the carrier's physical network elements (fixed and mobile) and the applications used by the carrier's customers.
The big picture
It's worth noting, before going further, that this CARTS discussion spun out of Afshar's description of AT&T's "network of the future," a network that should eventually allow customers to pick services completely independent of the access technology. This network would also let users switch access technologies while conducting live network sessions. Interestingly, this has been the stock and trade of some recent AT&T commercials, where folks shift effortlessly between watching an on-demand movie at home to watching the same movie, uninterrupted, on a cellphone while in the back of a cab. (See AT&T's Trick Play.)
The new CARTS
The first version of CARTS, a 180-page whopper of a manual, has already been digested by the world's major telecom and IT vendors. But now AT&T needs to revise its CARTS 1.0 documentation to take into account the integration of Cingular and BellSouth into AT&T.
CARTS, though, sounds very much like IMS –- a blueprint for systems that can deliver any service to any user over any access network. But Afshar says IMS is just one part of it.
The idea for CARTS originated years ago when AT&T developed the next-generation network vision that spawned its "Concept of One" idea -- a single, MPLS-based network that would unify all of AT&T's services. (See AT&T’s New Gods.)
"We had created an architecture for infrastructure that could do that. For example, in 2000 we defined RSOIP [Real-Time Services Over IP] to the vendor community. Then IMS came along. Some of the vendors would tell you that AT&T invented IMS three years earlier than the 3rd Generation Partnership Project (3GPP) ," says Afshar.
IMS, though, doesn't give AT&T everything it wants, says the AT&T man, so IMS functions are being deployed for session management within CARTS.
The underlying message here, that IMS isn't going to meet carriers' needs alone, is appearing more and more these days as operators start to come to grips with service development in a world where the likes of Google (Nasdaq: GOOG) and Yahoo Inc. (Nasdaq: YHOO) are emerging as the key competitors to the traditional telecom players. (See Carriers Surf the Web 2.0 Wave and IMS's Web 2.0 Problem.)
So what is CARTS if it's more than IMS? Well, the whole architecture also includes service delivery platforms and OSS -- all the non-infrastructure and non-applications elements that are needed to allow services to be developed and made available to customers.
With that vision in mind, AT&T created its initial magnum opus, describing the architecture and asking vendors to develop the technology that would make CARTS a reality and enable AT&T and external developers to create new applications that run over the architecture.
And there has been early progress. Afshar says two vendors -- one from the IT community, one from the telecom world -- have delivered to AT&T a Service Logic Execution Environment (SLEE), one of the key elements in CARTS.
Afshar says "we have defined a standard SLEE and have communicated this to the major vendors. The SLEE is based on SIP servlets and HTTP servlets. It’s the same applications logic for SIP and Web 2.0 interaction. IT vendors and telecom vendors are responding -- a number have developed the SLEE, and we are experimenting with those. A major part of the SLEE is the integrated way of dealing with SIP and Web services. All of the applications require a session part and a Web interaction part."
Now a team of "several tens of people, not quite 100, from AT&T, BellSouth, and SBC" are working on a document that defines CARTS 2.0 "for the new AT&T," says Afshar. That document, which "describes functions, not products," will hit vendors' desks early in 2008, he predicts.
In the meantime, is there anything applications developers can be working with? Another major carrier that is keen to unlock the potential of the Web 2.0 develop community, BT Group plc (NYSE: BT; London: BTA), has made a software development kit (SDK) available to external developers, giving them the tools to develop services that can run over the U.K. operator's 21CN next-generation network. Does AT&T have the same thing?
"On the SDK, BT is ahead of us," says Afshar. "The issue for us is: How are we going to open up to third-party developers? It's a business and a technical decision. You can't allow third-party applications to come on and take up network resources without it being OK. We have the right functionality in the architecture to enable this -- now the decision rests with the business team. They will decide when the time is right."
— Ray Le Maistre, International News Editor, Light Reading
You May Also Like