The IP Multimedia Subsystem (IMS) is pitched by proponents as a way for carriers to use IP to better integrate voice, data and video applications across wireline and wireless networks. But skeptics view it as a plan to allow carriers to gain more control over customers and their ability to access applications and services.
Conceived by the Third Generation Partnership Project, IMS is an architecture that essentially takes the place of the control infrastructure in the traditional circuit-switched telephone network, separating services from the underlying networks that carry them. It uses the Session Initiation Protocol (SIP) as its signaling method for setting up calls and handling data sessions, and enables services such as text messaging, voice mail and file sharing to reside on application servers anywhere and be delivered by multiple wired and wireless service providers.
At the Next Generation Networks 2005 conference here this week, several sessions discussed the implications of IMS on current and future networks, including its ability to serve as the foundation of the next-generation public IP infrastructure -- a role now served by the Internet. All sessions, however, quickly evolved into debates on the merits, implications and intentions behind IMS.
"The integration of instant messaging, voice and video will drive IMS," says Sam Christie, chief architect at Nortel, an IMS vendor to the carriers. "It's a good way to potentially open up a network."
Or a good way for carriers to close a "walled garden" around their customers, others note.
"There's a feeling of a need of the carrier to have better control over the customer," says Scott Bradner, university technology security officer at Harvard University.
Bradner notes that the ability of IMS to separate the control and transport functions in a network can also be the hook that requires a customer to enter into a business arrangement with a particular carrier to gain access to certain applications or services.
"Having the carrier involved is not necessarily good for the customer," he says. The end-to-end nature of IP "conflicts with the carrier involvement" that IMS assumes, he adds.
Panel moderator Dave Passmore, research director of the Burton Group, asked whether IMS was merely a marketing and standards game that offered the years-old softswitch architecture as "old wine in new bottles." Nortel's Chrisitie replied that it was only a control plane entity that was deployed only in lab environments and not generating significant revenue for anyone.
"Old wine in new bottles? Partly yes, mostly no," Chrsitie replied. "It allows the incremental addition of new application servers where the softswitch architecture does not."
But the architecture of IMS is overly complex and complexity can be used to lock in customers, says Scott Brim, senior consulting engineer at Cisco. Brim spoke Tuesday at NGN on behalf of the IETF.
"Fourteen intermediaries to set up a call? That seems a little complex," Brim said, referring to IMS. "People would like capabilities instead of complexity. If you build walled gardens, customers will tunnel under the gardens."
The idea of IMS as a walled garden is a misperception, according to Helmut Schink, vice president of systems engineering at Siemens who was speaking on behalf of the ITU as a counterpoint to Brim and the IETF.
"IMS is not closed, it's an open door," Schink said. "It provides some level of protection. The user has the flexibility to leave the protection of the walled garden."