IP Multimedia Subsystem—the software our telephone calls have been waiting for

Rethink IT, April, 2005

Someone asked me the other day why his home phone line didn't have text messaging and why he couldn't pick up his mobile voicemail on his PC? These are perfectly valid questions and they sit side by side with 'why can't we videoconference from a new 3G mobile to a home-based TV set?' and 'why, can't photos be sent to a home landline phone from a mobile phone or digital camera?'.

The answer to this and multiple other 'why not?' questions is, of course, because all the networks that this traffic wants to travel over are not built in a compatible way.

Why can't calls find you, no matter what phone or PC you are on? Why can't video signals that are too big for your phone be rerouted to a networked screen near you? Why can't six other people join in a conversation and see the web pages we are looking at? These are all reasonable questions if all of us were using the same networks, or if they were homogenized in some way, for instance all running over IP. But even then a certain amount of standardization of protocol stacks and middleware would also be required.

SOFTWARE TO BREAK THE DEADLOCK

A piece of software that is likely to break this deadlock is IP Multimedia Subsystem or IMS, which is now halfway between its fifth and sixth editions and is finally about to be installed in anger.

Earlier editions never made it into real life installations, but now the scene is set to answer all the above questions with a unified IP system.

Because the standardization issues were dealt with from a mobile network point of view, many people still view IMS as a purely mobile piece of technology, which of course it can't be.

If people want home landline phones to have the same advanced facilities as mobile phones, they have to be built to carry the same services, and in turn the software APIs that enable them need to be implemented in servers on their wireline networks.

CONVERGING NETWORKS

It is pretty much established that services such as video conferencing are far less desirable when only applied to mobile handsets, and far better when the connection is made between a handset and a broadband line attached to a PC or TV screen. So IMS is really about converging networks, especially wireless with wireline and potentially future broadband wireless, with both of these.

Already, [O.sub.2] in Europe and Sprint in the US have major IMS trials, and others, for instance Telecom Italia Mobile (TIM), Vodafone, Cable and Wireless, Belgacom, KPN, TeliaSonera are either about to begin trials or have started trials using IMS enabled equipment from vendors such as Lucent, Nokia, Motorola, Siemens and Alcatel.

The biggest driver for the implementation of IMS for mobile networks is easily implementing new services, in a standard way, to drive up Average Revenue Per User (ARPU), but also to converge services across mobile and wireline networks. As such the natural place to start is among operators that run both kinds of networks.

But at the same time wireline incumbents need to be able to initiate services that both drive up ARPU, as the cash generation value of voice only networks comes down. To do this they need to move into services that offer a high cost of entry for their competitors, which right now are hurting wireline telcos' bottom lines.

It doesn't mean that any telco, whether it's a mobile provider or a fixed network provider, can stop worrying about its raw network.

Anyone wanting triple play operation must still implement MPLS routing to accelerate packets between a service and a DSLAM, it must still offer some kind of tunneling and Quality of Service protocol stacks at the client end to segregate traffic types and offer them guaranteed service levels. Overall network capacity still has to be delivered, whatever network you have built.

Instead, where IMS starts making things easier is at the applications layer, service layer and control layer of a network and relies on the connectivity layer already being man enough for the job.

There are some generic service enablers defined within IMS such as presence and group list management, which are implemented as services in a SIP (Session Initiation Protocol) Application Server.

CONTROL FUNCTION

The control layer is built from network control servers for managing call or session set-up, modification and teardown. The most important of these is the Call Session Control Function, effectively a SIP server. This layer has to contain support functions, such as provisioning, charging and operations management. Interworking with other operators' networks and or other types of networks is handled by border gateways.

In the pre-IMS world, services would be specified and supported by a single logical node, performing specialized features just for the new service you want to launch. Because of this, when services want to operate together (video, web pages and voice for instance in a conference call), then four different service servers have to be made to operate together. This function has to be retrofitted or driven through complex protocols that you install in the network on specialist servers built for that purpose.