Bookmark & share

By Thomas Magedanz

Experience in development and provisioning multimedia services in the telecommunications and internet worlds illustrates that it is difficult to find the right compromise between user adoption and revenue generation. Typically telecommunication services set a clear quality of service and are designed to generate revenues. Conversely, on the internet most of the user accepted services are free, besides the fact that users have to have some internet access and many services are sponsored by some kind of advertisements. In the telecoms world control is in the network, whereas the internet it is the end systems. So far, the internet world has been more successful in gaining user acceptance.

IMS represents a "natural" system evolution, that is, a combination of internet technologies and protocols with well-known Intelligent Networking concepts from the classic telecommunications world. In this regard, the IMS primarily defines a harmonized session control and AAA overlay architecture based on "carrier grade extensions" of flexible general purpose IETF protocols, such as the Session Initiation Protocol (SIP) and Diameter protocols, to maintain control of IP-based service provision on top of mobile and fixed networks.

To respect the competitive nature of the emerging multimedia services world, no specific services have been standardized in the IMS on purpose.

Therefore, we can say IMS today defines just a "docking station" for application servers, in order to control IP-based communication and information services and implement applications flexibly, to meet any emerging service demands. The SIP-based ISC interface for session control and the Diameter-based Sh interface for application server administration, as well as the Diameter-based Ro and Rf interfaces for online and offline charging, provide a minimum standardized "plug and play" capability so as to avoid a vendor lock.

Furthermore, the IMS standards give some indications of what type of application servers can be plugged in, such as legacy - and thus not really sensible - Intelligent Network/CAMEL service control points (connected via an ISC2INAP/CAP adaptor), emerging (but still unproven as carrier grade) SIP application servers typically based on SIP servlet technologies, and so-called open API service platforms such as Parlay, OSA, and JAIN platforms, enabling a protocol-agnostic, and thus seamless, implementation of services for most recent information technologies. Furthermore, the later technologies are the only ones which enable third party programming of network services and try to meet the mind set of the existing application programming community in the definition of these interfaces.

Anyhow, regardless of which combination of these is used, some important challenges remain. It is clear that we will not find in the near to mid-term some kind of commercial killer application, justifying the introduction of the IMS platform itself. Rather we will see - as we can learn from the internet - a myriad of different, highly customised services. This means that the IMS can only become a success if it can support many different services to its users. This raises the question of how to avoid emerging feature interaction problems in a world of multimedia service diversity. Furthermore, the current adoption of service-oriented architectures, associated web service technologies and the emergence of Web 2.0, means that services are created by "orchestrating" service components provided by potentially different providers. The success of the IMS - driven finally by the provision of successful services - will depend on the positioning of the IMS in this overall trend.

This means that fixed, mobile and converged network operators have to start thinking how they can provide, based on IMS and its desired flexibility in application server design, so-called service enablers to the application developer community. It would be fatal to believe that with IMS (instead of IN and OSA/Parlay as in the past) the network operators can develop some kind of multimedia services supermarket with dedicated partners in a walled garden approach. Rather, an early and open promotion of available service enablers to a variety of market players would be the right approach. Otherwise, thanks to IP technology and always-on flat rates, the internet community will design their own multimedia services.

The activities of the Open Mobile Alliance (OMA) in the context of the OMA Service Environment (OSE) seem to be quite promising, as they start from the OSA, Parlay, JAIN thinking but take into account the trend towards policy based service (enabler) provisioning. This is paving the way for a new wave of thinking about how operators can offer access to their assets to third parties.

Finally, coming back to the main question: Can the IMS architecture deliver real network programmability? But what is the network? An IP-based carrier network? An access network - to the internet? The internet itself - as a set of interconnected networks? However, in principle, yes! SIP and Diameter are powerful and flexible enough to provide for control and flexible charging of any voice, video, messaging, or content sharing sessions, as well as enabling group management, presence and location services on top of any IP-based network. In contrast to any legacy network programming approach, the IMS provides the necessary flexibility and openness.

However, the major challenge is the internet as a programmable network itself! And here the question is - who is finally programming it? And who is able to develop revenue bringing services? And what will end users finally pay for? The IMS is a reasonable, but probably the last, unified approach of the telco industry to implement a controllable internet. If the IMS fails, the open internet and related application and content providers will take over and the network operators will provide the bit pipes. But for sure, not for free.