You can go read my latest outpourings on IMS over at IMS Insider.
UPDATE: For those with a larger budget, you can get another dose of Martin on IMS in IMS Opportunities and Challenges, published by Informa Telecoms & Media group.
Posted by Martin Geddes at 01:06 PMTrackBack URL for this entry:
http://www.telepocalypse.net/cgi-sys/cgiwrap/mgeddes/MT/mt-tb.cgi/699.
I saw your "insight" into IMS and fundamentally disagree with most of your points.
I'll take it one at a time.
Point #1. Agree and Disagree. Making legacy apps fit this architecture is difficult. Probably better as a green field. It is a shame that so many companies have invested in closed architectures to-date. That will probably have to be thrown away, unless the carrier wants to apply heavy customization. You assume these legacy components will not be replaced; I have already seen legacy messaging platforms being deployed with SIP compliance/interfaces. I am extremely curious about the connectivity/service concept. What are you saying? You have to have connectivity to communicate. If you are in the silo, how can you expect to not be served by the silo? You can open the services across access networks, even visiting ones, but you cannot get around connectivity or other physical limitations.
On the last point, as a customer I want convergence. Am I going to get convergence from Skype? How do I as a consumer benefit from an environment that is such a mess of access dependent, device dependent services? It is one or the other.
It is my hope that the operators deploying IMS will build an open, robust framework. I think they can balance the fat pipe with enhanced service offerings, possibly from third parties. Let's face it, all of your killer apps (don't like the term but it is universally understood) are going to come from a third party vendor. So, what IMS helps providers do, is first of all build an IP network that supports any IP media. It gives them a common framework/standard to measure all network equipment providers. It gives them flexibility to change network behavior, virtualize capital, etc. I do not feel like typing this all out right now...
I do not think the telcos ultimately grasp what IMS means, but I think that some of what you are throwing out there is tantamount to FUD. Ultimately, I have not seen a more impressive concept to deal with converged, session based IP services. There are still a ton of hurdles, but they have all but completely replicated legacy telco operations into the IMS standard and will be able to truly start focusing on next generation, convergent services. Once the major voice, video, and data platforms are in place, we should get to see services converge and it should be very cool.
I am looking for a competitive offering to make me think otherwise, but nothing is emerging.
The comments about lack of innovation on IN are telling. There have been APIs for years (Parlay etc) that have attempted to bridge the divide, but the innovators aren't interested.
It's like a government department trying to recruit creative graduates by offering them a great pension plan! Just not going to happen...
I am of the opinion that the main reason for lack of new features in IN is handling the feature interaction problem. Even though a an individual feature could be developed in isolation, a service provider has to agree to do "regression testing" to ensure proper feature interaction is maintained. This effectively kills the possibility of "thousand flowers" blooming. By the way this has nothing to do with intelligent vs. stupid network. The so called client-server SIP will also have the same problem, until and unless, the features are developed at the end-points, in which case there are revenue opportunities for the service provider.
Posted by: at April 26, 2006 12:12 PMAswath makes a good point that I was not able to articulate. IMS gives the framework to deploy traditional telco services as data services. Therefore, intelligence within the network is heightened and we can become more interactive. I can now deploy a network address book (relatively simple), etc. Traditional telco services are hard to integrate into other environments, beyond basic triggering. Now that we can deploy voicemail on IP, maybe we can wrap many more functions into it, just an example. The previous telco model tied functions directly to network equipment providers, and it was very hard to break them up into discrete funtions that could be manipulated.
Posted by: at April 27, 2006 03:55 PMMartin,
I agree with a lot of your thoughts about IMS. I recently spoke with an operator who has IMS up and running for Video calling. They said it was great but few people used it. The next IMS service they wanted to launch was Push To Talk. They thought this may be more attractive. Thinking they were on a roll I asked "What comes next?". Answer "Oh we do not know, we will wait for our vendor to tell us". Now if that does not sound like the IN days I do not know what does!