By Tsahi Levent-Levi

Whenever I visit customers in Asia Pacific, it always amazes me how different companies that are developing the same product can go about it in so many different ways. OK, maybe not exactly the same product. One handset weighs in at 170 grams and their competitor’s handset at 164 grams. One handset could also have a metallic color, and the other shiny gray. Go figure.

In trying to understand this phenomenon, I came to the conclusion that all these different approaches are a result of a fundamental difference in R&D philosophy.

Let’s assume you are a developer, and you have been given the daunting task of developing a mobile handset. And YES; it should be an IMS handset. The applications you need to support promise you a major headache:

– VoIP calls (audio and video)

– Presence

– IM (Instant Messaging)

– VS (Video Sharing)

- VCC (Voice Call Continuity)

By the time you’re finished, you assume the whole world speaks in “acronym”…

And now for the 50 million dollar question: How would you approach this development task?

Rest assured I won’t let you think too much while reading this post. Take it from me… I’ve seen primarily two kinds of companies – the “top down” kind and the “bottom up” kind.

Essentially, if you need to build an IMS client (that’s a mobile phone that does IP), first and foremost you need a solid SIP IMS implementation. Not your average-Joe “IMS-ready” or “IMS-lite” one. You also need engines for all those fun applications that will enable you to do VoIP calls, presence, etc. And to top it off, you need a user interface that can be stitched into Windows Mobile, Symbian or whatever operating system you’re using.

Illustrating it in a diagram, you get the following result. If you are one of those “bottom up” companies, then you believe that infrastructure is the key. You search for the right SIP stack, with all those nasty IMS extensions. You make sure your RTP implementation can handle all those necessary bandwidth and retransmissions for IMS. And you select a good codec vendor that can deliver on the promise of encoding H.264 in CIF resolution using only 80 MIPS (I wish…). Once you have that part figured out, the next step is to look for the engines you need to build your applications!

If you are one of those “top down” companies, you believe that the world is a set of applications you can mix and match… and all is well in the land of computing. You see SIP, RTP and those codecs as commodities. And for you, IMS is just another one of those things that you’ll figure out in the future. So out you go to seek your fortune in the market. You find a slick brochure of engines and applications, and choose what looks nicest. Presto! You have a really cool demo application running that really works. Until you look under the hood or try to plug it into a real network with other users.

In case you haven’t figured it out, I tend to side with the “bottom up” guys. I’ve seen too many projects leave the gate and start out running too fast… just to find out (pretty soon) that they were heading in the wrong direction. And now they are left with a lot of catching up to do.

What kind of company do you work for?