Tuesday, December 27, 2005
VOIP is a giant leap forward in dragging us "Call Processing" types out of the dark ages to the modern world. SIP is the motivation.
While SIP improved the interoperability between systems, it did not take away the inherent complexity of developing voice applications. Many of us VOIP application vendors basically built proprietary constructs from scratch (a HUGE benefit) that helped develop applications rapidly.
The net result is that voice applications continue to be developed by a select few.
Don't get me wrong. We all use the latest and greatest available today: Object oriented design patterns for vastly improved software quality, SOAP for remote data access, CPL or equivalent for simple routing, VXML and MSML/MOML for media control, and a services oriented model for application delivery. However, they continue to be proprietary to the application vendor.
The best we can do is claim SOAP or SIP as the "API" for developing applications. Want to integrate a cool conferencing application as a "converged" application? Sorry, besides SIP "interop", no can do!
Carriers are wising up to this.
They are demanding an open environment that doesn't lock them into a particular vendor. Initiatives like IMS are forcing vendors to rethink and offer an open and integrated solution.
How far away are we from being able to place a call with a simple function call from any computing or communication device? Time will tell.
The best bet the industry has now is SIP Servlets. This technology has the provisions in place for rapid application development, an asynchronous programming model, and state of the art data access. Reputed vendors like BEA, IBM, and Oracle have application servers that are battle hardened in the Web world. They are scalable, robust, and boast of a stable developer community. The "traditional" PSTN community is also attempting to converge on a similar effort with JAIN SIP. This really completes the picture.
The new year looks promising! Is it time to talk about "Call 2.0" ?