SOA - What's Left to Say?

One might indeed wonder how much there is left to say about Service- oriented Architectures. By now pretty much every software product or tool has been SOA enabled. The SOA hype machine has been running at full speed for long enough that we had to add a "2.0" suffix in order to do it justice. We understand that loose coupling is the way of the future, remote- procedure calls are so 1998, and that there is a WS-* specification to solve about any problem, including ones we have not faced yet.

So how come we have a whole track on SOA? Because there is a lot left to be said. And a lot left to be done. Major shifts in development attitudes and programming models take more than a few PowerPoint presentations to take hold. More than a decade after object-oriented programming became popular the champions of this programming model definitely have a trick or two left to teach us. The same is true for service-orientation. The apparent simplicity of service-oriented interaction belies the difficulty of getting it right. We need to understand where the service-oriented model works well and where it does not; how it interfaces with synchronous, object- oriented applications; how to find the appropriate balance between top-down control and bottom-up evolution; how to define expressive contracts. The list goes on and on.

This year's SOA track tackles many of these issues head-on, traversing the complete stack of service-oriented interaction. Replacing vendor hype with real-life experiences we'll hear about capturing knowledge in SOA patterns, orchestration and business process management (BPM), SOA governance, modeling, and using SOA from scripting languages.