When I review them, I'm reminded that there are two very different views of SOA, which in my opinion are both equally true. I call this the SOA Dichotomy, because these views seemingly contradict one another.
One of the views is that SOA makes things easier for the enterprise. Certainly this view has a great deal of merit and validity. Fully realized a service-oriented architecture allows an organization to fully leverage their investment in the real intellectual property of software - the service that it provides - without closely coupling that service to a single application. It's hard to overstate the impact this can make; it's truly a transformation of the way IT provides information services to the enterprise. Applications can now be assembled, and business processes composed from services and altered rapidly in response to changing business conditions. The workplace can be transformed - workers no longer need to be dependent upon a series of applications to do their work. Instead they can have a single composite application that meets all of their needs without ever having to leave to transfer to another application. Anyone who's ever been on hold with a call-center representative while they said, "Can you hold for a second, that's in a different application," knows how valuable this can be. Call centers know it as well - they know to the penny just how much a single second's delay costs them. Even before an enterprise completes its transformation to a fully deployed SOA environment, the benefits of interoperability and increased agility grow dramatically with every application that is decomposed into services in an effect similar to the very familiar Network Effect. The more services deployed, the more valuable they become. Usually, this is attributed to the fact that SOA makes things simpler, easier to do.
