You have to be living under a rock not to see the ripples from Anne Thomas Manes' recent post declaring that "SOA is dead," but that there is still a need for it. In essence Anne was stating out loud what we already knew, but were unable to admit: The complexity of SOA was just too much for most global 2000 companies, and the recent round of budget cuts killed many SOA projects.
Okay, if we're not going to do SOA, than back to the basics I say. I think that it may be a better approach to focus on concepts and approaches that are more simplistic in nature, such as data integration, and perhaps use that as a jumping off point to SOA, or SOA like substance in the future.
Thus, it may be a much better approach to architecture to get your master data management and data integration house in order first, not thinking about more complex architectures, and then move up the stack from there. Indeed, if data worked better in many organizations, the architecture would naturally be more valuable and agile as well. That's something.
So, get back to basics, get back to the data.