With the big question mark that people are putting after SOA these days, many are viewing SOA with a more suspicious eye. The essence of their issues are around the extreme hype surrounding SOA in the past, the fact that they had to do actual work to make SOA function, and that it was a complex problem to solve, in contrast to the magical things that were perhaps expected to occur when implementing SOA circa 2002.
In looking back at all my books, articles, podcasts, and blog posts, I was clearly more the realist when it comes to architecture, and I understand, as do many others, that there is no free lunch when doing this and that architecture is not easy. I also knew that with some realistic expectations and some well-planned execution, SOA does indeed work.
[ Earlier this year, SOA was given an obituary by Burton Group analyst Burton Group analyst Anne Thomas Manes | Learn all about the cloud-SOA connection. ]
What I see happening now is a movement to what I call "New SOA." Or SOA as it was with all of the hype and unreasonable expectations removed. Also, some of the more silly technology out there finally falling from grace, such as a subset of the SOA governance tech out there as well as old school integration servers and message queues rebranded as SOA-in-a-box. At the same time, some of the technology that held more value, such as process configuration and/or abstraction technology, is having a good run these days.
New SOA is all about doing and not buying. It's all about working from the ground up instead of from the hype down. It's all about learning how SOA works as an architecture pattern and how to recognize and leverage best practices. It's about results and traceable benefits to the business.
I urge you to look at New SOA; it's perhaps a lot less sexy than Old SOA, but it won't break your heart.