Are Web services the same thing -- in effect -- as SOA?
Industry analysts and vendors say no, no, a thousand times no. But Aberdeen Group research shows that about 50 percent of those involved in moving to service-oriented technologies say otherwise and, in fact, are doing SOA via the Web services route.
The other half move to SOA by investing in middleware infrastructure: The ESB, registry, governance software and so on.
Which is right?
You're probably expecting me to be a diplomat and say, "Neither" or "Both." But according to a recent Aberdeen white paper, the smarter choice is to invest in a full SOA infrastructure.
"SOA Middleware Takes the Lead: Picking Up Where Web Services Leaves Off," surveyed those moving to SOA and found that the best-in-class organizations are twice as likely as the industry average to focus on SOA middleware infrastructure rather than Web services.
I mentioned the report in passing yesterday, but after reading the full white paper, I thought it worth its own post today --particularly since it offers some insight into key questions about SOA implementation.
Fifty-six percent of Best-in-Class are focusing primarily on SOA infrastructure, versus 31 percent of the Industry Average. Only 11 percent of the Laggards are building and deploying SOA applications.
Further, 61 percent of the best-in-class have invested in SOA infrastructure middleware.
Only 17 percent said they used a combination. In fact, after looking at the pay-offs of SOA versus Web services, Aberdeen concludes it's best to pick one or the other. Mixing Web services with SOA infrastructure is too complicated, leading to unnecessary stumbling blocks.
That said, I should note that BEA Systems -- a supplier of enterprise infrastructure SOA software -- underwrote Aberdeen's research. That makes me wonder why the report didn't spend a lot of time talking about one finding that supports Web services as a means to SOA: 83 percent of best-in-class companies have invested in Web services-specific application development tools.
What other practices differentiate the best-in-class from the industry average? According to the white paper:
But here's the most telling important finding of the report: It's paying off. Big time.
One hundred percent of best-in-class companies reported reduced application development costs. That's every single one of them reporting that they're spending less on application development. How many of the industry average companies reported similar reductions? Fifty-nine percent. Meanwhile, 23 percent of industry laggards actually saw an increase in application development costs.
More best-in-class also reported improvements in user satisfaction and reduced application maintenance costs.
You'll find a lot more information about what leaders are doing and how SOA's paying off for them in the full report, which you can download for free in exchange for your contact information.