In the past few months, most of the news about SOA has been pretty positive. Why, just last week I wrote a post asking, "Can We Call SOA a Success Yet?" after an Amberpoint survey showed only 1.5 percent of SOA projects ended in failure, and nearly four out of 10 enterprises report their SOA implementations met all their goals.
"We are close to the point where, as an architect, you need to justify why you would NOT use SOA for the integration portion of a project," wrote an enthusiastic commenter, Kirstan Vandersluis, of Xaware.
But a recent InfoWorld article paints a much different picture of SOA.
Companies spent more than $1.4 million on SOA-related software and services last year, according to the article, which sources a recent AMR Research report.
This statistic from the AMR Research press release is even more startling: 45 percent of SOA adopters reported spending over $500,000 on SOA software and services in 2007. A half-million dollars isn't small cheese, even for IT.
Those are big, scary numbers. But what AMR Research consultants had to say about them in the InfoWorld interview is an even bigger wake-up call:
"Hundreds of millions of dollars will be invested pursuing these markets in 2008, much of it wasted," said AMR analyst Ian Finley. The AMR survey found that most companies don't really know why they are investing in SOA, which Finley said makes long-term commitment iffy.
It goes on to say that a lack of clear purpose for SOA implementations is calling its momentum into question.
Suddenly, January's popular blog discussions about SOA and budget cuts seems much more relevant.
If SOA is going to be move beyond the early adopters, Finley told InfoWorld, the benefits needs to be clarified.
Frankly, I wish Finley had clarified his own point about wasted money. I don't doubt there's waste, but the article does nothing to elaborate. Was the money spent on solutions that didn't work? Did they spend too much on vendor-ware by trying to buy a SOA? Or did they make too many mistakes? Perhaps if we purchase the full report, it tells us -- but even that's not clear.
On a more practical note, the article says the AMR's report found service reuse is not a real benefit to SOA -- just as David Linthicum and others warned. Organizations often focus on reuse because, let's face it, it sounds great: Write this bit of code, bundle it as a service, and you'll never have to recode this function again. But it just isn't working out that way.
OK -- so what can we say for certain SOA achieves? I've seen two concrete benefits from SOA repeated over and over:
1. Agility. According to Finley, the real benefit organizations are seeing from SOA is agility -- a hard-to-measure quality. Basically, Finley said it means companies being able to deploy new IT projects faster. An Aberdeen report last year also linked SOA using middleware -- as opposed to just Web services -- to reduced application development costs. You might want to review that study, as well.
2. Integration. An SOA won't magically solve your integration challenges, but integration -- particularly of legacy systems -- is one of the top reasons for moving to SOA. The Amberpoint survey found that most -- nearly 74 percent -- reported SOA best addresses "inflexible, hard-to-integrate systems." The second choice -- "restricted information flow due to information stovepipes" -- also relates back to integration.
It's probably going to take more time to figure out SOA's definite benefits. Perhaps the real question isn't, "What are SOA's benefits?" but whether SOA survives the crash of its own hype cycle long enough to take root in the IT canon.