Watching the adoption curve of service-oriented architecture is an interesting phenomenon. You get the sense that some IT shops are aching to stuff the SOA genie back in the bottle.
Paradigm shifts are no fun. It forces you to do things differently and you very well might have been pleased with the way you were doing things before the paradigm shift.
The slow, steady crawl of SOA into the IT mainstream aids that sort of willful opposition. Since it's not happening in one fell swoop, maybe it will just lose momentum and you can start to worry about the next buzzword.
The thing is, while SOA most certainly is a buzzword, it's also representative of an enterprise architecture movement that most surely isn't going away. Loose coupling, abstracting business logic from implementation and reusability are here to stay. Call the collection of principles embodied by SOA what you will – bizitecture or, perhaps, Steve – but they're going to be around for a long time.
Dave Linthicum, a well-respected consultant in this field and SearchWebServices site expert, recently caused a stir when he made the following statement: "Five years from now, we won't be talking about SOA… It will all be folded into EA."
Does that give you license to ignore SOA? Not at all. It just means that SOA, or more importantly the principles SOA embodies, will become pervasive. Yet the spin on these things can get severe so Linthicum felt the need to re-explain his comments, writing "So, just to be clear, I'm not announcing the death of SOA, but that it will have a core systemic value going forward….always, within architecture."
Meanwhile, over at his ZDNet blog, Joe McKendrick makes the case for SOA adoption being driven by business opportunity. If this is truly an era of business-driven computing then that makes perfect sense. He notes: "The lesson here is that SOA's value is only realized when the metrics and measurements are put in place. Without measurement, no one knows what exactly SOA is delivering to the business. It could be doubling profit margins — but who would know?"
Wading into the pool, proving SOA's value to the business with every step, might not sound like as much fun as a big plunge, but it would fit right in with a movement that's constantly trying to skewer its own hype.
Just recently, Burton Group bemoaned spend-happy worst practices in the SOA space, while noting it could take 20 years to realize the benefits of service-orientation.
Dennis Byron last week did a nice job of putting the whole SOA matter into perspective, writing: "IT guys get either bent out of shape or a good laugh out of this stuff because basically the IT whole world is going to be based on SOA by about 2019 no matter what, ending a process that began around 2001, just as almost the whole IT world today is based on n-tier client/server computing, almost at the end of a process that began around 1989. No amount of surveys or advocacy groups is going to make it happen any faster or slower."
That's a refreshing way to look at it. The Ventures' classic admonition of "Walk, Don't Run" applies. You can loathe and resist the hype, but you're still likely to end up doing it because it's going to have staying power and people are going to tie it to business value. SOA isn't going away because it never was designed as a thrill ride. It's sober and effective, content to skip the party and instead make you money.
