Bad Reasons and One Good Reason Not to Embrace SOA


Recently, I asked an IT programmer, who wished to remain anonymous, what his company thought about SOA.


"My boss can't spell SOA," he said. "He's a COBOL programmer who hasn't written a line of code in 10 years."


I suspect there are more than a few companies that aren't adopting SOA for the same reason-a lack of awareness and a general aversion to change. But even if you can get your organization beyond those barriers, there's still a long list of reasons why companies don't adopt SOA, as a recent ebizQ discussion shows. (I was going to post there, but they changed their log-in process, and I just didn't have the time or patience to jump through registration hoops today.)


The list ranges from "a bad image" and "over-hyped" to the fear factor.


"Business and IT managers are the major inhibitors because they are afraid of SOA, which highlights efficiency (or its absence) of their work," writes blogger and enterprise architect Michael Poulin.


Tarak Modi, an enterprise architect and consultant, lists not one, but six primary hindrances he's seen over the years, including power struggles, failed bottom-up implementations and poor governance.


I particularly liked Joe McKendrick's point:


"Many companies may think they have 'SOA,' but more likely have JBOWS architecture, or 'Just a Bunch of Web Services.' They may have been led to believe by their vendors that their latest solution gets 'SOA' up and running for them, where actually it is just one piece of the evolution. By the way, there's nothing wrong with JBOWS, it too is part of the evolution on the journey to SOA. The fact is, few companies have actually reached the stage where they have a full-functioning SOA."


You would think vendors would be embarrassed at this point to claim they're selling SOA, when they're really selling Web services but, alas, no. I had two vendor interviews this week in which the company representative actually said they were offering a "complete SOA solution," but when you got down to it, what they meant was they're selling a tool for developing and deploying Web services, with a bit of consulting thrown in.


And, unfortunately, vendors seem to be the primary educators on SOA for many companies. As Chris Harding, head of The Open Group's SOA Working Group, recently explained, this leads to a lot of confusion about what is and isn't needed for SOA:


"... you have messages coming from them (vendors) saying, 'SOA is wonderful and here is our enterprise service bus, which is something you can use to implement SOA.' You come out of that with the feeling, even if they didn't say it absolutely, that all you need to do is buy an enterprise service bus and you have SOA. Of course, it's not like that and buying an enterprise service bus - or a service registry, or an event processor or adopting an event-driven approach - may or may not solve your problem."


It must be really frustrating for companies to be told they're buying a full SOA solution, only to find out afterwards that they're just getting started. I can see why they'd balk at pursuing it further.


Which leads us to another major reason why companies might not embrace SOA: You may not need it. Oh, sure, it might help you achieve your strategic goals and eliminate integration projects down the road. There are great success stories and now we know how to calculate SOA's ROI.


But really, if your current approach is basically working and you're keeping your head above water, it's not like you have to implement SOA. It's not the only option; sometimes SOA's not justifiable.


Nobody's going to take away your birthday if you don't ever embrace SOA.


In fact, for certain, there's one very good reason not to do so: When you don't know what the business case is. As David Linthicum recently wrote:


"Folks, you can't figure out if SOA is going to have any business value without doing a business case up front. This means understanding your core needs and how SOA will create an architecture that solves actual problems, and not just looking to push out an SOA because it seems like the right thing to do."


Eventually, I think SOA will succeed. As McKendrick recently pointed out, even in the current economic climate, SOA is projected to grow 17 percent each year over the next several years. As he puts it, that's not too shabby.


I suspect in many companies, SOA will slip in like a lamb, slowly and quietly.