Loraine Lawson did a good job in rounding up all of the optimism around SOA lately. Indeed, I've been talking about how the slower economy actually benefits SOA, and other pundits have been saying similar things. Clearly illogical, but it's indeed what I'm seeing out there. However, I can see where I'm asking some to suspend disbelief, but the conclusions are sound nonetheless.
"First, SOA doesn't have a clear return on investment -– or even a clear path to it. Linthicum says he's proven 'many times' that SOA can lower costs over time using ROI analysis -– and he even has a white paper you can download to help you figure your own ROI. But most of what I've read is much less confident in SOA's ability to deliver an ROI, particularly in the short term."
The issue is that the ROI around SOA is complex and holistic, and people typically attempt to calculate it within a particular domain, not enterprisewide and not over a longer period of time. Indeed, if you look at just the cost of inefficiencies within the architecture now, it's typically 20 to 40 percent of the total IT spend, and when you consider the impact on the business the figures become even more compelling. Notice, I've yet to consider reuse or agility. Mileage and ROIs may vary depending on the existing state of affairs, and there are some enterprises where SOA is contraindicated, clearly.
"Second, many IT divisions still aren't really sure how to implement SOA. Call me crazy, but I can't imagine this is the time your CIO will sign off on SOA by trial and error. You could, of course, hire a consultant, which would be advisable, but SOA consultants aren't cheap. And neither is supporting technology."
I very much agree with this. The core issue around SOA is the lack of talent within the existing enterprises and the unwillingness to hire the talent as new employees or consultants. In times like this, it's easy to blow SOA off as something that the company cannot afford, but again they are looking at the little picture. If you manage your resources properly, it's doable and will have some short term impact that can be brought to management.
"Third, just a few months ago, the SOA news was all about Anne Thomas Manes' largely fruitless search for SOA success stories. Now, we've found a few here and there, but that's the sort of thing that's hard to shake. How can you justify redirecting your IT architecture and spending value money -– and let's not forget, time -- on something that's experienced a lukewarm reception, at best?"
Again, good point. What Anne found is that SOA success was dependent on the culture, people, and processes. Thus, if you do drive toward SOA and move toward fixing the architecture, you have to have the proper leadership and talent in place. Those that don't will fail. This is a commonly understood pattern now, both within my research and what's being externalized by Burton, Gartner, and other firms that have been tracking this.
There are lots of reasons to avoid looking at SOA now, but most of them are not valid. Those companies who learn to become proactive around architecture will find a pot of gold at the end of the rainbow. Trust me.