No doubt about it, 2008 was a rough year for service-oriented architecture and the SOA faithful. SOA consultants found themselves in court trying to define SOA during cases involving failed implementations, Anne Thomas Manes of the Burton Group revealed SOA successes were hard to come by and, finally, Gartner revealed fewer companies plan to adopt SOA.
But it wasn't all for naught, as ZapThink senior analyst Ron Schmelzer noted when he shared his predictions for 2009 by e-mail:
The recession really hit hard in 2008 (and hopefully it won't get worse in 2009), but already we've learned a few valuable lessons. First, investing in architecture is investing in improving processes. A few companies rolled out SOA initiatives that actually decreased their overall IT expenditures because it allowed them to get rid of ineffective, inefficient processes.
In fact, I think we learned a great deal about SOA this year-even more than the two examples cited by Schmelzer.
So, before we skedaddle out of 2008, I thought it might be helpful to review seven SOA success strategies learned this year:
1: Take baby steps with SOA governance. SOA governance became a hot topic this year. Several experts suggested you start small to avoid the major power plays that hang up so many governance initiates. Todd Biske, an enterprise architect and author suggested you approach SOA goverance as an education process, rather than an excuse for top-down mandates.
2: Make sure someone's manning the SOA shop. SOA expert David Linthicum first noticed that in many companies, the enterprise architect lacked organizational and budgetary authority to guide a SOA deployment. He and others urged companies to put someone-either the EA or by creating a new, service-manager role - in charge of SOA and give that person the authority to manage services.
3: Form an SOA Center of Excellence. I actually saw this piece of advice repeated often, and I think it goes well with ensuring that someone is in charge of your SOA initiative. An SOA Center of Excellence might fill that role, but more importantly, it will help get everyone on the same page and ensure you learn from your mistakes and successes.
4: Get the CIO tuned in to SOA. A Software AG study on SOA governance revealed that the CIO's office played a role in the SOA steering committees at only 18 percent of companies participating in the study. That's a lot of SOA implementations being done without the CIO's involvement. Talk about a bad idea. You have to wonder how an enterprise-wide approach like SOA can succeed without the CIO's sponsorship.
5: Alternatively, if you can't get the CIO on board, you could get a new CIO. If the Software AG survey indicates CIO participation at large, it's no wonder that research by the Burton Group revealed a new CIO often made the difference between SOA success or failure. Of course, this isn't necessarily a strategy everyone can adopt.
6: Don't pay the middleware middleman. This piece of advice came from Schmelzer's ZapThink colleague Jason Bloomberg, who argued that many companies weren't seeing the agility and decreased integration costs they'd hoped for because they were over-deploying middleware in their service-oriented architectures.
7: Finally, forget the debate about SOA and do what works for your company. That's what Craig Sutter, technical director for VetSource, did. Sutter wasn't overly concerned about whether his SOA deployment fit the strict definition of SOA. Instead, he focused on his company's business needs and how to solve them with a service-oriented approach.
For more ideas on creating SOA success, you can also review Mike Kavis's piece on the eight characteristics of successful SOA implementations.
Despite SOA's rocky year, I suspect it will endure in 2009, particularly if we remember the lessons of 2008.
Of course, it won't hurt that we're also starting to hear more about how SOA can save companies money. Again, to quote Schmelzer:
SOA is not a technology. SOA is not a product. So why people see SOA as a product or technology investment still boggles us. SOA is a means to an end ... a means to a more efficient end. Those companies that can use this time of belt tightening to actually improve their processes will see huge returns and recover what is otherwise lost money. We're already starting to see that.