What? Linthicum is talking about canceling SOA? How can this be?

As we discussed in previous blogs around defending SOA during budget cuts, there are indeed instances when a SOA project should be shut down in the interest of the business. This does not mean that SOA is bad, nor enterprise architecture, only that sometimes people do SOA for the wrong reasons, and there is no clear benefit.

Many SOA projects are created out of hype, not need. Clearly many enterprises are "managing by magazine" and are more concerned about the doing something cool rather than doing something helpful. You know the difference, and I'm sure there are both types of projects in your organization today.

So, should you cancel a SOA project? Here are some tests:

First, no strategic vision around how that projects fits in the overall enterprise architecture. This is the big one. If there is no roadmap, no plan, no vision, you're off to a false start. Better stop, do some planning inclusive of understanding the business, then restart later. Typically this means the first project was moving in the wrong direction…what you get when taking a shot in the dark.

Second, it's not SOA. Many projects are called SOA, but are really not SOA. Often I see portal development projects that are funded with the title "SOA" but have nothing to do with services, agility, reuse, or architecture, they are just portal development projects. Portals are great, and valuable to the business, but they are not SOA unless part of a larger plan and strategy. It's architecture, at the end of the day.

Finally, not enough resources. Many SOA projects are under funded, typically because they are considered strategic and long term, with more resources directed at short term and tactical. I would rather cancel the project and start again later, than have a lot of well intentioned people break their picks on making changes to the architecture that will have any value.

Hopefully you're not working on a SOA project that needs to have the plug pulled, but if you are you may have some tough calls to make.

Posted by Dave Linthicum on January 26, 2008 05:03 PM