I spotted this article in Redmond Developer by John K. Waters, outlining the outlook for SOA in 2009. This publication focuses on the Microsoft world.
As Microsoft steps up talk about providing tools that let organizations build service-oriented architectures (SOAs), it appears SOA adoption is on the verge of decelerating, thanks to the year-long recession that became all but official earlier this month.
SOA projects are particularly at risk of being cut from budgets due to the difficulty in determining the ROI (return on investment) of such projects. One in five developers actively working on a SOA or Web services app cite providing ROI as the most challenging part of a SOA project, according to a study by Evans released last month.
The article is largely about the inability for IT to define and communicate an ROI around SOA, typically why the projects are underway. For that, I can't disagree. However, this is not the fault of the approach or the technology, but the fault of the SOA practitioners. Indeed, you need to define the ROI up front, and it should be a gating factor for moving into SOA in the first place.
Just to be clear, the ROI for SOA is agility, first and foremost. Defining the value of agility is difficult, but not impossible. Indeed, I've written about the process of defining the value of agility many times. Thus, if the value is not there, you don't move forward with the SOA.
The issue is when SOA practitioners don't do the ROI analysis at the early stages and, when money is tight, can't respond to the stakeholders with a business case, and they get their budgets pulled. Thus, money and time wasted, and not much to show for it. This is occurring now, and will occur in 09 as the downturn continues.
The advice that I have is pretty clear:
- First, define the business case and ROI up front and make sure you communicate with those funding the project.
- Second, focus on shorter term tactical wins, with a clear business benefit, that moves you in the direction of SOA. Those that take on too much, and thus it takes too long, are at the largest risk of having their budgets pulled.
- Finally, if your SOA project will be shut down, take steps to minimize the damage, such as documenting what's been done, and how to pick up the pieces when times are better.
2009 will be just fine for SOA, if you do your homework.