Dave -

Saying there is too much technology is like saying there is too much food. What is wrong with having great options? Been to a Whole Foods market lately?

Admittedly many have trouble with too many choices, as obesity statistics will attest. However with respect to SOA, I agree with you that the challenge is not too much technology, rather it is poor strategies for implementing this technology within an SOA.

In December you wrote two blogs on the importance of data in SOA. Steve Lemmo in a recent DM Review article http://www.dmreview.com/editorial/newsletter_article.cfm?articleId=1074470 adds to your point of view.

SOA leaders need to focus their SOA strategies on fundamentals such as data first. Then they can apply the wonderful set of technology options to help them instantiate a data services layer in their SOA that will provide immediate and lasting benefit.

- Robert Eve, Composite Software

Robert,

But there are two issues with the "too much food" analogy which are also IT issues.

1) Obesity - People gorge both on food and IT, they have more than is good for them and they get fat and slow and eating/IT becomes "the thing" to do.

2) Food v Recipes - If you go to a supermarket and concentrate on buying the "best" food this doesn't mean that you will have the best meal. They cook with bits of animals in Southern France that the rest of the western world wouldn't consider to be ingredients but the final taste is amazing, but that is down to the recipe and the skill of the chef. Its the same with IT, its about picking the right tool at the right time, and that means concentrating on the recipe first (i.e. the business) and then going out shopping.

Data is another place the techies feel safe, but I'd argue that again its not a real place to start. Understand the business first, that helps to put everything in context and avoids such disasters as single canonical form being dreamed up by the IT department as a "smart" idea.

"Data first" is a good concept; but SOA should be implemented by looking at business processes first, then ensuring that the right data is being used to maximize ROI on those business processes. Then look at technology to provide the services to expose and share the data to make it all happen.

The problem with the "Whole Foods Market" thought is that you need a grocery list before you even go in; otherwise, you're trying to figure out what you need while you're shopping. This often leads to "impulse buying," and thus my agreement with "too much SOA technology." Very inefficient.

Steve and Don:

You are right. Business requirements come first. But most of those requirements will be fulfilled with existing data, especially in an SOA world built on top of / around legacy architectures. So SOA technlogy buyers need to think about data early and often.

The good news is that the "Data" thought leaders like DM Review are finally addressing the SOA data challenge. And the SOA visionaries like Dave Linthicum are pushing hard on the value of a data abstraction or data services layer. This is where SOA technology investments will pay off quickly, with value compounding with every new project.

Robert Eve, Composite Software