Now that we know what fails, let's talk about what works.

• The enterprise has invested in education. The professionals who are building the SOA are schooled in the approaches and technology needed to solve the problem. You can count on spending up to $ 30K per person to get them up-to-speed.

• The enterprise has invested in hiring the right people. Many organizations attempt to leverage their existing staff to work on their SOA, not considering their aptitude for the approach. No matter how much training you provide, some just don't get it. Investing in staff is as important as investing in technology, and the best and the brightest—albeit, they cost a bit more—payback huge dividends.

• The enterprise understands their requirements as well as their business case. The requirements gathering portion of understanding your SOA domain is perhaps the most important thing you can focus upon. Once understood, you can back in the appropriate technology set to solve the problem (from above). Moreover, when creating a business case, you also understand how much value the SOA is bringing into the organization.

• The movement toward SOA (notice I did not say 'project') is driven by a commitment from the top. Nothing is worse than attempting to do something innovative in a highly territorial environment that spans into those environments. SOA spans territories. Thus, you need buy-in from the top of the organization, and they must have the political will to embrace change.

• A proof of concept is leveraged to validate the technology. You don't know if a car works for you unless you take a test drive. The same can be said about SOA technology. You must also put performance and security into the mix.

So, learn from the mistakes of others, and from the successes as well. Go do good SOA.

Posted by Dave Linthicum on August 21, 2007 03:15 PM