In thinking about the Webinar I'm doing on Friday, I just wanted to put together a list, based on my experiences, of the 5 things you should consider before you start your SOA project. Clearly, not the only things to consider, but as I'm working SOA projects out there…my top 5.

1. Consider the people. Most SOA projects die due to lack of talent, not lack of need or resources. Make sure you have the right people on the job, and they are trained properly. As Ron and I discussed on the Podcast this week, the SOA talent shortage is hurting this emerging area of technology.
2. Consider the Buy-in. Most SOAs only live with approval and support from the top. Thus, projects that don't have sponsors in the right places are doomed, typically due to politics, not technical failure. SOA is a huge change in the way you approach IT, and this change requires resources and support. If you don't have them, don't try SOA.
3. Consider your own needs. Most SOA implementers have tendency to dash out there and select technology before understanding the problems they are looking to solve. Your needs, for your enterprise, are unique and require solutions that are customized for your issues. One size does not all…SOA is something you do, not something you buy.
4. Consider the approach. You can't iterate your way to a successful SOA, and you need a good definition of the project steps before beginning your own SOA project. Typically this means discovering metadata, services, processes, etc., and then defining your own SOA that's meets your needs.
5. Consider the business case. While most people think SOA is always a fit, in some cases it won't have the impact required to justify the cost. Thus, you need to stop and think about the business cases at some point, and this will help you establish creditability with the sponsors, as well as shine a profitable light on the project when asking for resources.

Posted by Dave Linthicum on April 11, 2007 06:55 AM