I have just gotten back from the Object Management Group’s SOA Consortium meeting. This is a very good advocacy group consisting of primarily corporate customers with a few collection of assorted vendors and integrators. It is a very thoughtful and pragmatic group. I have been a member of the steering committee for the past year and enjoyed participating in the weekly conference calls. We have discussed important issues ranging from maturity models, to SOA governance, to many different best practices. Learning from each other is a powerful theme and focus.
What made this meeting fun for me was being able to share the podium with Sandy Carter, head of marketing for SOA for IBM. We both presented our observations about the direction of SOA for the group. I thought I would share some of Sandy’s comments and then some of my own.
Sandy’s talk focused on two major areas: the learnings from more than 5700 SOA engagements at IBM and the overwhelming requirements for SOA skills and ROI experience with SOA engagements.
So, here are some of the key takeaways from Sandy’s talk:
1.While everyone talks about how programming skills are being outsourced, IBM’s extensive surveys have shown that the an overwhelming number of the respondents found that the biggest inhibitor to the adoption of SOA is the lack of skills — primarily the ability to understand technology tied to the needs of the business. IBM found that 80% of these companies are going to increase their SOA resources this year and 60% of the companies are planning to retrain their existing staff for SOA.
2. Sandy mentioned a new study IBM had sponsored by the London School of Economics that found that companies that implemented SOA increased revenue by 2% — companies that focused on changes in automating business process increased revenue by 18%. Pretty powerful numbers.
3. In another study focused on interviewing hundreds of CEOs, IBM found that SOA is top of mind with the hundreds of CEOs that IBM interviewed. Sandy pointed out in her talk that these CEO from some of the largest companies in the world see the direct link between SOA and business agility.
One of the most interesting issues that Sandy mentioned in her talk was the emerging a Service Science major at major universities. IBM has established a partnership with 200 Universities to help create this new discipline. Teaching MBA students about the SOA principle of creating a service oriented approach to business/technology is critical I would love to see this type of program grow dramatically. I think that many universities are still stuck in the 1980s when it comes to teaching about the intersection of technology and business.
My talk at the meeting centered on SOA lessons from 2007 and predictions for 2008. I’ll give you a synopsis of what I said about this year. My next entry will be a full set of predictions for next year.
My main observation about 2007 was that it was a year of learning about SOA. Our book, SOA for Dummies had just been published at the very end of 2006. It was becoming clear that there was a hunger for information about what this new technology approach was all about. So, 2007 was a year for learning. No market or technology can mature without starting with a lot of missteps. It was also a year when people made lots of mistakes including:
1. Let’s code thousands of cool web services and see what happens — guess what…no one knew what to do with them!
2. Let’s create a corporate wide SOA implementation this year — what’s wrong with boiling the ocean? (too obvious to make a comment on this one)
3. If we implement an Enterprise Service Bus we are all done with SOA…right? — wrong!
4. Hey, we are reusing a service in the same application but we’re not getting very much value….(try reusing in a different application)
Clearly, these aren’t all the mistakes we made in 2007 but it gives you an idea of what I have been seeing. One of the biggest issues that I saw in 2007 was the lack of collaboration between the business and technical sides of the business. In addition, there was simply too much political jocking for control.
But before you think that I am just in a negative mood…I saw many big successes with SOA in 2007. Many companies are understanding that SOA is, in fact, a business strategy based on codifying business services that represent best practices for business policy and process. These companies are taking a long view — not expecting instant results. Many of these organizations are finding strong returns on investments but they would rather not tip off the competition. Before starting one SOA project, our team had to sign three different non-disclosure documents!
I think that we are at the end of the over hyped stage of the SOA market. It is inevitable in any new market that it begins with unreasonable expectations. When customers start using the approach to solve real problems, it is always harder than the hype would suggest. The reality is that transforming software from purpose driven, single use applications to flexible, agile, and reusable services that are loosely coupled is hard work. In fact, the fact that we are getting over the hype phase actually means that SOA is real!