SOA adoption is a little like love and marriage, suggests Peter S. Kastner, vice president of enterprise integration for the Aberdeen Group. It starts out with the glow of romance and ends up with a reality that while it is far from perfect, it doesn't mean it can't be made to work.
As the lead for Aberdeen's fact-based research on SOA, the analyst looks at data he has gathered from a thousand-plus companies and concludes that if 2005 was the year of SOA romance, by the end of 2006, the honeymoon is over. The promise of SOA cost-savings remains its biggest lure, he said, looking over the research he has gathered during six trips out into the field where he visited IT professionals who are actually trying to make SOA work.
"The cost of integrating applications is roughly 40 percent of the IT budget," Kastner said. "That's 40 percent of a $1.3 trillion global IT budget. It's just an enormous factor in overall costs."
Millions of lines of COBOL code are becoming more and more expensive to maintain, he said, adding that in many corporations, "COBOL has reached the breaking point and beyond."
One popular answer is composite applications.
"The answer plain and simple is that if I can keep the business logic of how to do the transactions, how to do the database in my existing legacy code, and then with SOA and Web services turn it into a composite, I can extend the life of that application indefinitely," he said. "I can clean up the front end with a variety of new Web development tools. For example, I get rid of all of the user inefficiencies of having to deal with multiple green screens and give people the rich Windows user interface that they are used to and can get a lot more information per square inch of screen real estate."
Composite applications sound good, but Kastner says that approach is basically "putting lipstick on the SOA pig." It doesn't deliver on the full-promise of SOA, but for many IT departments that are just beginning to work with Web services applications, it may be their most doable option.
He notes that most of the IT departments he's visited are in the first year to six months of trying to work with the new technology. In many cases, they do not yet have the skills and knowledge to build a complete SOA infrastructure, but he found some that have made meaningful progress. While it is difficult and will take years, a big payoff looms for those who put in the effort. Kastner found some leadership in the banking industry.
"They are buying not only into the lipstick approach of putting a pretty face on an old tired application with Web services," he said, "but they're also moving toward a much more controlled, governed and smaller code base by populating a services-oriented architecture."
He said one New York banking institution, which demanded anonymity before taking part in his research, is in the midst of a major project that involves replacing 1,500 individual applications, representing literally billions of lines of code.
"When they get done a year or two from now, they will end up with 200 services and an order of magnitude less code," Kastner explained. "They should see dramatically decreased software maintenance costs and hence lower application development lifecycle costs. That is a huge driver for companies."