Hard Lessons from SOA/Web Services Early Adopters

Loraine Lawson

A recent InformationWeek survey on SOA and Web services revealed that one in three companies are disappointed with their results.


Now, some people would look at that stat and think, "Uh-oh. That's not good." But it's important to note that of the 278 professionals surveyed, the majority, or 58 percent, said the projects lived up to their expectations, and another 10 percent said the projects exceeded their companies' expectations. So: 68 percent felt they got what they wanted or better.


The problems really surfaced, though, when Information Week drilled down a bit on whether SOA and Web services - don't you wish surveys wouldn't collapse the two? - achieved the desired business goals. Of those who were disappointed, 27 percent reported that SOA failed to provide the expected level of integration.


I wonder about the integration results, since even early adopters are still in the beginning stages of building services. It makes sense that few companies would report integration benefits at this point. Perhaps it's still too early to see integration benefits.


Even so, overall, the survey results are sobering. The findings suggest companies have overestimated what SOA can do, while underestimating the cost - 30 percent say it costs more than they expected - and complexity - 58 percent reported SOA lead to more complexity into their IT systems. This certainly casts some doubts on the assumption that SOA doesn't require a budget.


The results of the Information Week survey became more bothersome once I read this post by SOA for Profit blogger Martin van den Berg, the Architecture Service Line Manager at Sogeti Nederland B.V. and the author, along with his fellow bloggers - the self-titled SOA Muskateers - of an e-book called, not surprisingly, SOA for Profit.


van den Berg reports he and his colleagues uncovered similar results when they surveyed 12 clients in the Netherlands. Though the survey number is tiny, these clients had an average of three years of experience with SOA - essentially making them SOA veterans. Though he didn't ask specifically about integration, he did ask about reducing complexity. Seven out of the 12 reported that this was one of their main reasons for implementing SOA, but only four in 12 said their SOA projects had decreased complexity.


The other eight reported that SOA had increased complexity.


In the InformationWeek survey, 58 percent of those who said SOA fell short of their expectations cited "introduced more complexity into IT" as the reason.


van den Berg draws two conclusions from the results of the Netherland survey and the InformationWeek survey: 1. SOA isn't going to solve all your IT problems, so set realistic expectations. 2. IT still isn't doing a very good job of measuring progress on new projects.


But the InformationWeek article does offer some good news, as well as more specific clues to successful implementations.


First, 31 percent have already seen a return on their investment, with 44 percent expecting to see ROI within 12 months.


Second, the article shares what UK telecommunications company BT has done to reap benefits from its SOA and Web services implementation. George Glass, the chief architect at BT, reports that BT is nearly 85 percent finished with building its services - which suggests the company is well positioned to speak to the pros and cons of a more mature implementation.


Glass believes more IT divisions will realize a SOA/Web services payoff once they've built out enough services to shut down legacy applications. According to the article, BT shut down 205 legacy systems the first year of its SOA implementation, 710 its second year, and 260 in the first quarter of its third year.


Contrast this with the 36 percent of InformationWeek respondents who reported they did not replace any legacy applications after building services. It notes that 36 percent of the survey respondents did not replace older technology with SOA. What's unclear, unfortunately, is which 36 percent. Is there a correlation between the 32 percent who were disappointed with SOA/Web services and the 36 percent who didn't shut down legacy applications?


The article doesn't say. But if Glass' experience holds true for other companies, there's still hope that SOA will solve many more IT problems than it causes.


And perhaps those satisfaction numbers will go up as more companies learn from the experience of early adopters: Temper your expectations by looking at the long-range plan, create a realistic budget, and identify ways to reduce infrastructure complexities as you build out services.

Add Comment      Leave a comment on this blog post

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.