The Pros and Cons of Web-Oriented Architecture Explained


InformationWeek produced an in-depth report on Web-Oriented Architecture this month, which I found late last week via this blog post.


You'll note the enticing title: "A Better Way than SOA." I'll warn you right now that title ended up being a tad misleading. Although the article spends a significant amount of ink on SOA's flaws, once you read the full six-page article, you'll find it actually cautions it's not an either/or proposition:

"The WOA versus SOA debate is unnecessarily contentious. It's obvious that both REST and WS-*-style SOA have their place, so IT groups should stay focused on architectural issues rather than implementation minutia."

It's a really good read for several reasons, and not just because you can download it and read it in the carpool line.


First, it does a great job of explaining WOA and how it differs from SOA:

"SOA and WOA work at different layers of abstraction. SOA is a system-level architectural style that tries to implement new business capabilities so that they can be consumed by many applications. WOA is an interface-level architectural style that focuses on the means by which these service capabilities are exposed to consumers."

Second, it discusses how, when, where and why you could choose WOA over SOA.

Third, it provides examples of real-world companies that have opted for WOA and REST over SOA and SOAP with great results -- including Amazon for its Amazon Web Services.

Finally, it includes a discussion on the shortcomings of WOA and REST, including that REST is not as flexible, and tends to mean you're hard-coding business processes into the infrastructure and using HTTPS for security.

After reading the piece, you may very well decide you'd rather try out WOA and grow into SOA instead of starting with SOA. Essentially, that is the thesis of this article - and it makes an excellent defense of that position, with the help of statistics such as these from an InformationWeek survey of 278 IT pros conducted last year:

  • 32 percent say SOA projects fell short of expectations.
  • 58 percent said their SOA projects introduced more complexity into their IT environment.
  • Just 10 percent said the results exceeded expectations.

Unfortunately, the article doesn't discuss how that decision might affect you long term, as more vendors service-enable their products.


It'd be interesting to see how those statistics might change as SOA implementations mature. As blogger Assaf Arkin humorously pointed out after reading the article, statistics like these don't necessarily make SOA a failure -- just average:

"Considering industry standard, that doesn't look all that doomy or gloomy. It looks ...just average ... From all my sources, it looks like SOA worked out as well as you can expect any good technology to work out. And I'm going to share a secret with you. Keep it private. REST is no silver bullet either. Five years from now, you'll read the exact same article, only this time using the same failure rate and ROI, just blaming a different technology."