Can Web Pixie Dust Save SOA?


Have you ever noticed how magical the Web is?


I mean, sure, it's technology, not magic. But it's almost as if Web-enabling something gives it a new life, a new depth that just makes applications more dynamic, more fun, more ... magic.


Maybe that's why WOA Web-oriented architecture is generating such a buzz, even though, at its technological heart, it's basically Web-enabling SOA.


Back in September, when I interviewed Gartner vice president Nick Gall, he offered a simple formula for describing WOA:


WOA = SOA + REST + WWW Gall first coined the term back in 2005, so in a way, he's entitled to define it although, to be fair, Dion Hinchcliffe has contributed substantially to the effort.


When I spoke with Gall, he mentioned Gartner was working on a research note about WOA. It published this week under the title, "Tutorial: Web-Oriented Architecture: Putting the Web Back in Web Services."


If you're rolling your eyes, you're probably not alone. The report's co-author, Anthony Bradley, even blogged about the report under the insightful, if somewhat sarcastic, title, "I Just Learned SOA and Now I Have to Learn WOA"?"


I can see how some would think WOA is a shell game. Gartner defines "Web-oriented architecture" as sub-style of service-oriented architecture, based on the Web. And just like it's SOA sire, WOA suffers from a lot of the same problems: An imprecise definition, vague statements about "architectural style" instead of specifications or standards, and few actual real world deployments. Some say it really just boils down to choosing REST over SOAP.


Dion Hinchcliffe posted a very thorough explanation of the functional difference between SOA and WOA earlier this year. But what I haven't seen is a simple, straightforward and short take on why businesses should care about WOA, particularly given the disillusionment with SOA.


It looks like this might be one of the issues the Gartner report tries to pin down, given that the report's summary includes this:

"WOA's goal is to transform traditional application-to-application integration from a 'rat's nest' of specialized interfaces into a generic web of globally linked hypermedia."
That's about the best explanation I've seen on why businesses should care about WOA. To simplify even further, I would offer this summary, taken from my readings and interviews on WOA: WOA offers the same benefits of SOA, plus all the pros of the Internet ---URI-identified endpoints, HTTPS and, perhaps most importantly, mashup capabilities.


You'll notice that what's missing from that summary are the downsides of WOA. I am certainly not trying to pretend there aren't downsides although at this point, the cons are pretty fuzzy. I'm just trying to give you an idea of the business case for WOA and why it's attracting everyone's attention right now.


Obviously, Gall isn't going to republish a $495 research note in his blog, but he does post Gartner's five fundamental, generic interface constraints of WOA:

  1. Identification of resources
  2. Manipulation of resources through representations
  3. Self-descriptive messages
  4. Hypermedia as the engine of application state
  5. Application neutrality
The post elaborates more on the meaning of "application neutrality."


Those are technical constraints, as they should be. After all, WOA, like SOA, will be built with technology and planning, not magic.


Still, a little web pixie dust might be just the thing to fix the disillusionment with service-oriented architecture Web or otherwise.