Have we been making a mountain out of an SOA molehill?
I've been having an interesting dialog (with follow-up here) over at the Enterprise 2.0 site with Bob Serr, CTO of Parlano, who questions the "disruptive" claims being made about SOA. At best, he stated in a recent blog post, SOA is a great methodology for changing or streamlining the way things are done, but is not yet “earth shattering.”Web 2.0 services may or may not have an SOA foundation -- but whether they do or not really doesn't matter.
SOA simply has not delivered any of the new or revolutionary changes the industry hype would want us to believe, Bob points out. Plus, there is no substantial connection yet between SOA and Web 2.0, a paradigm that is a bit closer to revolutionary.
"To me SOA is an evolution of concepts that have been around for decades. Software developers have long been figuring out ways to componentize what they build so that they can utilize these components in different ways. The fact that we now access these components through HTTP and Web Services is great, and makes programming much easier and more extensible. But is this disruptive over something like the CORBA or RMI or HTTP programs we wrote years ago? Back then we could write distributed components that programs from multiple languages could access, and these services could be published into directories. Now we have programs based on Web services, and these programs can be published using Web Services Description Languages (WSDL) in various services directories. And we have applications that allow semi-fluent technologists to easily incorporate logic from these services into programs. But is this disruptive? Does this change markets or create new ones? I don't think so. At least not yet."
Bob also makes the point that there are potentially disruptive Web 2.0 venues on the rise out there, and they may be built on SOAs, but this is irrelevant. "Sure, it is possible that Linked-In is built on a Service-Oriented Architecture. But it might not be. For all we know it might be a simple CGI application that is monolithic and has all of its logic written in one place. And what about Wikipedia or any of the new Office 2.0 applications that you can run from companies such as Google? Are these built using SOA? Again, maybe they are, maybe they are not. But does it matter? End-users don't care."
As a response, I pointed to a post I made on this site that examines areas in which SOA may cause some disruption, including paving the way to "loosely coupled businesses" built on bundles of services, greater opportunities for microbusinesses that can offer pieces of applications or SaaS through services provided to the infrastructures of larger companies, and the fact that Web 2.0 has many of the same goals as SOA outside the corporate firewall.
Bob, however, cites ZDNet blogging colleague Dion Hinchcliffe's own calculations about the actual number of APIs and mashups currently in existence, which don't suggest the revolution is at hand anytime soon. Dion had stated in a recent post, that as of December 13, 2006 there were 348 APIs registered and 1,350 mashups. Bob observes that "while these numbers are more impressive than say 0, they are nothing compared with the number of Websites that were created in the previous revolution, which was the World Wide Web."