« RFID research reports from ABI Research   |   Main   |   Q3 Sales of Smart Phones, PDAs show strong growth »

Was Gartner’s endorsement of SOA at Symposium newsworthy?

At the opening keynote at Gartner’s Symposium conference, I was struck by the opening panel’s announcment that SOA was the most important trend. Now, I’m not minimizing it - but the delivery of this was basically that the speaker (I think Gartner Distinguished Analyst, Gene Phifer) basically looked at the audience and pronounced “SOA” like it was new information and all the attenees - largely CIOs - needed to take away from the conference. Here’s “>the research piece Gartner produced to go along with their keynote theme.

I’m just surprised that they pronounced it as if it were unheard of in the industry, instead of the lively topic of hype that it already was (see such indications of preSymposium indicators of SOA’s importance as this Survey: SOA prominent on 2005 budgets) Still, the Gartner endorsement has let loose a more serious consideration of what SOA means by the tech media. Two pieces I found interesting are these from Infoworld - Coming to Terms with Service-Oriented Architectures - a more general, big picture piece - and Coming to Terms with Components - which gets more into the details of businsess models and implementations.

Sure, SOA is important, but it’s a concept, a framework not all that different than the vision behind CORBA back in the ’90s (and how successful was that? what about EJB?). The emergence of SODA (services oriented development of applications) as a methodology takes us one step further toward realizing this goal by giving us a path to follow. Make no mistake - implementing SOA involves rearchitecting applications, data centers, networks and systems. It can be implemented incrementally, as enterprises roll out new applications - but this fundamentally conflicts with the kind of projects enterprises are taking on. They’re largely incremental improvements and staged roll outs. The current climate in enterprise IT spending doesn’t point to a huge uptake in undertaking projects to implement SOA.

There are significant benefits to be gained from adoption of SOA and component-based web services. It reduces integration costs for future applications and for changes to the system - essentially increasing agility and adaptability. But we can’t forget that it involves building a fundamentally different infrastructure.

Now vendors are talking about this a lot. Package vendors are retooling their applications to include web services and support SOA - but that’s because SOA helps mitigate one of their harder-to-avoid sales objections - the sheer cost and disruption of upgrades and new installs. But we have to remember that what vendors want to sell and what customers want to buy can be quite different. Especially with lagging IT spending and a widely held sense that there is no ‘next big thing’ out there to drive spending, vendors are eager to hop on anything that will strengthen their selling propositon. And SOA - as a “revolutionary” approach to software development, integration and delivery - implies a multi-year investment cycle, similar in strength and scope to those seen with the migration to client/server and the web.