Aw, Phil, say it ain't so... Fellow ZDnet blogger Phil Wainewright says that SOA is passe, observing that anyone who's anyone is piling on SOA these days, from Tim Bray, to John Hagel and Vinnie Mirchandani. Even Gartner is bailing -- analyst Daryl Plummer opened up the discussion a couple months back in a commentary in Optimize that questioned the SOA paradigm in enterprises.
Phil notes that "when everyone is lining up to trash anything as comprehensively as this, you know it must be reaching some sort of bottom. We haven't yet reached the point where vendors start to renounce their commitment to SOA, but it can only be a matter of time. This is what Gartner terms the 'trough of disillusionment' in its technology hype cycle. If SOA were a quoted equity, it would now be at least 90% down from its peak valuation. Everyone is bailing out." Maybe reaching the trough is actually good news for SOA, since the next stage is advancing up the slope of enlightenment to the plateau of productivity
IMHO, it doesn't seem like the hype is anywhere close to dying down, as a current Google news search of "SOA" will attest. But, if Phil's perceptions are correct -- and Phil knows this stuff very well -- this may be one of those cases where a technology/philosophy is sunk before it's even tried. There are actually very few actual SOA implementations out there, unlike other technologies that have spent time wallowing in the trough.
I have not heard many horror stories of millions being drained on SOA-based projects that aren't delivering. So far, the stories relate to efficiencies seen in one or more areas of the enterprise. Not necessarily massive savings -- or massive failures for that matter -- because massive SOA projects simply aren't there yet in mainstream enterprises.
This being said, maybe reaching the trough is actually good news for SOA advocates. As Gartner observes with its hype cycle model, the "trough of disillusionment" is followed by an upward "slope of enlightenment," followed by a "plateau of productivity." (Climbing a slope of enlightenment sounds fun, but what could more charming than life on the plateau of productivity? Are we talking about Walton's Mountain here, or what?)
In the post-hype era, after the trade press and me-too vendors milk it and move on to something else, enterprise planners, managers, developers, and administrators are left to roll up their sleeves and figure out how to make it all work. This has certainly been the case with post-hype e-commerce, CRM, the application service provider model, and even many ERP implementations, which are humming along fine and profitably for their owners.
Phil says as much: "As inevitably as SOA has fallen from grace, sooner or later it will bounce back," but perhaps under a different name. I agree. The movement to standardization and abstraction is an unstoppable force. We're not only decomposing and componentizing applications, but businesses as well. We've come too far to turn back. It's just a question of whether we'll continue to use the unfortunately overworked term 'SOA' or call it something else.
Perhaps the ultimate measure of success of a technology or philosophy is when no one talks about it anymore, when it becomes a ubiquitous part of day-to-day business. In a recent TalkBack post to this Weblog, a reader, parasubvert, drew an interesting analogy between perceptions around SOA and object orientation:
"OO was just as intangible, over-hyped, and had quite a few contradictory definitions," the reader observed. "There are plenty of people today that would claim OO has 'failed,' even though the majority of software is written that way today, and most non-OO languages often wind up adding OO-like features to them after peer pressure. So, even if we don't solve the world's IT challenges, if loose coupling, autonomy, explicit contract governance, and canonical models all become built-in and 'normal' to the next generation of IT infrastructure, won't that really be what we're looking for? To raise the waterline?"
So, on to the plateau of productivity!