Creating Silos

I came across this Infoworld article about Gartner's predictions for the business intelligence market (aside: I think business intelligence will be a much bigger part of this year's Defrag conference), and this specific bit hit me (emphasis mine):

"So business units will increase spending on packaged analytics, including corporate performance management, predictive analytics, and online marketing BI wares. In so doing, however, they "risk creating silos of applications and information, which will limit cross-function analysis, add complexity, and delay corporate planning and execution of changes," said Nigel Rayner, research vice president at Gartner, in a prepared statement."

There's something very interesting happening in enterprise software. Business units are starting to purchase software (as a service) that uses the web as a platform (obviously). And because of that, the "silo" problem is going to get radically worse - and fast. The problems that result are not just business process oriented, but are also fundamentally architectural in nature.Of course, the larger problem isn't just in the enterprise. The "problem" (opportunity) lies in the web becoming the platform.The IT function inside of enterprises spent an enormous amount of money and time in the late 90s trying to solve the "enterprise application integration" problem. Huge companies were made in the process (Oracle, CA, heck - just about everybody). But just as that problem is at least partially under control, we're now facing the web application integration problem. (sidenote: I'm officially claiming "WAI" as an acronym.)Web application integration is the meta-theme for Glue. It's what brings together cloud computing, APIs, WOA, data integration, identity, context, the open social stack --all of it. It is important that we realize that this trend is being driven not only by the innovation that's occurred around the web as platform over the last few years (call it "web 2.0"), but also by the adoption of software by *business units*.This fundamental shift will remake IT in the enterprise as we know it (over the next 10 years). We can already begin to see that in statements about how business units have "lost faith" in IT's ability to deliver what business units need. And what's the result? Business units are now just providing it for themselves. The architectural implications of that bit alone are enormous. But I'm assuming that IT units are not just going to go quietly into that dark night. CIOs and enterprise architects aren't just going to give up their livelihood. Nor should they. The problems created by business unit adoption (complexity, lack of cross-app functionality, a completely un-orchestrated approach to IT) can ultimately lead us to a place where technology is giving business ZERO productivity gains.So, what's an enterprise architect to do? Begin to understand the shovels and picks and tools that will help "glue" it all together. It doesn't really matter what the software *does* (it could be BI, or expense management, or CRM, or identity management), what matters is how it interacts and intersects with other applications.Getting a handle on it means understanding the nuts and bolts of things like scaling APIs, XMPP, event driven architecture, running apps in the cloud, cross-app context, etc. It means that some of the seemingly "consumer facing" problems and solutions that exist today will be the foundation for enterprise solutions tomorrow.If IT simply allows silos to be created all over again on the web, the problem will be much worse this time around. The web is open, networked, and, in a sense, not tame-able. It calls for the kind of radically distributed architecture that SOA has promised and (largely by no fault of it's own) failed to deliver. It's the big opportunity that Glue is addressing. And I'd love to hear your thoughts about how to address it properly.