SOA is useless without good data

Time to pay more attention to enterprise information management

Fellow ZDNet blogger Dana Gardner has just connected the dots between enterprise data management and SOA, and I think this is an important connection that needs to be made. As Dana puts it: "the stakes are much higher for the comprehensive and coordinated attainment of data as SOA becomes common. You can't really get to SOA without the data act together. Having a bunch of services and orchestrating them with some independent logic is fine and dandy. But if the comprehensive data on what the services and processes themselves reflect is not available or timely, well … architecture is a destiny I'd rather not have as a CIO in that instance."

Imagine an SOA, or set of Web services, that you can click on to get at your most up-to-the-minute sales numbers from all locations. What if you can't trust those numbers? What if they're drawn from different systems, are refreshed at different times (and some maybe not at all), and have different shades of meaning at their source locations?  That's a pretty useless service.

It may seem like common sense to get data sources in order, but some in the industry say we're worrying too much about SOA at the expense of enterprise information management, when the two actually need each other. Gartner's David Cearley went so far as to say that SOA is only half of the equation, and that every forward-looking organization needs an enterprise information management strategy to go along with SOA. Adopting such a strategy would increase the chances of success for SOA by at least 70 percent, he said. Cearley noted, however, that companies are still planning their enterprise information management attributes, and that metadata management and XML are still in the early stages of evolution.

I also recently posted my own thoughts on the connection between data and SOA over at Webservices.Org, and this thinking was propelled by IBM's recent announcements in this space.

"Don't put the SOA cart in front of the metadata horse," Dana advises. "If anything, err on the side of the data investment. That way your SOA will readily reap the rewards of your hard work in data management and optimization for maximum process efficiency and knowledge commerce."

Information management is all about getting at the right information at the right time. By being able to sift and sort through data, decision makers can identify and predict new product demand, inventory flows, and spot potential fraud even before it happens. Likewise, SOA is all about getting at the right information at the right time.

No EIM strategy will deliver value without the capability to integrate with disparate silos across the enterprise, or across the globe. Conversely, there will be no value behind SOA without an effective data model that can deliver the right information. The two need each other.