IBM advanced its Information On Demand initiative October 16 by introducing an Information Server and supporting services (see "IBM Moves Beyond Data Management"), but are businesses ready for a new level of abstraction? Ambuj Goyal, General Manager, IBM Information Management Software, discusses IT preparedness, SOA imperatives and Big Blue's stake in the BI market.

Are companies prepared for the scope and discipline required in information management projects?

We see two broad classes of customers. One class of customer is saying, "I have a problem in X area, but I want to start small, get the problem solved and then move on to the next internal process that everybody agrees upon." The other class of companies, and I would say they are the minority (maybe 10 percent to 20 percent), say "let's get our arms around the data first."

If you want to start by assessing the state of information in your enterprise, we created an IBM Center of Excellence jointly between IBM Global Services and the IBM Software Group [that can help with everything from] simple, self-assessment tests to long-term consulting projects to help you organize your internal information processes. It can take a long time, but many organizations have been extremely successful by doing that. We also work with companies that say, "let's get started with one project." So, for example, you can start with [the information integration challenges tied to] mergers. Once you're successful, you can deploy it across many other lines of business, and that's when the processes [established in the initial project] start coming into play.

Services-oriented architecture is blurring a lot of boundaries around processes and applications. How does information management fit in?

If you're deploying composite applications, they need composite data from different sources. If the application holds the data, you can't use that same data in another application. The Information Server supports the approach of delivering information as a service so it can be consumed in multiple applications.

For example, when Wachovia creates the master view of the customer across multiple structured and unstructured information sources, that gets fed into 40 or 50 different applications, but they have one single version of the truth about their customers, reconciled across multiple environments. Now let's say they want to change the address for a particular customer; that needs to go into 50 different applications. A single address change used to require 10 calls, but now it's a single call because the Information Server that is being deployed is bidirectional. By changing it once, it then gets updated into every billing system and on every operational system.

Oracle and Microsoft have been ramping up their BI capabilities. Will IBM follow suit?

We are not in what we call the query analysis, reporting and analytics business. Infrastructure needs to feed into BI systems, and there are multiple vendors in the market with many billions of dollars worth of revenue. Customers need to feed reconciled information into those systems in the same way that they need to feed consistent, reliable information into call centers and supply chains. The typical application server or database server doesn't do that. That's where information is stored, but it's not reconciled information from across databases and applications and operating systems, delivered in context to that particular application.