RealTime IT News
First Fruits of IBM's Xperanto Experiment
By Clint Boulton
February 04, 2003

Rallying on the finding by Dataquest that its business integration efforts represent a $10 billion market opportunity by 2006, IBM Tuesday revealed new data management software tailored to help customers integrate information in multiple locations as if it were stored in one location.

DB2Information Integrator and DB2 Information Integrator for Content were forged in the company's Xperanto research project, a research and development effort focused on tackling rapidly changing data management needs, said Nelson Mattos, director of information integration at IBM Data Management Solutions.

Ideal for customer relationship management (CRM), where call center operators pull together information about customers residing in multiple databases, DB2 Information Integrator software supports two different programming models. The DB2 Information Integrator is tailored to the SQL-based developer community; the DB2 Information Integrator for Content supports a content management programming model. With the products, financial services organizations can tie together in-house customer bank records with investment information from the Internet.

Mattos discussed IBM's federated approach as a key distinguishing characteristic from other data integrators' products.

Mattos told internetnews.com IBM has taken a different path to data management for the database than rivals such as Microsoft and Oracle. While Oracle has been championing the idea of centralizing information in the database so users can pull data from one resource pool, IBM relies on a distributed, or federated, computing approach, where there are several databases managed by a single interface from which to grab information.

With the federated approach, DB2 Information Integrator software reduces the need for organizations to move their data or replace their current IT infrastructure, a key differentiator versus competitors who promote a rip and replace and centralized strategy. For example, Mattos said, a customer can access and integrate relational data in DB2 and Oracle, images in Documentum, e-mail in Lotus Notes, spreadsheets in Microsoft Excel and Web services generated by WebSphere Application Server using just one query. This data is then presented to the customer in a consolidated view.

"While Oracle believes centralization is a relevant approach, there are many situations where centralizing data does not work," Mattos said. "Often, customers are dealing with heterogeneous data, where the value of data changing constantly. Stock quotes are a good example of this. Or areas such as life sciences, which generate gigabytes of data. With the centralized approach, you run out of space pretty soon."

Mattos said IBM applies the access-on-demand approach to obviate this problem, but acknowledged that both centralized and distributed approaches work in certain situations, such as when centralization is used to bring customer relationship management (CRM) data into a warehouse.

Analyst take
Giga Information Group Research Director Philip Russom praised the fruits of Xperanto, but cautioned that the market for enterprise information integration (EII), as it's known, is far too nascent to predict whether IBM's on-demand computing strategy will work with the new DB2 products. Small pure plays in this space include Nimble Technology. MetaMatrix and Enosys, although BEA Systems offers Liquid Data, a comparable integration tool.

Russom said one of Big Blue's key differentiators is the fact that the new software offers allows businesses to access and integrate both structured (text documents) and unstructured (e-mail messages and flat files) information, as if it were stored in one place. Most rivals, he said, only handle structured data.

"IBM's EII platform is much more comprehensive than small-to mid-size vendors," Russom told internetnews.com. "One difference that is in your face is that everyone besides IBM is offering real-time data integration for structured data, but IBM also provides this for unstructured data. Another is that a customer can say 'I just want to use EII for queries.' IBM says 'we can do that'."

Russom said most EII systems are read-only, but IBM's Information Integrator allows customers to write back.

IBM's Mattos, who said research shows that companies spend as much as 40 percent of their annual IT budgets on integration, also talked of the cost-effectiveness of on-demand and EII.

Russom, a bit skeptical, said those aspects remain to be seen, but that it is conceivable that customers could enjoy cost savings.

"IBM's on-demand [play] requires a lot of IT," Russom explained. "The thing about EII is that it is so very new. "Very few have implemented it and there is very little data about what it costs. Until more companies implement it and we see what the payroll costs are, we don't know. However, it is conceivable that, compared to older approaches to competitive analysis, such as data warehousing and data processing, EII could easily similar results for less. I have pretty good confidence in EII."

As for competition, Russom said no one can compete with IBM right now in EII.

"IBM is very focused on Oracle and Microsoft," he said. "Any company that's a vendor with a mature database management system has some of pieces of EII, but neither Oracle nor Microsoft have put it together yet."