DB2 Information Integrator, a family of data integration products that IBM shipped in beta last week, holds the promise of saving enterprises time and money by averting the need for deploying application integration or data warehousing projects.
DB2 Information Integrator Version 8.1 and DB2 Information Integrator for Content, the two members of the family, are the first fruits of the Armonk, N.Y., companys years-long Xperanto initiative.
The software encapsulates IBMs goal to help customers access, integrate and analyze all forms of data within and beyond the enterprise. DB2 Information Integrator 8.1 is tailored for SQL-based, or structured, data, while DB2 Information Integrator for Content is geared primarily for unstructured data.
IBM pledges that the new software, due to be generally available by July, will help developers slash the amount of laborious hand coding required when integrating two or more relational databases. This is accomplished by creating a metadata index that indicates where data is stored. The so-called federated approach to data management leaves data in its native applications and allows users to query it as if it were in one place.
Information Integrator beta customers, many of whom work in the biosciences and are steeped in the world of structured data, are giving the release a thumbs up. Most point to good performance; the cost, labor and time savings reaped from avoiding enterprise application integration and data warehousing; and the time savings bestowed by eliminating previously clunky ways of comparing data from diverse sources.
Craig Stewart, director of research and academic computing at Indiana University, in Indianapolis, said the federated approach is the "only tenable approach" to life sciences today.
"Particularly in biomed and genomics, you see tremendous distribution of data sources worldwide," Stewart said. "Theres just no way to try to create a centralized warehouse and keep it up-to-date and manage it effectively."
But users of rival databases arent convinced. Mark West, CIO of Electronic Arts Inc., in Redwood City, Calif., and a user of Oracle Corp.s namesake database, said the federated approach is "great in concept but difficult in practice.
"The promise is theyll bring together all this nonuniform data in a uniform way," West said. "In reality, its difficult to achieve because you dont know how [a given] spreadsheet has been changed or if a new version was created. ... Thats where federated falls down."
Although Information Integrator supports accessing and integrating XML documents as data sources and generates XML documents as query results, it doesnt support XQuery, the XML query language. Officials said such support will be added when XQuery becomes a standard, likely next year.
According to Nelson Mattos, IBMs director of information integration, XQuery didnt appear in the first Xperanto release because IBMs goal is to simplify the work of application developers. "That means we need to provide an information integration environment that is tailored to their skills," said Mattos, in San Jose, Calif. "There are very little skills available for XQuery—Id say almost none."