Other enhancements will come in the form of technologies that can exploit unstructured data and help organizations deal with the explosion of data from Web 2.0, the CTO said.
"For unstructured info, we need very high-quality information extraction capabilities to be able to determine names, addresses, phone numbers ... and to be able to determine the sentiments, the causalities, etc.," Jhingran said. "IBM scientists are working on these hard problems and we have made great strides to be able to bring this extracted information back into the world of databases where the behavior is slightly different and the reasoning [and] analysis needs to be too."
As an example, he continued, "a sale in a quarter is a sale in a quarter. But is this document about IBM, about databases or about the computing industry? Probably all three-so how do we account for uncertainty in the database analysis? That is a major question that our scientists are answering."
Peter O'Kelly, an analyst with the Burton Group, said the DBMS market is no longer a strictly relational DBMS market. He noted that the top vendors-IBM, Microsoft and Oracle-have moved to multidata model DBMSes with support for the extended relational data model along with specialized storage subsystems for XML content, file streams and analytical data domains, and while some thought leaders in the DBMS market have suggested that the "one size fits all" DBMS era is ending, the latest DBMSes from those vendors are central to enterprise application infrastructure partly because those vendors have seamlessly extended their DBMSes to accommodate nontraditional database requirements.
"I believe the single most important contribution DB2 made to the DBMS market was in establishing the credibility and effectiveness of relational DBMSes for real-world enterprise data management scenarios," O'Kelly said. "IBM, Informix, Ingres, Oracle, and Sybase all played an intense game of RDBMS feature/function leapfrog during the 1980s and 1990s, and each vendor contributed to overall RDBMS innovation, but I believe DB2's biggest influence was in irrefutably establishing the fact that RDBMSes were ready for enterprise prime time.
"This is all taken for granted today," he added, "but it was a radical and hugely controversial proposition 25 years ago, when enterprises were still focused on earlier-primarily hierarchical and network-DBMS architectures."