The upcoming release, code-named "Viper," is due out in the middle of 2006, according to Bernie Spang, director of database marketing at IBM. Viper will be able to search and analyze structured data in a relational data repository and unstructured data such as multimedia files, Excel spreadsheets and Word documents in an XML repository without the need to reformat it.

That will help save time for developers at organizations moving to service-oriented architectures, in which data tends to be stored in multiple open formats. Viper will support the XQuery language to process XML data, along with standard SQL.

"XML is certainly picking up steam, and there is a lot of growing interest," said Noel Yuhanna, database analyst at Forrester Research Inc. While Oracle Corp. had a "head start" in supporting XML, Yuhanna said he expects that the top three vendors will continue to try to best each other in XML support. He also said he doesn't expect XML to replace SQL as the data format of choice -- despite estimates that the amount of unstructured or XML-formatted data is larger and growing faster than structured relational data in most corporations.

Michael Olson, CEO of open-source XML database vendor Sleepycat Software Inc. in Lincoln, Mass., agreed.

"IBM, Microsoft and Oracle are always going to dominate the data center, as far as I'm concerned," he said. But Olson said embedded XML databases such as his company's Berkeley DB, which run on devices or applications working on the edge of the network, will continue to be speedier and easier to use for such purposes. "Large heavyweight databases, even with XML support, can't compete as effectively in this space.