Database giants battle with XML

With Oracle's annual OpenWorld conference looming, database vendors are preparing for battle once again.

With Oracle's annual OpenWorld conference looming, database vendors are preparing for battle once again.

This time around, the big three - IBM, Oracle and Microsoft - are brandishing Extensible Markup Language (XML) as the weapon for making their databases faster as an anchor for Web services.

Microsoft is readying its charge into the enterprise-class arena with the forthcoming version of SQL Server, code-named Yukon. Yukon is being hailed as an XML-savvy, back-end engine for Microsoft's .Net Web services initiative.

The second, and perhaps more important, design goal for Yukon is language independence, according to Barry Goffe, group product manager of the .Net enterprise server group. "We've had this vision for a long time: to have a multilanguage database," Goffe said.

To create that multilanguage database, Microsoft is arming Yukon to host XML natively in the database and is making XML a definable column type, which enables XML data to more effectively be searched and retrieved, said Stan Sorenson, director of server marketing at Microsoft.

Although the delivery date for Microsoft's next generation of SQL Server has been a closely guarded secret, Sorenson said, "We expect [Yukon] to ship in the first half of the calendar year 2003," adding that Yukon will go into beta in the second quarter of next year.

Microsoft has been adding support for emerging XML standards in a series of Web releases, the latest of which, SQLXML 2.0, came in October. SQLXML 2.0 contains support for XML Schema Definition (XSD), a specification designed to ease data integration from the World Wide Web Consortium (W3C).

Oracle and IBM, meanwhile, are also sharpening their XML battle-axes. Jeremy Burton, vice-president of worldwide marketing at Oracle, said that Oracle would offer an XML database support) at the company's annual Oracle OpenWorld conference in late November.

Not to be outdone, IBM officials say they have a head start on Oracle and Microsoft, noting that it has already delivered key database-related technology.

IBM is stressing that the combination of DB2 and its XML Extender provides the functional equivalent of Oracle's XDB technology, said Jeff Jones, director of strategy for IBM data-management solutions at IBM's labs.

"Our XDB is DB2 plus the XML Extender. We have not packaged that and called it something new. We have extended DB2 to manage XML content by creating an XML Extender and an XML data type and by integrating them all very closely with the engine," Jones said.

For the moment, Microsoft, IBM and Oracle are all in a race to implant XQuery - an XML-based querying standard in their core databases - for accessing very different data sources, most notably those based in either XML or SQL.

"We are driving XQuery support into DB2 so we will be able to think of DB2 as bilingual," Jones said. "It would allow for receiving XML-or SQL-based requests and give you back data results in either one." IBM has a prototype product with XQuery support integrated and working in its labs.

The big three are not the only ones supporting XML. Relational database vendor Sybase is riding the XML train, as are pure-play XML database providers such as Software AG and Ixiasoft.

Some users are already starting to reap the rewards of XML. Analysts have said that in addition to improving performance, supporting XML in the database gives software the capability of integrating with other applications and systems. "The X in XML stands for extensible, but it might as well stand for exchange, because that's how people use it," said Philip Russom, an independent industry analyst.

But XML is not an easy fit with relational data. "One hitch is that XML data is hierarchical, which makes it akin to a round peg in the square holes of relational databases," Russom said.

In a push towards a language-independent database, Microsoft's Yukon will work with the Common Language Runtime, the company's technology for multilanguage support across more than 20 non-Microsoft programming languages. The database will be tightly integrated with Microsoft's Visual Studio.Net toolbox for developers.

Oracle and IBM, not surprisingly, have also been working to equip their databases for unstructured data - Oracle with its Internet File System (iFS) and IBM through its extenders, content management software.

The ultimate goal of combining structured and unstructured data sets is to provide a view of all the information that a company has on any given topic. This includes aggregating all the discussions, notes, meetings and information that resides in office applications, data warehouses and enterprise applications such as customer relationship management and enterprise resource planning, said Microsoft's Goffe.

Read more on Business applications