Oracle, IBM Gain Database Market Share on NT

Oracle Corp. and IBM Corp. gained market share in 1998 as Microsoft Corp. lost nearly an eight point share for database sales on Microsoft’s Windows NT operating system, according to preliminary figures.

The numbers represented in a Dataquest ( study will serve as a baseline for judging the impact that Microsoft’s SQL Server 7.0 will have on the market in 1999 and 2000. Version 7.0, which Microsoft positions as a defining release that will put SQL Server on the enterprise map, debuted at Comdex in mid-November 1998.

The study put Oracle’s market share at 46.1 percent, up from 40.7 percent in 1997; Microsoft at 29.7 percent, down from 37.8 percent in 1997; and IBM at 9.7 percent, up from 7.1 percent in 1997. Sybase Inc. dropped from 7.5 percent share in 1997 to 3.0 percent in 1998.

The market as a whole grew 43 percent, from $841 million in 1997 to $1.2 billion in 1998, according to the report released in late March. The growth exceeded forecasts, which had been built on expectations that Y2K concerns would limit new purchases.

Oracle and IBM rushed to provide their perspectives on the report. IBM officials were eager to talk about another finding of the report, that IBM had regained its overall lead in the worldwide database management systems market, which includes mainframes and mid-range systems. Janet Perna, general manager of data management at IBM Software Solutions, also wanted to emphasize the momentum of IBM’s DB2 Universal Database on Windows NT.

IBM’s uptick in market share, coupled with the strong overall revenue growth of the database market on NT, works out to about 100 percent revenue growth for DB2 on NT. "It grew faster than either Oracle or Microsoft’s database on that platform," Perna says. "And I expect it to grow faster than the other vendors on NT in 1999 as well."

By comparison, Oracle’s revenue from NT databases grew about 55 percent and Microsoft’s grew 12 percent in 1998, according to Dataquest’s figures.

Perna credits much of IBM’s growth on NT to the breadth of software IBM offers to support DB2, and the integration of DB2 with IBM e-commerce software. "We now have over 10,000 applications enabled to DB2 UDB on the Unix and NT platforms," Perna says.

Charles Rozwat, senior vice president of Oracle’s server technology division, dismisses IBM’s growth, grouping the company as "down in the noise level" with Sybase and Informix Corp. "We were a little bit surprised by how much market share we picked up from Microsoft," Rozwat says.

Sounding some of the same themes as Perna, Rozwat credits Oracle’s gains on Windows NT in part to e-commerce. Rozwat says customers turned to Oracle for the reliability, scalability and high availability they needed to support e-commerce and data warehousing. Oracle retained its lead in the overall RDBMS market -- a 38.5 percent market share -- and maintained its commanding lead on Unix platforms -- a 60.9 percent market share.

Rozwat refuses to acknowledge that competing with SQL Server 7.0 in 1999 will be any different than competing with SQL Server 6.5 in 1998. "Going forward into ’99, Microsoft released SQL Server 7.0 and we’ve released Oracle8i," Rozwat says. "Given the fact that we believe Oracle8 is technologically at least three years ahead of SQL 7.0, with 8i we’ll even widen our gap."

Analysts seem to have other ideas. In the report, Dataquest analyst Carolyn DiCenzo notes Microsoft’s success in attracting interest in SQL Server 7.0 from ERP vendors such as SAP AG, PeopleSoft and Baan Co. She also mentions that Microsoft beefed up its Application Development Customer Unit from 40 to 600 people to encourage independent software vendors to focus on the platform. "Oracle is still the company to beat. But 1999 should be an interesting year as Microsoft and IBM continue to focus on this platform," DiCenzo writes.

A few months ago, Carl Olofson, an analyst at International Data Corp. (, predicted in a report that SQL Server 7.0 would "eat away at, and possibly eliminate Oracle’s NT RDBMS lead in 1999, unless some significant technical problem arises." Olofson’s conclusion was based on SQL Server 7.0’s ability to address the immediate practical needs of medium-sized shops and on application vendors’ preference for SQL Server for fear of Oracle’s own applications business.