Whose victory?

History buffs recognize Shiloh as the name of the Civil War's second major battle. For Microsoft BackOffice and SQL Server users, however, Shiloh means the next major release of Microsoft SQL Server. Microsoft knows it has to offer a better story on the reliability and scalability fronts, where Oracle and IBM focus when they compare their products to SQL Server 7.0. But Microsoft can claim terabyte-sized data warehouses and can boast 99.986 percent availability (during a nine-month period beginning late in June 1998) for the well-known TerraServer site, which sells satellite images.

Clustering is one area in which Oracle and IBM both are eating SQL Server's lunch. SQL Server Shiloh will support more than current two-node clustering. Bigger clusters will help SQL Server compete in both scalability, with more than two nodes, and reliability, with more sophisticated failover. One disadvantage of SQL Server being so tightly integrated with Windows NT is reliance on NT for clustering support. But an advantage is that SQL Server's log shipping (automated warm backup) will benefit from enhanced cluster support because you'll be able to configure log shipping simultaneously from several nodes. (For more information on clustering, see Brian Moran and David Sapery, "SQL Server Clustering," June 1999.)

Microsoft also has been making noise about data mining, another area in which both Oracle (with its ConText and customer relationship management applications, and its recent acquisition of high-end data-mining vendor Thinking Machines) and IBM (with its Intelligent Miners for Text, Data, and Relationship Management) shine. Data mining is still bleeding-edge technology with far too many small players hoping to emerge as market leaders. True data mining—not simply advanced querying—is compute-intensive, so assume that Shiloh will have improved support for parallelism.

Because data mining is viewed as part of business intelligence (BI) in general, support for data mining will expand SQL Server's competitive position in the BI market. OLAP Services, of course, represents the foundation, but data mining and better natural- language support will make SQL Server an even more potent competitor. For example, Shiloh will ship with a more powerful, easier-to-use English Query tool. (See Web Dev, "The Amazing English Query Tool," April 1999, for more information about English Query.) English Query 7.0, which shipped with SQL Server 7.0, currently provides support for OLAP Services by generating MDX expressions to query OLAP cubes.

Shiloh-generation SQL Ser-ver also will include (finally!) a thin version of SQL Server for use with Windows CE devices. Sybase (SQL Anywhere UltraLite) and Oracle (Oracle Lite) have been shipping lite versions of their products for some time. Even IBM has announced that DB2 UDB 6.0 (Universal Database) will ship with a thin version of UDB called IBM DB2 Everywhere sometime this year.

So who will win the second Battle of Shiloh? I suspect the SQL Server community will.