It took more than a decade for SQL Server to gain across-the-board credibility in the enterprise relational database management system (RDBMS) market. It's taken much less time for SQL Server to achieve a leading position in the OLAP space. The OLAP Report, an independent resource for OLAP technical information, recently crowned SQL Server 2000 Analysis Services the market leader in the highly competitive business intelligence (BI) space. You can read the entire report at http://www.olapreport.com/Market.htm , but here are two of the report's most interesting findings—as well as a prediction of my own.
First, according to The OLAP Report, Microsoft claimed 24.4 percent of the OLAP market in 2002, with Hyperion grabbing 22.9 percent of the market, and Cognos coming in at 12.6 percent. No other vendor had more than a 10 percent market share. The top two vendors accounted for almost 58 percent of the market, leaving a surprisingly large number of vendors fighting it out for the remaining 42 percent.
Second, Microsoft was the only major RDBMS vendor to have a significant piece of the OLAP market. Oracle had 4.7 percent, whereas IBM weighed in with 2.2 percent. And Oracle and IBM both lost market share, whereas Microsoft gained a few percentage points. "Microsoft's SQL Server Analysis Services grew faster than the market and therefore increased its market share again," noted Nigel Pendse, lead author of The OLAP Report's 2002 market-share report. "As in previous years, usage of Analysis Services grew faster than SQL Server itself, as Oracle and DB2 sites continue to select Analysis Services to build BI solutions on top of their existing relational databases."
And here's my prediction: Long term, BI will have huge ramifications for the competitive position of the major database vendors. The ultimate purpose of BI—whether you're using OLAP, data mining, or an old-fashioned spreadsheet—is to help users and companies make better decisions. So, it stands to reason that almost all companies should be using data analytics (or BI) in some capacity.
A somewhat artificial distinction exists today between the online transaction processing (OLTP) and OLAP worlds. When you get down to it, data is data, and most OLAP solutions have their roots in the OLTP world. After all, someone has to collect the transactions and raw data in the first place. And eventually, people use some sort of extraction, transformation, and loading (ETL) process to put that transactional data into a data warehouse, where other people analyze it and split it into smaller data sets. Because of technical limitations, doing complex analytics directly on live transactional databases simply isn't practical today. But I believe that software and hardware will eventually improve to the point where OLTP and OLAP data stores start to blur. When that happens, why bother with a complicated ETL solution if you can effectively build and query your data marts without it?
Industry reports tell us that the OLAP market is growing steadily—and typically much faster than the RDBMS market. And according to The OLAP Report, Microsoft is the only major RDBMS vendor that's successfully capturing OLAP market share. As the lines between OLAP and OLTP continue to blur, I predict that Microsoft will reap huge competitive advantages over its traditional OLTP competitors, who aren't taking the OLAP space seriously.
