A Spoonful of Arbor Helps OLAP Medicine Go Down

Despite the promise of unlocking data stores and making valuable data more easily available to corporate end users, OLAP thus far hasn’t had a broad-based impact on the industry. With the development of Microsoft Corp.’s "lite" OLAP engine, Plato, slated to be included in version 7.0 of SQL Server, many analysts believe that OLAP is ready to make that impact. OLAP pioneer Arbor Software Corp. (Sunnyvale, Calif.) hopes to get a start on a possible OLAP revolution by releasing the Arbor Integration Server, an OLAP server tool that can automatically create new OLAP applications, as well as transparently modify existing ones.

"The OLAP marketplace is an interesting kind of market segment right now," observes Robert Craig, director of data warehousing and the business intelligence practice at the Hurwitz Group (Framingham, Mass., www.hurwitz.com). "My perception is that Microsoft is going to have a big impact on this marketplace when Plato becomes available later this year. The key strategy is that Microsoft is going to bring OLAP to the masses, so we’ll see how that works out."

Dan Druker, director of product marketing for Arbor, agrees with Craig but notes that Microsoft’s Plato engine will be closely tied to SQL Server, and not to other relational database platforms, thereby creating an opportunity for a vendor who can access relational stores from vendors such as Oracle Corp. and IBM Corp. According to Druker, the OLAP capabilities of Plato very likely could serve to whet the appetite of many a database administrator (DBA), and this could create a demand for more robust, higher performance products. "What Plato’s going to do is create the demand for OLAP -- but on steroids," Druker maintains. "Most of my customers are using Oracle and IBM DB2 and Sybase; the data’s just not in SQL Server yet."

The Arbor Integration Server is a suite of graphical tools and data integration services that Arbor’s Druker says can reduce the time and expense that an enterprise customarily takes to create, deploy and manage analytic applications.

The Integration Server comprises OLAP Architect, a graphical tool that can map relational tables into logical dimensional schemes; OLAP Builder, a graphical tool that lets users assemble the reusable structures created by the OLAP Architect into analytic applications; and OLAP Integration Services, a multithreaded server application that automatically creates analytic applications from data and metadata stored in relational databases.

What it all adds up to, says Arbor’s Druker, is enabling enterprises to fully exploit an investment in a data warehouse. "The integration server is directly designed to address the issues of an organization that has data warehouses and wants to be able to more fully leverage them," says Druker. To this end, the Arbor Integration Server defines a relational store that functions as a central catalog to which all or a subset of all the data in a data warehouse or data warehouses can be mapped.

"We can look at the data warehouse and define, for example, 15 dimensions across the data warehouse," Druker explains. "I can create common structures and store these in a big catalog, so I get centralized definitions that I can use later to build analytic apps."

Druker acknowledges that a "data management guru" is initially required to set up the central metadata store. However, the Arbor Integration Server features built-in intelligence, which can create custom analytic applications using dimensional information stored in the metadata repository and can automatically remap dimensions and applications as dimensions change. Thus, DBAs with little data warehousing experience can oversee the Integration Server’s daily operation.

Hurwitz’s Craig sees this functionality as a big plus for Arbor. "It simplifies the whole process of building OLAP cubes for Essbase customers," Craig observes. "It’s easier to build new cubes, because you can just go against one of these pre-existing dimensional cubes and use it."