Repository Plays Central Role in Microsoft Data Warehousing Strategy
BOSTON -- At the Microsoft Focus on Windows NT conference held on June 24 and 25 in Boston, Steven Murchie, Microsoft Corp.’s group product manager for SQL Server, focused his keynote address, "The Impact of Windows NT on Data Warehousing," on the part Microsoft Repository will play in Microsoft’s data warehousing strategy.
For now, Microsoft Repository is best known as a tool for storing and managing components for application development. It ships primarily with the Visual Studio application development suite. But, according to Murchie, Repository is being groomed for a central role in the data warehouses that will be built using Microsoft SQL Server 7.0 and the company's forthcoming OLAP engine, Plato. In that context, Repository will be used to store and manage metadata.
Microsoft Repository and its data warehousing extensions will enable broader access to data warehouses by facilitating the sharing of metadata and by making it easier for data warehouse users to obtain information about the data in the warehouse, Murchie says.
Two kinds of metadata are relevant for data warehouses, Murchie says: technical metadata, which includes information about data sources, targets, transformations and mappings; and business metadata, which describes the data’s lineage, the queries, reports or charts that include the data, and any other references made to the data. For example, "users can ask the Repository what it knows about sales data," Murchie explains.
Microsoft Repository will also be used to manage SQL Server 7.0’s Data Transformation Services (DTS), a tool for transforming data destined to populate data warehouses, Murchie says.
Microsoft has been working to extend Repository to enable it to perform its assigned task within the data warehousing framework since December, when it announced a design review process for gathering industry feedback on the extensions. Since then, Microsoft has held several conferences with ISVs and others to solicit feedback on the extensions.
The Meta Data Coalition (www.mdc.com), a group of vendors focused on the interoperability of metadata, has released its own standard, the Metadata Interchange Specification (MDIS). However, it has also made available an MDIS-to-Microsoft Repository translator. The Metadata Coalition has also been involved in Microsoft’s design review process.
While Murchie pointed out that recent figures from the Meta Group (Stamford, Conn.) show that more than 50 percent of new data warehousing/data mart implementations are running on Windows NT, he added that, in the data warehousing space, efforts such as the design review and the Metadata Coalition’s participation are necessary because "it’s always a combination of multiple vendors that need to come together to make things work."