Microsoft announces a plan for making data warehouses easier to build, manage, and standardize.
Today at the Networld+Interop trade show in Atlanta, Microsoft announced the Alliance for Data Warehousing, together with partners including application vendors SAP and Baan, data warehousing software makers Platinum Technology, Praxis International, and Informatica, query tool vendors Business Objects and Pilot Software, hardware vendor NCR, and systems integrator ExecuSoft Systems will present a unified model for building midsized data warehouses based on Microsoft's SQL Server database and Windows NT operating system.
While database makers, including Oracle, Sybase, and Informix have generally aimed their data warehousing products at large corporations for building huge, multiterabyte data stores of corporate information, Microsoft will target its program at smaller companies building so-called data marts. Data marts typically are much smaller than data warehouses, and may include a fraction of the data stored in larger systems, and are targeted at users in small businesses or in divisions of large companies.
Microsoft, and its partners, is also formulating an ActiveX-based framework, called the Active Data Warehousing Framework, that will make information stored in a data warehouses based on the company's SQL Server database easily accessible and manageable via tools from multiple vendors.
IBM--which established the market for data warehousing more than six years ago--is also rolling out its own low-end, workgroup-based data warehousing plan at Networld+Interop, according to the company.
But Microsoft's scheme comes with industry backing, and analysts say, the time is right for Microsoft's plan. "This is a ripe opportunity for Microsoft to come in and put its stamp on the market," said Dan Kusnetzky, an analyst with International Data Corporporation. "There is a huge opportunity for a vendor to stand up and say, 'We are building data warehouses for the rest of us.'"
The proposed framework will define a standard API to allow access to metadata--the information that describes the type and location of data stored in a warehouse--so that tools from multiple vendors will be able to obtain the same data stored in SQL Server. No metadata standard currently exists, which means that each vendor's tool must define a metadata layer for an underlying data warehouse before it can get the data. That's a time consuming process that can slow down data warehouse implementation.
The framework will also define interfaces for other common procedures for assembling data warehouses, including data acquisition, transformation of data into a single format, data distribution, and replication. Microsoft said the next version of SQL Server, code-named Sphinx, and due to ship next year, will include several data warehouse-specific enhancements, including parallel query processing, advanced data indexing, and data partitioning. Also due to be added is the ability to replicate data via the Internet, said Microsoft.
A new Excel add-in will be released later this year allowing users to create reports based in data contained in SQL Server. The add-in will be available free of charge from Microsoft's Web site.
Microsoft is also working on ActiveX technology for getting to data sources through Web browsers. The company will use its OLE DB data access API, along with middleware--obtained through the acquisition of NetWise last fall--to build the technology.
The combination of well-packaged technology and backing from established players could give Microsoft an opportunity to grab a big share of the data warehousing market, said Kusnetzky. "There are people looking at the low end and the high end of the data warehousing market; but nobody is looking at the whole scale. Microsoft, being a leader in the desktop community, could easily put together a partnership to build a warehousing ecosystem to do that."
What's unclear is whether the company's proposed standards will win any support from database competitors?whose databases manage the vast majority of corporate data. Analysts see IBM, Oracle, Sybase, Informix, and others resisting a proposed standard based on Microsoft ActiveX technology.