What does Microsoft need a research group for? The short answer is to probe the leading edge of computer science to develop new insights for the industry and innovative products for consumers.
In 1991, Microsoft became one of the first major software companies to start a research group when it hired Rick Rashid from Carnegie Mellon University to start Microsoft Research. The group was tasked with supporting computer science research without regard to the timelines for any particular products. The purpose for this approach was to develop new technologies that future software generations could build upon.
Today there are six Microsoft Research labs in the following locations: Redmond; Silicon Valley; Cambridge, Massachusetts; Cambridge, England; Bangalore; and Beijing. I first heard about Microsoft Research in 1997 when Jim Gray demonstrated the TerraServer application in New York City at Microsoft’s “Scalability Day” marketing event to herald the soon-to-be-released SQL Server 7.0. The TerraServer project was envisioned and commissioned in 1996 by Paul Flessner, who was the general manager of the SQL Server group at the time. The SQL Server team was redesigning the SQL Server product from the ground up because it had basically reached a scalability wall—the current architecture just couldn’t handle more users and more data. At the time, businesses and consumers were really starting to make use of the Internet. The mission of the TerraServer project was to build an interactive, Web-based, database-driven application to showcase the scalability of a single database server running Windows NT 4.0 and SQL Server 7.0. The goal was to have at least a terabyte of data available. The TerraServer drove home the power of data on the Internet more than anything I had seen up to that time. Everyone I showed it to immediately understood how to use it, what its value was, and how much data needed to be available and easily accessible.
In the fall of 2000, Jim and Tom Barclay (the lead developer and project manager for the TerraServer project) met with SQL Server MVPs at the Professional Association of SQL Server (PASS) Conference in San Francisco and gave us an inside look at how the TerraServer stored and accessed the massive data volumes it was working with. They also told us about some of the other incredible contributions that Microsoft Research has made to Microsoft products and the computer industry as a whole. Technologies developed by Microsoft Research have led to more efficient search algorithms, phonetic audio searching, answering machine detection, and strategies for enhancing gaming environments (for Halo 3 in particular). To see a list of some of the other amazing technologies contributed by Microsoft Research, go to
http://research.microsoft.com/aboutmsr/pastpresentfuture/contributions.aspx.
Microsoft Research’s next facility is going to be an advanced development laboratory in downtown Madison, Wisconsin, which is slated to open this fall. The lab will be named the Microsoft Jim Gray Systems Lab after Jim, who was the first director of the Bay Area Research Center (BARC) in San Francisco. He disappeared while sailing outside of San Francisco Bay in January 2007. Jim was a true scientist. He loved doing research, and when that research led to exciting new products that was just icing on the cake. Products were a side benefit, not the goal, of Jim’s work. You can read about many of his research projects at http://research.microsoft.com/~gray/.
I'm pleased and honored to have known Jim. I'm also proud to be able to work so closely with Microsoft as it continues to support such a widespread and comprehensive list of research projects and research facilities, never knowing which ones will lead to better products and which ones will lead to better knowledge and understanding of computers' capabilities.
