Paul Flessner, the company's senior vice president of server applications, on Thursday sketched out the software giant's design goals for the next two versions of SQL Server. He added that sales of the recently released SQL Server 2005 have "stunned" company executives, contributing to 20 percent revenue growth during the past two fiscal quarters.
He said Microsoft this summer will release an early version, or "community technology preview," of a new database, called SQL Server Everywhere Edition, set to be completed in the second half of this year.
Paul Flessner
Microsoft senior vp
server applications
SQL Server Everywhere is an "embeddded" database, used to store data on small devices, such as mobile phones, rather than require users to connect to a server to access information. Flessner said Microsoft had already developed the embedded database for internal use but will now release it as a commercial product.
Other companies, such as Oracle, IBM and Sybase, already offer embedded databases. In addition, there are several open-source options such as Sleepycat, which was recently purchased by Oracle.
Looking ahead at future products, Flessner said Microsoft has identified a few trends in data and storage that will dictate the planned features for the next two versions of SQL Server.
Specifically, Flessner said forecasts indicate that data will continue to multiply rapidly; the cost of storage will drop dramatically, to the point that 1 terabyte of data can be stored on $100 disks by next year; and powerful devices, such as phones, cameras and digital-music players, will store more and more data.
To address these changes, Microsoft is investing in techniques to better store images, music files and other types of "unstructured data," Flessner said.
In addition, he said database engineers are working on improving tools to analyze data and to administer databases.
To make it easier to share information between client devices and servers, Microsoft will invest in tools for developers to ease synchronization of data between machines, Flessner said.
"We're not naive enough to believe that all data will be in one central database," he said. "People are not going to walk around with a terabyte in their pocket and not make use of it."
