In-Depth

Data Trends Foreshadow DBA Challenges, MS SQL Server Direction

More unstructured data and an emphasis on business intelligence will keep DBAs busy

Four big data trends will impact an enterprise’s data strategy, according to Paul Flessner, outgoing senior vice president of server applications at Microsoft. As keynote speaker at the 2006 PASS Community Summit conference held in Seattle last week, Flessner noted these trends are what also drive development efforts of the next version of his company’s SQL Server product.

At the top of his list of data trends worth watching: the digitally-born data explosion. Flessner says enterprises are now talking in terms of exabytes and beyond, and that most of that new data is unstructured. From video monitoring files to medical images, audio and video files to types yet to be invented, enterprises are not just storing that data but need to analyze it.

Storage and memory innovation will also have an impact on data managers. Citing a 95 percent drop in the cost of storage since 2000, and possible DRAM prices by 2010 down to a penny a megabyte, Flessner says enterprises can expect to do more analysis in memory, speeding results to end users.

That data is no longer just in the data center, he warns. “Lots of data is being stored on laptops; storage growth in the data center is not where it’s happening.” Flash memory is replacing hard drives, and mobile devices will especially benefit from larger data stores and power savings. Users want more offline experiences, with data analysis performed on mobile devices. The upshot: DBAs should expect to spend more time tackling synchronization issues.

Also on his trends list is the evolution of applications, from connected, database-driven programs using a single data source to “occasionally connected”, multi-source, information-based applications. Key to the new technology: SOA, which lets you couple a variety of information sources and scale easily. Applications, he notes, are increasingly being defined by a workflow engine.

Flessner’s group has been paying attention to these trends, he says—most particularly to the move beyond relational databases. “We’re moving from words and numbers to sights and sounds. That will have a big impact on your architecture.” It will also impact what queries users will want to perform. They’ll be looking for patterns in files, for example. That fancy analysis you see on CSI is great, but we’re not there yet, Flessner noted, but users will increasingly demand such analytical capabilities.

Perhaps the biggest emphasis for Microsoft is the connection between SQL and business intelligence. “Enterprises have paid for SQL in order to put their data in. Why would they want to pay to get their data out?” asked his audience. That could explain the improved toolset for data analysis exhibited at the company’s booth at the PASS trade show. Microsoft was demonstrating everything from how to combine OLAP cubes and SharePoint to data analysis capabilities of products from its ProClarity acquisition.

Flessner admits that the current five-year update cycle for Microsoft SQL Server has to be cut, and said the next release, code named Katmai, should be ready in two to three years from the last release (a year ago this month). A service pack, due in the first quarter of next year, will further enhance BI features, implement better compression (to catch up with competitor Oracle, no doubt), and improve the manageability of the product, a key complaint of members, according to PASS President Kevin Kline.

About the Author

James E. Powell is the former editorial director of Enterprise Strategies (esj.com).

Must Read Articles