News

Microsoft Unveils Data Warehouse SQL Server Solutions

• By Kurt Mackie
• 02/23/09

Microsoft and its hardware partners on Monday rolled out preconfigured data warehouse reference architectures that incorporate Microsoft SQL Server 2008. The new offerings are part of a "SQL Server Fast Track Data Warehouse" program and feature hardware server products from Bull, Dell, and HP.

The solutions promise scalability for enterprise data warehousing applications. The best hardware-software solutions can scale up to 32 terabytes, according to an announcement issued by Microsoft.

Microsoft has made considerable progress in the data warehouse market in competition with IBM and Oracle, according to James Kobielus, senior analyst at Forrester Research and author of "Forrester Wave for Enterprise Data Warehousing Platforms." He estimated that Oracle Database, IBM DB2 and Microsoft SQL Server are used in about one third of data warehousing apps. Two thirds of those deployments are for transactional apps, such as "OLTP, ERP, CRM" and others, he added, in an e-mailed response.

Microsoft is also claiming that the SQL Server Fast Track Data Warehouse offerings represent an integral step toward the massive scalability promised by Madison, a Microsoft code-name project that aims at integrating parallel processing technologies acquired from DATAllegro.

Microsoft closed its acquisition of DATAllegro back in September and is planning to launch a product that can scale to hundreds of terabytes of data sometime in the first half of 2010.

Kobielus explained that Microsoft is currently "playing catch-up with rivals" that already have a scale-out technology called "shared-nothing massively parallel processing." However, with the completion of its Madison project, Microsoft will be able to run with the pack.

"We’re confident that 'Project Madison,' when rolled out in beta later this year or early next year, will provide a very credible platform for that vendor to compete more fully with Teradata, Oracle, IBM, Greenplum, and others for the petabyte-scale data warehouses that customers in some verticals -- such as telecommunications and Web 2.0 -- are deploying for various applications," Kobielus wrote.

Various hardware products have been tested under the SQL Server Fast Track Data Warehouse partnership program.

Bull offers data warehouse capability through its NovaScale R460 E1 and NovaScale R480 E1 servers using four-core and six-core Intel Xeon processors.

Dell has two options. The Dell Power Edge 2950 server has two Intel Xeon quad-core processors with a total of eight CPU cores, supporting four terabytes of storage. There's also the Dell Power Edge R900 G5, which has four Intel Xeon six-core processors for a total of 24 CPU cores and 12 terabytes of storage.

HP offers three server options. The HP ProLiant DL385 G5p server has two AMD Opteron quad-core CPUs for a total of eight CPU cores and four terabytes of storage. The HP ProLiant DL585 G5 has four AMD Opteron quad-core CPUs for a total of 16 CPU cores and eight terabytes of storage. Lastly, HP offers its top model, HP ProLiant DL785, with eight AMD Opteron quad-core CPUs for a total of 32 CPU cores and 16 terabytes of storage.

Microsoft's announcement suggests that the SQL Server Fast Track Data Warehouse program provides performance and scalability "out of the box with minimal tuning required." The prices start at $13,000 per terabyte.

Microsoft is also partnering with system integrators to deliver "solution templates specially tailored for the new reference architectures," according to its announcement. Those system integrators include "Avanade, Cognizant Technology Solutions, HP and Hitachi Consulting Corp."

Does business really need the kind scalability promised by Microsoft's SQL Server Fast Track Data Warehouse offerings? Kobielus thinks the answer is yes. He's authored a new report to be published next month by Forrester on the topic, which predicts a need for data sets ranging from the high gigabytes to thousands of terabytes (petabytes).

"Enterprises everywhere continually struggle to keep their enterprise data warehousing (EDW) environments from being swamped by the inexorable pressure from growing data and usage volumes for business intelligence (BI), predictive analytics, data mining, and other key applications that hinge on information maintained in the EDW," he wrote.