Intelligent Enterprise

Brenda Castiel

Dan Brown

You're thirsty. You go to the refrigerator, open the door and retrieve your beverage of choice. That was easy. You rely on that appliance to be there, ready to meet your needs, anytime.

Similarly, data warehouse appliances " quick to set up and install, easy to manage, light on support costs " are just there, ready to satisfy the business users' never-ending thirst for information. The past year has seen significant growth in the sales of these "containers of knowledge," with one vendor having doubled its customer base in 2006.

A data warehouse appliance is specially built hardware and software that's integrated and packaged to support business intelligence applications. The main distinction from other types of data warehouse solutions is the configuration of servers, memory, storage and software, with redundancy to provide high availability. Emerging vendors in this market also tout appliance advantages including ease of use, low acquisition cost, and economical power and cooling requirements.

Where did these new gizmos come from and why are they appearing now? More importantly, what needs do appliances fill and when are they the right choice for your organization? This article provides a detailed analysis of the characteristics and architecture of these products and offers guidelines on how and when to choose this increasingly attractive warehousing alternative.

Setting the Stage

IDC predicts that the market for business analytics will grow at a rate of 10 percent over the next five years, and Gartner and IDC surveys find that BI is either the number-one or number-two priority for most organizations.

Compliance has increased interest in analytics and BI because companies need to have confidence in their numbers and be able to access business information faster. Another trend driving interest in BI is the desire to improve employee accountability and performance management. This is pushing BI out to more users and inspiring more companies to invest in analytics applications and tools.

Meanwhile, data warehouses are getting bigger. "Managing Data Warehouse Growth" documents that the largest data warehouse in 1995 held one terabyte, but 10 years later the largest was 100 times larger. The growth curve is accelerating dramatically, with data warehouse volumes tripling every two years. Indications are that 500 terabyte data warehouses are not far off.

The data-growth trend demands more than just huge databases. These masses of information have to load quickly, too, as customers demand ever-more-current data. Often the data must be available for analysis while there is still time to affect the business transaction in question, so continuous, fast loads are becoming the new norm.

Query complexity is also increasing, with customers running " or hoping to run " queries that would not have been attempted in the past. Complex analyses such as those used in insurance-claim-fraud detection, Web-clickstream analysis, and customer-churn analysis are now so important to organizations in terms of ROI and competitive advantage, they must be run.

Thanks to regulatory compliance demands and increased tactical use of BI, data warehouses and DBMS systems must accommodate mixed analytic, operational and transactional workloads.