Categories Software

By Gabrielle Gagnon

Data warehousing is the technology trend most often associated with enterprise computing today. The term conjures up images of vast data banks fed from systems all over the globe, with legions of corporate analysts mining them for golden nuggets of information that will make their companies more profitable.

All of the developments in database technology over the past 20 years have culminated in the data warehouse. Entity-relationship modeling, heuristic searches, mass data storage, neural networks, multiprocessing, and natural-language interfaces have all found their niches in the data warehouse. But aside from being a database engineer's dream, what practical benefits does a data warehouse offer the enterprise?

When asked, corporate executives often say that having a data warehouse gives them a competitive advantage, because it gives them a better understanding of their data and a better understanding of their business in relation to their competitors, and it lets them provide better customer service.

So what exactly is a data warehouse? Should your company have one, and if so, what should it look like?

Essentially, a data warehouse provides historical data for decision-support applications. Such applications include reporting, online analytical processing (OLAP), executive information systems (EIS), and data mining.

According to W.H. Inmon, the man who originally came up with the term, a data warehouse is a centralized, integrated repository of information. Here integrated means cleaned up, merged, and redesigned. This may be more or less complicated depending on how many systems feed into a warehouse and how widely they differ in handling similar information.

But most companies already have repositories of information in their production systems, and many of them are centralized. Aren't these data warehouses? Not really.

Data warehouses differ from production databases, or online transaction-processing (OLTP) systems, in their purpose and design. An OLTP system is designed and optimized for data entry and updates, whereas a data warehouse is optimized for data retrieval and reporting, and it is usually a read-only system. An OLTP system contains data needed for running the day-to-day operations of a business, but a data warehouse contains data used for analyzing the business. The data in an OLTP system is current and highly volatile, with data elements that may be incomplete or unknown at the time of entry. A data warehouse contains historical, nonvolatile data that has been adjusted for transaction errors. Finally, since their purposes are so different, OLTP systems and data warehouses use different data-modeling strategies. Redundancy is almost nonexistent in OLTP systems, since redundant data complicates updates. So OLTP systems are highly normalized and are usually based on a relational model. But redundancy is desirable in a data warehouse, since it simplifies user access and enhances performance by minimizing the number of tables that have to be joined. Some data warehouses don't use a relational model at all, preferring a multidimensional design instead.

Gabrielle Gagnon is a free-lance journalist and frequent contributor to PC Magazine.

Next: Types of Data Warehouses

Published as Enterprise Computing in the 3/9/99 issue of PC Magazine.

Related Links
• ZDNet Enterprise