Very Large Databases

Very Large Databases (VLDB)

Imagine if all the snows falling so soundlessly across the land were data. Pile up in 20-foot drifts it might, but no one would dare dump it into New York's East River. Sure as the seasons, databases grow bigger every year. More data begets more data; and like snowflakes, each piece is unique and meaningful. The recent attention paid to decision-support systems and data warehouses is evidence that companies understand the value of data: It is the raw material of information, which is today's biggest competitive advantage.

Very Large Databases (VLDB) is all about stretching database technology to its limits. At the 1996 VLDB Summit (April 28 to May 1 in Chicago), Nautilus Systems' president, Dr. Jen Que Louie, presented a paper on the metrics defining VLDBs. We learned that sheer data volume is only one measure of VLDBness; others include number of users, number of rows, and level of availability.

VLDBs are measured in three ways:

Total storage under the management of the DBMS (measured in gigabytes),
Total number of rows or records (counted in millions), and
Total number of concurrent online users.

The Winter Group conducts an annual survey of VLDBs that focuses on database size, activity, products used, practices, production status and use of the systems. There have been 120 databases registered as a result of the survey, two of which are large transaction processing systems that run under Unix. The ten largest Unix data warehouses range in size from 150GB to 700GB, the largest of which is run by Value Rx Pharmacy and supports 400 concurrent users. Caliber Technology Inc's transaction processing system supports 2,300 concurrent users on an IBM ES9000 using the Model 204 DBMS. The system tracks freight for Roadway Express.

Nautilus Systems' experience with VLDBs started with acoustic analysis and evolved to include business data analysis, decision-support systems (DSSs), executive information systems (EISs), Geographic Information Systems, and multidimensional operations.

[ Clients | Hardware and Software Experience ] [ Computer and Network Security | Geographic Information Systems ] [ Data Warehousing and Data Mining | Competitive Intelligence ] [ Recommended Books ]