TPC Announces New Decision Support Benchmark; TPC-D measures how quickly real-world business questions can be answered.

SYDNEY, Australia--(BUSINESS WIRE)--April 5, 1995--The Transaction Processing Performance Council (TPC) Wednesday announced a new benchmark, the TPC-D Decision Support benchmark.

Decision support is a rapidly growing class of business applications characterized by a few users executing complex queries against large data volumes. This differs significantly from other commercial environments such as on-line transaction processing (OLTP) where transactions are executed at a high rate by many users to support the operational end of a business. The new TPC-D benchmark models the analysis end of a business where data produced by OLTP applications is queried to compute trends and aid in the decision making process.

"To remain competitive in the 70s and 80s, companies made a tremendous investment in computerizing their internal and customer-related operations and building good OLTP systems," said Kim Shanley, TPC's Chief Operating Officer and Administrator. "OLTP will remain an important competitive differentiator, and consequently, the TPC plans to launch three new benchmarks covering different aspects of OLTP in 1995. But in the 90s and into the next century, decision support will become an increasingly important competitive factor. Decision support is about taking the data being produced in the course of doing business and producing the knowledge necessary to do a better job of running that business. And there are benefits for end-users and customers as well. The better and more quickly companies can analyze their customers' changing needs, the better and more quickly they can satisfy those needs."

TPC-D Queries

The 17 queries that make up the benchmark originate from real-life decision support applications, and they were chosen for being representative of such applications. The queries also cover a wide range of complexity and execution times. Each query is described with the help of a real-world business question it seeks to answer. These business questions are related to activities such as pricing and promotions, supply and demand management, profit and revenue reporting, customer satisfaction study, and market share analysis. They answer questions such as, "Determine the market response to a promotional campaign."

TPC-D Data

The benchmark recognizes that when it comes to decision support, the size of the data set is just as important as the complexity of the queries. The power of a system can be applied toward processing more data, more complex queries, or a larger number of queries. For this reason, test sponsors can choose the size of the database they want to measure against. They can also choose the number of users executing queries.

TPC-D Metrics

The benchmark has three primary metrics:

--The query processing power metric, OppD, which provides a measure of raw query execution and shows how fast queries can be executed when all the available power of the system is concentrated on a single query at a time.

--The query processing throughput metric, OthD, which provides measure of optimal execution of concurrent queries and shows how many queries the system can execute per hour.

--The overall price performance metric, or cost per query per hour, $ per OphD, represents the 5-year cost of ownership divided by a composite of the system's power and throughput metrics.

The name of the metrics include the size of the data set used for the test. For example, a test run against a 10 gigabyte database has a query processing power rating expressed in QppD@10GB.

Database Build Speed

The new TPC-D benchmark also gives, for the first time, information on the speed at which the test database is built. Tests sponsors are required to report the time taken to create the database and to populate it with a predefined set of records. This is a crucial new piece of information that will give users insight on how efficient systems are at building a new database, a somewhat frequent operation in the decision support world.

45 TPC Members Worldwide

The TPC is a non-profit corporation founded to define transaction processing and database benchmarks and to disseminate objective, verifiable TPC performance data to the industry. The TPC was established in August 1988 by eight leading software and hardware companies, and now has 45 members from the United States, Japan, Australia, and Europe. The TPC members are: Amdahl, AT&T, Australian Government, Bull, Compaq, Convex, Cray Research, Data General, Digital Equipment Corporation, Electronic Data Systems (EDS), EMC, Encore Computer, Fujitsu/ICL, Hewlett-Packard, Hitachi, IBM, IDEAS International, Informix, Intel, Intergraph, ITOM International, Microsoft, Mitsubishi, Motorola, NEC, Nikkei Datapro, Novell, Oki Electric, Olivetti, Oracle, Pyramid, Samsung, SCO, Sequent, Siemens Nixdorf, Silicon Graphics, Software AG, Sony Corporation, Stratus, Sun Microsystems, Sybase, Tandem, Toshiba, Tricord and Unisys.

Note to Editors: TPC, TPC-A, TPC-B, TPC-C, TPC-D, TPC-E, tpsA, tpsB, and tpmC are trademarks of the Transaction Processing Performance Council (TPC).

CONTACT: TPC

Kim Shanley, Tracy Derossett, 408/295-8894