Performance measurement and data base design

Abstract

Good data base design is of increasing concern to systems designers and data base administrators. The rapidly increasing use of data base management systems to support an increasing number of dynamic information systems has brought data base performance measurement and evaluation into the limelight. The task is to provide as easily as possible the most cost effective data base structures (logical-physical structures, physical access paths, storage mappings, etc.) to support the given logical system requirements while meeting a number of constraints such as integrity, flexibility, etc. In environments where data base contents and transaction traffic are changing, the determination of the best data base structures is a complex task, particularly with the multiple criteria of access time, storage utilization, etc. This session strives to present and discuss quantitative-oriented performance experiences, experimental results, insights and tools available toward good data base design in actual systems and situations (particularly IMS, System 2000, and ADABAS).