Year

1999

Degree Name

Master of Science (Hons.)

Department

Department of Computer Science

Abstract

This thesis examines the proposition that because of recent huge increases in processing power, disk and memory capacities the commercial mainstream object relational databases may now be a viable option to replace dedicated real-time databases in industrial automation. The benefits are lower product cost, greater availability of trained manpower for development and maintenance and lower risks due to larger installed base and larger number of platforms supported. The issues considered in testing this proposition were performance, ability to mimic critical real-time database features, replication of the real-time database application development and administration tools and finally the low overhead high speed, real-time data compression facility available in real-time databases. An efficient yet simple real-time compression algorithm was developed for use with relational databases and benchmarked. Extensive comparative benchmarking has been done to convincingly prove the proposition. The results overwhelmingly show, that for a majority of industrial real-time database applications, the performance offered by a commercial object relational database on a current platform are more than adequate.

Share

COinS