ACTIVE DATABASE MANAGEMENT SYSTEMS

Abstract

Conventional passive database management systems are inadequate for time-constrained applications, because they either do not provide timely response to critical situations or compromise modularity. Active database management systems attempt to provide both modularity and timely response, by allowing event-condition-action rules to be specified declaratively; when events of interest occur, they efficiently evaluate the corresponding conditions, and if these conditions are satisfied, they trigger the corresponding actions. The development of active database management systems requires the solution of a number of research problems in the areas of knowledge modelling, execution modelling, condition monitoring, scheduling, system architecture, and performance evaluation. This paper describes the principal research issues in each of these areas, surveys the approaches being taken in a number of research projects on active DBMSs, and emphasizes the approaches we are taking in the HiPAC (High Performance ACtive DBMS) project.