An initial report on the design of Ariel DBMS with an integrated production rule system

Abstract

The design and implementation strategy for Ariel, A DBMS with a built-in production rule system (a trigger system), is described. Ariel is being built with the EXODUS database tool kit. The query language of Ariel is a subset of POSTQUEL extended with a new rule language. Effort is focussed on integrating the rule system with transaction processing, and making the rule system powerful and efficient. The current implementation of the rule condition-testing mechanism in Ariel is based on a variation of the Rete network, a type of discrimination network used widely in rule-based systems tools for artificial intelligence programming. The standard Rete algorithm has been extended to improve performance in a database environment.