Efficient rule matching in large scale rule based systems

The paper presents an efficient rule matching algorithm for a large scale production system. The matching algorithm is state saving and does incremental evaluation. Implementation details are presented which include optimization of join tests, and efficient buffering of data blocks. A cost analysis, both in terms of matching evaluation cost and storage cost for the saved state is presented. Results from the performance study show that substantial savings in matching cost are obtained with little space overhead for the saving state. Matching becomes computationally intensive in a secondary memory environment, and efficient algorithms are a must for successful integration of production systems and databases. The authors show how the OPS5 and relational model are compatible, and thus implementation techniques in one domain are applicable to the other