COST EVALUATION MODEL FOR PROCESSING DISTRIBUTED QUERY.

Abstract

The efficiency of query processing has a strong influence on the performance of distributed database management systems. Estimating the size of a derived relation is the essential components for finding an optimal distributed query strategy in a distributed database. In a local area network environment a query can be solved by a sequence of semijoin and join operations with the transmission of data between nodes. Different ordering of the sequence of operations will result in different data transmission costs. This paper introduces the notions of join and semijoin reducibilities and provides a model for estimating the size of the derived relation after operation. Thus, the evaluation of data transmission costs for processing a distributed query can be minimized. This model is also directly applicable to distributed query processing in database machines which are either multiprocessor interconnection structure or VLSI architecture implementation.