Optimization of search queries with aggregation functions for the conditions of massive virtual integration of databases

The concept of virtual integration assumes that during a query execution the data from several databases are transferred via network to the integration system server where they are processed jointly. One of the possible directions of query optimization is reduction of the transmitted data amount. In this paper a method of optimization for the queries which calculate the aggregation functions of the SQL language (the sum, average et al.) is offered. Algorithms of optimizing transformation over the query execution plan and those ensuring the function evaluation directly in databases are presented. The experimental results have shown high efficiency of the proposed method.