RDBMS Speculative Support Improvement by the Use of the Query Hypergraph Representation

Abstract

The paper concerns the methodology for speculative support for query execution in Relational Database Management Systems (RDBMSs). It discusses and develops our proposal of supporting the RDBMS query execution based on a graph-based analytic approach. This approach assumes using the results of speculative queries defined by analyzing a multigraph representation of a stream of input queries arriving to a RDBMS. The queries from the RDBMS input stream are permanently analysed using a Speculation Window moving on the query stream to define the optimized speculative queries for execution. More specifically, the current paper develops the basic idea of the speculative query support and shows how the use of the speculative query results by the awaiting input user queries can be improved by the analysis of a query hypergraph representation. The analysis of a joint hypergraph representation of the speculative and user input queries has been employed to better organize the use of the speculative query results in the stream of input user queries of a RDBMS. The advantages of the proposed approach are positively assessed based on the experiments with the simulated execution of the testbed sets of queries.