An Optimization Method of Data Communication and Control for Parallel Execution of SQL Queries

. This paper describes a method for optimizing data communication and control for parallel execution of SQL queries. The optimization consists in avoiding the data communication and control which are not needed for the global parallel execution consistent with the execution plan generated by the optimizer. The basis of the method lies in the propagation of the partition attributes and of the number of processors during the relational operations parallelization phase. Performance evaluation shows the impact the propagation process has on improving response time. The main contribution of this work is the homogeneous integration in the parallelization step of the propagation process for partition attributes and number of processors in order to decrease communication volumes. 1. Introduction and Motivation Optimization of parallel execution in a parallel architecture now represents a critical step point in the evaluation of simple and complex SQL queries. Two types of parallel databases s..