Hybrid simultaneous scheduling and mapping in sql multi-query parallelization

Abstract

In the context of the database query optimization on a parallel architecture, we focus here on the problem of dynamically mapping the tasks making up several SQL queries onto a shared-nothing parallel architecture. The main contribution of this paper lies in the proposal of an incremental parallelization strategy which carries out simultaneously both scheduling and mapping in co-operation with memory allocation in a dynamic multi-user context. The paper functionally describes in terms of components our dynamic PSA/MEG-based parallelization strategy in a multi-user context, as well as one of the two incremental memory allocation heuristics, called ModDeg, which are integrated into the MEG mapping method, and presents a sum-up of the performance evaluation of this heuristic.