Parallel Query Processing in DBS3

In this paper, we describe our approach to the compile-time optimization and parallelization of queries for execution in DBS3, a shared-memory parallel database system. Our approach enables exploring a search space large enough to include zigzag trees which are intermediate between left-deep and right-deep trees. Zigzag trees are shown to provide better response time than right-deep trees in the case of limited memory. Because DBS3 implements a parallel dataaow execution model, this approach applies to both shared-memory and distributed-memory architectures. Performance measurements run using the DBS3 prototype show the advantages of zigzag trees under various conditions.