Multiple query optimization in PBASE/3

We propose a multi-phase multi-query optimization approach (MMO) for parallel databases. MMO takes multiple objectives into account. It overcomes the limitations of two-phase and one-phase approaches, and can be easily integrated into the traditional optimizer. We discuss the strategies and algorithms for searching common operators, rewriting queries, and cost-based optimization in MMO. Theoretic analysis proves that the efficiency of the MMO optimizer is high and the searching strategies for common operators have little influence on the performance.