A Multiple-Plan Model for Rule-Based Optimizer
Wen-Yen Wang · 2002
In this paper, we propose a multiple-plan model extending database optimizer to yield multiple plans per query for run-time resource availability. Conventional query optimizers generate an optimal p1an for each query by making assumptions about run-time resource avai1ability. These assumptions are often not accurate at run-time. This research focuses on multiple-plan generation at optimize-time. In this work, the author bui1ds a rule-based optimizer that transforms a query into multiple plans, where each one is optima1 for different buffer availabilities. The mode1 using hot points as buffer requirements and defined in the Optimizer Specification Language (OSL) produces multip1e plans for each query.