Managing cache for efficient query processing

Abstract

This work discusses the opportunities and mechanisms to leverage query processing performance using caches. Our first set of techniques is realized in a centralized multi-user environment where we propose a non-speculative demand-driven caching: Conform-CoD, Scramble-CoD and Integrated-CoD, where the exact cost of investment and the return on investment are known, and the cache is certain to be reused. Also, we further extend the mechanism to exploit pipeline plans. Next, we move into a distributed environment, and explore the caches for assisting query answering. We propose 2 strategies: promotes the reuse of the cache content and improves query answering through recall-routing. Lastly, we propose a cache coherence method to update the caches for a hierarchy of network servers which includes 2 mechanisms: data scoping and delta profiling that work in conjunction with both the eager and lazy update models.