Practical Optimal Caching Using Multiple Virtual Caches in Multiple Query Optimization

Abstract

Databases today are increasing in both complexity in size and have gained an unprecedented level of importance in commercial and business applications. Of particular interest are the databases based on the client-server model where a centralized database services multiple clients over a network. Clearly, this means that a large number of queries are fired at the central database in a short period of time. Realistically speaking, it is not possible to process the queries in conventional methods and some optimization techniques are used to speed up the entire process. One way to do this optimization is to cache common results of sub-queries to avoid re-computation and redundancy. In the following paper, we demonstrate the use of multiple virtual caches, a novel concept which allows us to practically emulate the optimal caching algorithm with the limited knowledge of the future known by the optimizer. The approach presented significantly speeds up the entire process of query processing and improves both the utilization of the cache and throughput of the CPU. Towards the end of the paper, we will show the results of our performance evaluation, drawing comparisons with the existing approaches that are currently employed.