Strategies for semantic caching

Abstract

One major problem with the use of mediator-based architectures is long query response times. An approach to shortening response times is to cache data at the mediator site. Recently, there has been growing interest in semantic query caching which may generally outperform the page and tuple caching approaches. In this paper, we present two semantic-region caching strategies with different storage granularities for mediators accessing relational databases. In contrast to most existing approaches, we do not only cache the projection result of queries but also the condition attributes, resulting in a higher cache hit rate. Additionally, we introduce the Profit-based Replacement algorithm with Aging Counter (PRAG), which incorporates the semantic notion of locality into the system.