Graph based query trimming algorithm for relational data semantic cache

Abstract

Semantic cache enhances the capability of conventional (page/tuple) cache by adopting the dynamic strategy to group the contents and semantics of already processed queries. Query processing and cache management are two major activities for semantic caching. Semantic caching demands efficient, correct and complete algorithms to process incoming queries. Efficiency of query processing activity in semantic cache mainly depends on query trimming. Query trimming process consists of two sub processes; query matching and query splitting. Efficient query matching and query splitting process will ensures the optimal query trimming process. In this paper, we have proposed query trimming algorithm by defining query matching algorithm and using query splitting algorithm from the literature. To optimize the query matching process we have proposed graph based semantic indexing scheme. On the basis of proposed indexing scheme we have designed query matching algorithm (qTrim). We have proved that complexity of graph based query matching algorithm is reduced to linear from exponential. Comparison of graph based (proposed) scheme is done with segment based scheme. On the basis of comparison and runtime complexity we have argued that the proposed algorithm is an optimum to trim the queries into sub-queries (probe and remainder).