Access Structures for Predicate Caching Supporting E-Commerce Applications

P. Bodorik, D. Jutla, and Y. Fu (Canada)

Keywords

Distributed and Parallel Computing and Systems,Predicate Caching, Fast Access Methods, Data Structures

Abstract

Caching is an important technique utilized internally by DB systems and also externally by systems to cached data retrieved from DBs. Predicate and semantic caching are methods that manage data, cached from the DB, in units defined by queries' predicates. We use joining predicates to classify queries and allocate the limited space of an in memory cache towards separate buffers, one buffer to each class of queries. We devise efficient data structures to retrieve results of queries from buffers. The data structures avoid duplication of tuples and are also exploited in containment analysis. How effective the technique is evaluated using a modified TPC-W benchmark.

Important Links:



Go Back