Intelligent caching and indexing techniques for relational database systems

摘要

Recent developments in the design of database systems include proposals for several extensions to the basic model of relational database systems. In this paper we present some ideas on the design of intelligent support mechanisms for large databases where procedures in the form of database commands are stored in relation fields. First, we examine the idea of storing results of previously processed procedures in secondary storage (caching). Problems associated with cache organizations, such as replacement policies and validation schemes, are examined in depth. Another means for reducing the execution cost of queries is indexing. Conventional indexing techniques assume that all values are known, such schemes cannot be used effectively. As a solution to that problem, a new indexing scheme, Partial Indexing, is proposed and analyzed. Uses of partial indexes in conventional database systems are also described.