An Adaptive Materialization Method for Globally Optimized Query Evaluation in a Hierarchical Mediator System

Abstract

Semantically related mediators form a hierarchical structure based on the definition of views in each mediator, since a view in a mediator is defined by views in other mediators or base tables in several local information servers. A global query is transformed into a set of its sub-queries; and, each sub-query is the unit of evaluation in an information server or a child mediator. Therefore, it is possible to speed up the execution of a global query if the previous results of frequently requested sub-queries are materialized in a mediator. However, the integrating schema in a mediator can be incrementally modified; and, the evaluation frequency of a global query can also be continuously varied. In order to select the optimized set of materialized sub-queries, with respect to their current evaluation frequencies, the proposed method applies a decay factor for modeling the recent access behavior of each sub-query. As a result, it is possible to adjust the optimized set of materialized sub-queries adaptively according to the recent changes in the evaluation frequencies of sub-queries. Consequently, this paper proposes an algorithm for selecting the optimized set of sub-queries to be materialized in each mediator of a hierarchical mediator system adaptively such that the overall query evaluation cost of the hierarchical mediator system can be minimized at all times.