A Localized Approach to Query Optimization in Heterogeneous Database System

Abstract

Heterogeneities exist in multiple databases. For example, a real world entity may be differently represented in relations of different databases. In particular, keys of these relations may be incompatible. In this paper, we develop an entity join operator named EJ operator, which can be used to join two relations on their compatible/incompatible keys. By using this join, if an entity is represented in both relations, all its properties given in relations can be obtained. Furthermore, a neighborhood join operator named NJ operator is introduced, which joins relations allowing some degree of value uncertainty on the join attributes. The EJ operator, NJ operator and the previously proposed extended relational operators can be combined to support a more powerful set of operations on heterogeneous databases. Since the EJ operator "integrates" tuples from relations, whose attribute values may have conflicts, a natural way to process the operator is to transmit both relations to a site, resolve the conflicts and process the join, which is very costly. An approach which transforms a global query into local subqueries to process queries with EJs in multiple sites is, thus, proposed in an attempt to lower the query processing cost.