Matching techniques in global schema design

Abstract

A global schema is an integrated view of heterogeneous databases. Identifying common types of information from the various local Schemas is an important part of designing a global schema. After matching the entity types and attributes, the designer constructs the global schema by applying a set of integration operators. In this paper, we propose two types of matching techniques and discuss their use in global schema design. Entity type matching partitions the entity types of the local Schemas into subsets with common information. Attributes within each entity type cluster are then matched according to the similarity of their meaning and scope. We define assertion types for both kinds of matching and provide an algorithm to check the consistency and completeness of attribute assertions.