Title

Query clearing house approach for multidatabase interoperability

Date of Completion

January 1999

Keywords

Computer Science

Degree

Ph.D.

Abstract

The marriage of databases and advanced Internet technology is providing an unprecedented opportunity for direct access to databases. One obvious research problem is how to let a user access data from multiple databases as freely as possible, despite numerous problems associated with heterogeneities among databases. To support interoperability among those multiple heterogeneous databases, we propose a new model, called Query Clearing House (QCH), which functions as a mediator between the end users and databases. ^ One important prerequisite for building such a mediator is the organization of database schemas in a cohesive way so that a systematic method of determining data relevancy could be developed. A heuristic translation method is proposed to convert relational database schemas into canonical form expressions so that the resulting canonical form expressions can include more descriptive data semantics than the original database schemas. Next, we propose a heuristic method designed to construct meta-data and mapping libraries from those canonical form expressions. The meta-data encompasses all the database semantics plus additional semantic information, such as states and events information in database domains, thematic roles of concepts in event/states, and classification information. The mapping libraries store information about the relationships between meta-data and database schemas. Meta-data and mapping libraries are necessary for the QCH to perform mediation task between the end user and the databases. ^ Finally, we propose an intelligent user query interpretation method that decides which databases are relevant to the query and automatically generates local queries for the relevant databases. In the end, we describe a prototype system that is capable of interpreting user queries on multiple databases and generating SQL queries. ^