Abstract
FDBMS (Federated Database Management System), a new trend of DBMS, are capable of integrating several heterogeneous databases, which are managed by their respective rules, for search and queries. FDBMS integrate databases without any changes to the individual databases nor any need to treate another database. Only a few FDBMS, among those recently proposed in the information science field, support heterogeneous attributes.
We have developed a new FDBMS applicable to relational databases, that supports conversions of both schemas and attribute values and thus creation of a global schema, an integration of schemas of individual databases. This FDBMS consists of a GUI support system and a query processor system. The GUI system helps the user produce a global schema. The query processor accepts an SQL statement querying the global schema, converts into several SQL statements querying local tables, and reports the results. This FDBMS trialed in two cases, and its validity has been proved.
Case 1:
"Vegetable and fruit market information databases covering different time periods" The databases have different identification code systems, since they relate to different time periods, and the data structures have also evolved. We performed integrated queries which involved DBs from several time periods.
Case 2:
"Farming System Databases from Different Prefectures" Again, the databases do not share the same schema, rather each has its own rules. We performed integrated queries for information commonly found in these databases without intruding on each DBs control of its local autonomy.
