The role of expert systems in federated distributed multi-database systems/Ince Levent

Abstract

A shared information system is a series of computer systems interconnected by some kind of communication network. There are data repositories residing on each computer. These data repositories must somehow be integrated. The purpose for using distributed and multi-database systems is to allow users to view collections of data repositories as if they were a single entity. Multidatabase systems, better known as heterogeneous multidatabase systems, are characterized by dissimilar data models, concurrency and optimization strategies and access methods. Unlike homogenous systems, the data models that compose the global database can be based on different types of data models. It is not necessary that all participant databases use the same data model. Federated distributed database systems are a special case of multidatabase systems. They are completely autonomous and do not rely on the global data dictionary to process distributed queries. Processing distributed query requests in federated databases is very difficult since there are multiple independent databases with their own rules for query optimization, deadlock detection, and concurrency. Expert systems can play a role in this type of environment by supplying a knowledge base that contains rules for data object conversion, rules for resolving naming conflicts, and rules for exchanging data.