A conceptual framework for next-generation information systems
Sofia Swartz · 1989
With the tremendous increase in the electronic manipulation of data has come the heightened need for more effective storage, retrieval, and integration of information. Due to the evolution of computers and database technology, and the rapid and often unordered proliferation of databases, information no longer resides in one large central repository but is located throughout an enterprise in multiple databases. Researchers and database administrators are looking to distributed database technology to address this problem through a logically integrated interface to physically non-integrated databases. The task of distributed database integration typically requires substantial planning and design of a conceptual schema prior to implementation. The conceptual design adequately describes an integrated view of the databases but at a snapshot in time, and is therefore a static representation of potentially dynamic data. The challenge of this dissertation is therefore to design a conceptual framework for a distributed database system that will force attention away from static structures and schemas to those which will permit dynamic and reconfigurable information systems. This research serves to reevaluate the underlying approaches to multiple database systems. The needs of future information systems are proposed and the current approaches to information integration are critiqued as to their ability to support these needs. In light of the conclusion that the current approaches will not meet the proposed needs, a new approach is suggested, justified, and experimentally tested. The scope of effort is therefore to design and validate a conceptual framework to support flexible information systems through dynamic data integration. The architecture proposed has three main modules: intelligent user interface; dynamic conceptual view; and multiple companion databases. The intelligent user interface functions to interface with the user and assist in identifying and representing the information or transaction desired. The dynamic conceptual view module dynamically identifies and integrates data from the multiple distributed databases in the system. The companion databases accompany each database and contain metadata or self-describing data. This dissertation is not a simple expansion over current research, but rather an investigation of a new approach and architecture to data integration for next generation information systems.