Heterogeneous Databases: Proliferations, Issues, and Solutions

Abstract

The traditional data processing using files, tapes, and manual handling of transactions by operators is being replaced by modern database management using databases, disks, and automatic handling of transactions by database management systems (DBMS). Further, the use of DBMS allows online updates and ad-hoc queries which are not available to data processing. Consequently, the introduction and acceptance of DBMS in recent years have been phenomenal. It is entirely possible that in a few years no traditional data processing facilities will remain operational. Instead, the great proliferation and overwhelming use of DBMS will prevail. However, the proliferation and use of DBMS in a given organization have given rise to the phenomenon known as heterogeneous databases, which introduces new problems and likely issues. Without providing solutions to these newly created problems and to those likely forthcoming issues, the new database management era of heterogeneous databases will not become a reality, which may hinder the replacement of traditional data processing with modern database management. In this paper, we first characterize the notion of heterogeneous databases as the inevitable consequence in replacing the traditional data processing practice with the modern database management. We then articulate the current problems and future issues brought upon us by the great proliferation and overwhelming use of heterogeneous databases and their DBMS in an organization. Next, we survey available DBMS-solutions to these problems and issues. The limitations and advantages of various DBMS-solutions are examined. To this end a taxonomy of DBMS-solutions to the problems and issues of heterogeneous databases has been developed. Some of the current research and proposals on the heterogeneous-database solutions are discussed in the context of the taxonomy. Despite the presence of these problems and the arrival of new issues, promising DBMS-solutions to heterogeneous problems and issues are forthcoming, indicating that the era of effective and efficient heterogeneous database management will begin in earnest and that the momentum in replacing traditional data-processing practices with new heterogeneous-DBMS solutions will be inevitable and accelerated. Further, heterogeneous DBMS may also provide solutions to new database applications. © 1989, SDAIEEE. All rights reserved.