Implementing Relational Data Bases: Automating the Mapping Processes

Abstract

The design and implementation of data bases involve four major phases: deriving the business model, achieved by analysing the structure of an organisation and its information requirements; the formulation of the conceptual model, achieved by identifying entities, attributes and relationships; the logical mapping of the conceptual model onto the data structure of the target data base management system, and the physical mapping of this model into the access methods and storage structures used by that data base management system. The mapping between phases should be such that there is no loss of information. The accuracy of both the logical and physical mapping determine the performance of the resulting system. This paper describes algorithms and software tools which have been designed to facilitate the implementation of data bases. Some programs help as documentation aids. Others automate aspects of the processes. In particular rules for automatically mapping the conceptual data model to relational data bases are described, and algorithms are developed for implementing these models on Ingres and Mimer. Copies of the programs can be obtained directly from the authors.