Modeling a Large Scale Data Base.

This document reports the methods and results of a study of a large database through use of modeling concepts and simulation software. The simulation software had been developed using the concepts of the four level structure of the Data-Independent Accessing Model DIAM and the nomenclature and concepts of the Relational Model. The study was undertaken to assess the feasibility of applying these concepts to modeling a large existing database--in this case, the USAFs PACER database--and to demonstrate the use of existing components of query compiler software used to implement the concepts. The report describes a model of the information content of a subset of the PACER database the model is expressed as n-ary relations. In a subsequent phase of the study, the investigators developed a model of the access paths in the PACER system. The model uses the strings of the DIAM descriptive technique. Based on analysis of the results, the author concludes that DIAM concepts enhanced by Relational Model nonmenclature can serve well as the basis for study of large databases and that prototype software correctly compiles all paths applicable to an input query.