Abstract
A good database design brings total attention to data. Data have been recognized as a valuable resource, resulting in an increased use of database systems. A wide variety of models and techniques are available to aid in system development at the analysis and logical stage. Once the system is defined at the logical stage, many other decisions have to be made to define the construction, operation and maintenance of the database system at the physical stage.
Database design whether physical or logical is the process of determining the organization of a database. The concept of logical and conceptual modelling has been widely researched. However, few researchers seem to have considered physical modelling explicitly. Physical modelling is the conversion of a logical model to a physical model that will be suitable for a specific system with its own hardware configuration (Batini, Ceri et al. 1990).
The objective of this thesis is to propose a physical database modelling technique that will demonstrate a join between logical and physical design to improve the performance of a database system using physical tuning techniques. It will annotate an entity model with physical detail that will provide a framework for building a physical model.
To substantiate the relevance of the physical database model, the author looks at the physical storage structures in a database system and other physical database models that have been previously developed.
This thesis will use a moderately realistic database system case study. The case study is used as an example for interpreting possible physical problem areas in a "real" database system. The aim will be to model and optimise the physical design of the discussed case study.