The design of object orientated databases

Abstract

Object orientation is currently a hot issue in computer science. Supporters of this methodology see many benefits from using techniques based on object orientation : better organization and design of the system and an added possibility for reusability of many different parts of the system (reusability of code, design methods, whole application system frameworks). Object oriented techniques are already used in many different fields, e.g. object oriented programming languages, object oriented operating systems, and simulation of real time systems. This project will explore object orientation in database systems. It has two important goals : The main (most important) goal of the project is to explore the differences between two different techniques of designing and developing object oriented databases. The two techniques that we are investigating are transferring an existing relational database to an object oriented database and reengineering a system to an object oriented database. Transferring an existing relational database to an object oriented databases involves translating from the relational model to the object oriented model. This technique is very important now, when the object oriented model is becoming more prevalent and many systems are implemented using relational databases. Reengineering a system involves defining the requirements of an existing system and building the system, based on these requirements from scratch. Firstly we examine algorithms to transfer existing relational databases to object oriented databases. Findings from this process have a large practical value. Secondly we examine the process of designing and developing object oriented databases using purely object oriented modeling concepts. These findings will become very important in the future as object oriented databases become more mature. While examining this process, we will attempt to answer the following questions : how should this process be performed and what sub-processes should it consist of; what is a useful notation for object oriented database design and how should it be used; what is the final result of the object oriented database design process. The main question we are attempting to answer is : what arc the advantages and disadvantages of the two techniques and how do the final results of the two techniques differ. The second (less important) goal of this project is to perform a general investigation of the object oriented techniques in database systems : exploring differences between relational database systems and object oriented database systems. This goal is not as important as a first and is more an additional interesting investigation which may be performed during a work on the first problem. I will try to analyze the advantages and disadvantages of both the relational and object oriented database models. The relational database model is currently the most widely adopted commercially, so it is important to discover what is good in that model which should be preserved in the object oriented one. Also I will try to examine what is bad in the relational model, because it is also important to search for possible improvements.