A survey of commercial object-oriented database management systems

The object-oriented data model is the culmination of over thirty years of database research. Initially, database research focused on the need to provide information in a consistent and efficient manner to the business community. Early data models such as the hierarchical model and the network model met the goal of consistent and efficient access to data and were substantial improvements over simple file mechanisms for storing and accessing data. However, these models required highly skilled programmers to provide access to the data. Consequently, in the early 70's E.F. Codd, an IBM research computer scientists, proposed a new data model based on the simple mathematical notion of the relation. This model is known as the Relational Model. In the relational model, data is represented in flat tables (or relations) which have no physical or internal links between them. The simplicity of this model fostered the development of powerful but relatively simple query languages that now made data directly accessible to the general database user. Except for large, multi-user database systems, a database professional was in general no longer necessary. Database professionals found that traditional data in the form of character data, dates, and numeric data were easily represented and managed via the relational model. Commercial relational database management systems proliferated and performance of relational databases improved dramatically. However, there was a growing community of potential database users whose needs were not met by the relational model. These users needed to store data with data types not available in the relational model and who required a far richer modelling environment than that provided by the relational model. Indeed, the complexity of the objects to be represented in the model mandated a new approach to database technology. The Object-Oriented Model was the result.