A relational data base machine employing associative memories and transposed files

Abstract

This paper presents the design of a relational data base machine which exploits features of the transposed file structure and the search capability of an associative memory to enhance the performance of relational database operations. The transposed file storage organization facilitates a significant reduction in the amount of data transfer, that is other-wise required for the processing of queries on very large relations of the database. Application of associative memories to the implementation of relational database operations, although promising, has shown to introduce several drawbacks. Utilization of these powerful search devices can be improved if they are designed according to the requirements of the application. The design proposed here is that of a hierarchically structured associative memory and a parallel pipeline processor which meet the potential requirements of relational database operations.