Parallel processing of relational databases on a cellular tree machine

Abstract

This paper investigates the use of a cellular tree computer, the FFP machine, for parallel processing of relational databases. This machine directly executes Formal Functional Programming (FFP) code. A method to embed SQL statements in FFP programs is described. Schemes for parallel implementation of simple retrieval statements are presented for different I/O architectures. Then techniques for parallel implementation of more complex retrieval statements are given. Because the FFP machine can accommodate the almost unbounded parallelism of FFP programs, execution time analysis shows potential for highly efficient execution of relational database queries on this machine. Modifications of the FFP machine to speed up input/output operations are discussed.