Response time analysis of using a parallel join algorithm on a multiprocessor backend machine

摘要

The join operation is one of the most important operations of the relational data model. It participates in all queries which involve more than one relation. Owing to inequality join, the paper proposes a parallel algorithm which not only eliminates cross referencing but also maintains ordering reservation. To realize the algorithm, a multiprocessor machine called JAMAC (join-accent machine) is also proposed for executing the general join operation (includes equi-join and inequality join operations) efficiently. The response time in carrying out the join operation is derived, by comparing the processing rates of hardware modules, and analyzed.