Abstract:

The parallel algorithm for the JOIN operation has been one of me not topics in parallel databases. A series of  parallel algorithms for JOIN or their   improved forms have been obtained early or late. But when they are discussed in a comprehensive way, they have the problems of low communication efficie ncy and frequent load balancing. To overcome these problems, by analyzing and comparing the previous efforts, the paper improves the algorithm SABJ＋ and the algorithm ABJ＋ to obtain a highly efficient parallel algorithm of JOIN：the ABJ＋＋.

Key words: serial JOIN algorithm, parallel JOIN algorithm, parallel Hash JOIN algorithm, dividing bucket