Validation and performance evaluation of the partition and replicate algorithm

The partition-and-replicate-strategy (PRS) algorithm for distributed query processing is evaluated and its performance is validated. Although in principle PRS is better than single-site processing, early experimental results indicate the contrary. Based on experimental results, the factor which causes performance deterioration is identified and a remedy is provided. As a result, it is shown that the PRS strategy outperforms single-site processing in a realistic environment and that various parameters, such as the number of processing sites, partitioning speed relative to join speed, and sizes of the join relations, affect the performance of the PRS strategy significantly. Among these parameters, the algorithm is most sensitive to the partition speed.<
>