A load balancing approach for parallel database machines

Parallel database systems have become a major tool for high performance information processing. These systems require efficient load balancing approaches to partition each relation and to allocate them to the parallel architecture. If the database is not partitioned in a balanced way, the execution of database operations might waste resources, reducing the performance of parallel system. In this study, a multiattribute declustering approach is proposed for parallel database systems. This approach partitions the multiattribute file so as to generate blocks for any given number of processors. Then, an efficient allocation algorithm is applied to map adjacent blocks to neighboring processors considering the underlying architecture. Therefore, operations such as join, requiring high block interactions are executed efficiently on any network topology. Our algorithm is applied to both uniform and nonuniform data distribution cases and the results show that if balances the data load for any kind of data distribution.