The relation-partitioning approach to processing star queries in distributed databases

In a distributed database system, query processing involves considerable amount of data transmission among different computer sites. Since communication delay is substantial, minimizing inter-site data transmissions becomes an important issue. In this paper, we propose an abstract relation-partitioning approach to the distributed query processing problem. This approach transforms a query processing problem into a pseudo query processing problem, solves the query processing problem in the pseudo space, and then transforms the solution back to the original solution space. We then apply this method to derive optimal algorithms for star queries.