A Model for Equi-join Query Processing in Distributed Relational Databases

摘要

We develop a mathematical model to compute the minimum communication cost of a join-semijoin program for processing a given equi-join query. Some definitions and conditions upon which this paper is based are stated. We define a query processing graph for each equ-join query and characterize the set of join-semijoin programs which solve this query. A rule for estimating the size of the derived relation is derived. The parameters for estimating the size of derived relation form a consistent parameter system. With the assumption of communication cost dominance, the cost functions are linear in the size of data transmission. An optimization problem for distributed query processing is well formulated.