A Decision Support Tool for Distributed Database Design

Abstract

The efficiency and effectiveness of a distributed database depend

primarily on solving two interrelated design problems: data allocation,

specifying what data to replicate and where to store it, and operating

strategies, specifying where and how retrieval and update processes are

performed. We develop a distributed database design approach that

comprehensively addresses these problems, explicitly modeling their

interdependencies for both retrieval and update processing. We extend

earlier distributed database design models to include join order and

data reduction by semijoin, in addition to data replication, copy

identification, and join node selection. We demonstrate that join

ordering and data reduction by semijoin are important distributed

database design decisions that must be included in a distributed

database design algorithm if it is to determine an overall optimal

distributed database design.