A Simulator for Exploring Replication and Locality of Access in a Distributed Database

Abstract

In fault-tolerant distributed databases, one method of increasing data availability is through increased data replication. However, with increased data replication comes an increase in the time to perform transactions. What is needed is a way in which this and other problems involved with configuring distributed databases can be explored. This paper describes a simulator which provides an accurate model of many different distributed environments when given a set of parameters describing failures, data replication, and site organization. Such a simulator is a useful tool for exploring the behavior of possible distributed database configurations as well as for verifying theoretical results.