Relationship based replication algorithm for data grid

Data Grid is an infrastructure that manages huge amount of data files and provides intensive computational resources across geographically distributed systems.To increase resource availability and to ease resource sharing in such environment, there is a need for replication services.This research proposes a replication algorithm, termed as Relationship based Replication (RBR) that integrates users, grid and system perspective.In particular, the RBR includes information of three different relationships in identifying file(s) that requires replication; file-to-user, file-to-file and file-to-grid. Such an approach overcomes existing algorithms that is based either on users request or resource capabilities as an individual. The Relationship based Replication algorithm aims to improve the Data Grid performance by reducing the job execution time, bandwidth and storage usage.The RBR was realized using a network simulation (OptorSim) and experiment results revealed that it offers better performance than existing replication algorithms.