A Strategy for Data Replication in Data Grids

Data replication strategy and replica selection are the most important topics in the data grid. This paper presents a new strategy for data replication in the data grid environment. The strategy combines data replication algorithm with job scheduling policy and is able to select best data replicas in a dynamic grid environment by monitoring the data replication processes, keeping the efficient ones and abandoning the inefficient ones. Grid community-based data transfer and multiple replications of fragmented data from multiple data sources greatly improve the replication efficiency. The strategy can also support various data replication and job scheduling algorithms.