An on-line replication strategy to increase availability in Data Grids

Data is typically replicated in a Data Grid to improve the job response time and data availability. Strategies for data replication in a Data Grid have previously been proposed, but they typically assume unlimited storage for replicas. In this paper, we address the system-wide data availability problem assuming limited replica storage. We describe two new metrics to evaluate the reliability of the system, and propose an on-line optimizer algorithm that can Minimize the Data Missing Rate (MinDmr) in order to maximize the data availability. Based on MinDmr, we develop four optimizers associated with four different file access prediction functions. Simulation results utilizing the OptorSim show our MinDmr strategies achieve better performance overall than other strategies in terms of the goal of data availability using the two new metrics.