A Dynamic Optimal Replication Strategy in Data Grid Environment

Abstract

The data grid technology, which uses the scale of the Internet to provide storage capability for the mass data, has become one of the hot research topics. The high latency of the network is the bottleneck in accessing the files in data grid systems. The replication strategy can shorten the time of fetching the files by creating many replicas stored in appropriate locations. Due to the limited storage capacity of each node, it is important to have an efficient replication strategy to make sure of the overall performance. In order to solve the problem, we propose a dynamic optimal replication strategy (DORS) in this paper. First, we empirically infer a threshold from some former work to decide whether one file is needed to be copied to other nodes. Then, a measure is devised to get the importance of each file, which relies on the inner and external features of the file. Finally, the algorithm can decide which replica to replace based on the importance value. Results from the simulation show the better performance of our algorithm than other former work.