C. Choi and S. Kim (Korea)
Grid computing, Shared storage cluster system, Resource management, Data reorganization
Because the data reorganization is usually a time consuming service, efficient data reorganization is a key factor of the performance of shared storage cluster system. This paper proposes adaptive data reorganization classifying data reorganization into full/partial data reorganization for the performance of shared storage cluster system. The performance results are compared with those of deferred data reorganization using a Markov model. In higher workload intensities, our method provides better average rate of servicing operations than that of deferred data reorganization. Furthermore, adaptive data reorganization can be a viable local resource management of shared storage cluster system for Grid computing.
Important Links:
Go Back
