Affiliations: [a] Department of Computer Science and Applications, the Hashemite University, Zarqa, Jordan | [b] Department of Software Engineering, the Hashemite University, Zarqa, Jordan
Correspondence:
[*]
Corresponding author: Mohammad Bsoul, Department of Computer Science and Applications, the Hashemite University, Zarqa, Jordan. E-mail: [email protected].
Abstract: In a data grid, nodes have a limited storage space and cannot accommodate all the replicas they need. Therefore, data replication is used to copy only important replicas that may be requested in the near future. In this paper, a dynamic data replication strategy is implemented in order to increase the number of requests served locally and reduce both the total elapsed time and total bandwidth consumption. This strategy is based on Bandwidth Hierarchy based Replication (BHR) and modified BHR strategies, but superior to them. This is because of the features this strategy has. This strategy selects a region header in each region based on a modified version of closeness centrality and stores all the replicas in the data grid in the region header. Regarding the other nodes in the region, only important replicas will be copied to them based on their number of requests per mega bit (NoRpmb) that shows how much important the replica is in relation to its size.
Keywords: Data grid, replication strategy, bandwidth hierarchy based replication, simulation
