A Dynamic Programming Based Replica Placement Algorithm in Data Grid

Data replication is one of the well-known techniques to reduce the cost of data access and network bandwidth consumption, as well as to improve data availability in data grid The challenge in replica management is to select a set of suitable nodes for replicas and it is known as replica placement problem In this paper, we address replica placement problem in data grid under given traffic pattern and propose a replica placement algorithm based on dynamic programming It can help us find an optimal node set for replicas subject to workload constraint of all replicas while total communications cost in data grid is minimized Numerical example and experiments show that our algorithm is feasible and effective for the replica placement problem