Dynamic replication algorithm in data grid: Survey

Data Grid is an infrastructure that manages huge amount of data files, and provides intensive computational resources across geographically distributed collaboration. It is not enough to provide convenient accessibility to these data by only high speed network and large mainframe systems. For improving the performance of file accesses and to ease the sharing amongst distributed collaboration, such a system needs replication services. Data replication is a common method used to improve the performance of data access in distributed systems. In this paper, we present a survey of some related previous works and highlight some various algorithms that have been proposed by other researchers. A dynamic replication model based on mathematical concepts is proposed. The main purpose of this model is find out the popular file using the concept of exponential decay/growth. We estimate the next number of access for the file