A novel replication method in data grid

Data grid aims to provide services for sharing and managing large data files around the world. The challenging problem in data grid is how to reduce network traffic. One common method to tackle network traffic is to replicate files at different sites and select the best replica to reduce access latency. Since computing exact file transfer time is time consuming, so other methods such as prediction can be used for estimating file transfer time. In this paper, a cosine similarity predictor function is proposed for replication that uses path similarity between the previous and current file transfer paths. Based on this predictor function, a new replication algorithm is proposed. Experimental results with Optorsim show the proposed replication algorithm outperforms comparing to multi regression and neural network methods as number of jobs increases.