Replica Deletion Strategy Based on Gray Prediction Theory and Cost in P2P Network

Replication can improve reliability and availability of files by creating lots of replicas for files. But an overly large number of replicas spend a lot of storage resources, and have an adverse effect on network performance. In order to save and make the best use of storage resource, this paper presents a replica deletion strategy based on gray prediction theory and cost, to delete the useless replicas. Using gray prediction model for reference, the use value of replica, also called replica activity, is forecasted. Then the recovery cost of replica is analyzed and quantified. The survival value of replica is gained by integrating the gray prediction results, the recovery cost and other key factors, which decides to save or delete replicas. The experimental results show the gray prediction model can improve accuracy of prediction and prediction error lowers 48%. In addition, more comprehensive consideration when replica deletion cannot make it blindly, and rate of decay of replica number lowers 26%, which can reduce thrashing.