Frequent block access pattern-based replication algorithm for cloud storage systems

Abstract

Replication is a strategy in which multiple copies of same data are stored at multiple sites. Replication has been used in cloud storage systems as a way to increase data availability, reliability and fault tolerance and to increase the performance. Number of replicas to be created for a file and placement of the replicas are the two important factors which determine the storage requirement and performance of the cloud storage systems. In this paper, we have proposed a novel replication algorithm for cloud storage systems which decides replication factor for a file block based on frequent block access pattern and placement of that file block based on local support value. We have carried out preliminary analysis of the algorithms and the results indicate that the proposed algorithm can perform better than the replication algorithm followed in Hadoop storage system.