Cloud Storage Environments Replica Selection Strategy Of Maintaining Consistency

The rapid development of information technology leads to the explosive growth of dataaround the world. Cloud storage can solve the problem of mass data storage and managementin an effective way, how to improve the reliability and performance of cloud storage systembecomes the focus of current research. Although the replica technology effectively meets thedevelopment of the cloud storage system, at the same time, it has brought the problem ofreplication management.How to establish an appropriate replication management mechanism from multiple datastorage and select the best one to improve access speed of data resources in maximum fromthe multiple replicas has become the key problems of replica management.In the choice of the strategy of replica selection, on the basis of comprehensiveconsideration of environmental information and user preferences, the dissertation put forwardthe replica selection strategy based on the environment similarity and the user’s QoSsatisfaction. The algorithm’s main goal is to choose a replica of the highest overall credibilityaccording to the environmental information, user’s QoS satisfaction and the history evaluationinformation of storage nodes. Simulation experiment results showed that the strategy hadgood validity and timeliness.Under the choice of the strategy of the maintenance of replication consistency,considering the advantages and disadvantages of the strong consistency and the weakconsistency, the dissertation gave a definition about the ratio of heat and updating andtheoretically proves its relationship with the updated probability of replica. This paper makesfull consideration of the information, such as system load and the heat of the file access, combineswith the factors such as the number of the replica which has not been updated, adaptively to change theupdating priority and principle, puts forward the dynamic replication consistency maintenance algorithmbased on the information of heat and updating, Simulation experiment results showed that underthe circumstance of more replicas with heat access, the algorithm has better performance.