Game-Theoretic Strategy Analysis for Data Reliability Management in Cloud Storage Systems

Abstract

Replication is the simplest way to achieve data reliability in cloud storage systems. Nevertheless, replication incurs storage overhead to the cloud storage provider (CSP). To verify CSPs' reliability, users can audit CSPs with remote data integrity checking. However, the auditing incurs cost to users. Thus, CSPs and users involve a conflict situation, where users prefer less auditing and CSPs prefer less replication. In this paper, we provide a game-theoretic analysis to get optimal strategies for users and CSP. We use the pricing strategy adopted by Amazon S3 to explain our analysis. Our results show that a user should audit if CSP's reduced data copies are less than 1:81. If CSP believes lower user's staying probability, it should provide more discount or copies. According to this study, a user has the criterion for paying less auditing cost and CSP makes the reliability and pricing policy to keep users in business.