RRSD: A file replication method for ensuring data reliability and reducing storage consumption in a dynamic Cloud-P2P environment

Abstract

Abstract Data reliability, storage consumption and load balance have been widely concerned for current dynamic cloud storage. The traditional file replication methods can obtain better load balance and high data reliability via multi-replica, but leading to huge storage consumption. Although these methods reduce storage consumption by dynamically removing redundant replicas, data reliability is not ensured enough. To deal with the problem, this paper proposes a file replication method for ensuring data reliability and reducing storage consumption in a dynamic Cloud-P2P (RRSD), aiming to minimize the number of replicas to obtain better load balance while meeting data reliability requirements. RRSD uses the method of ”multiple times replica placement” and ”redundant replica deletion” to achieve the goal. It adopts the centralized manner to create minimal replicas that can meet the data reliability requirement according to file’s storage expectation to reduce storage consumption when a file stores on cloud. To respond to the time-varying dynamic cloud in time, this approach uses the manner of decentralized self-adaptive to dynamically create fewer replicas, and select optimal node as placement nodes to improve load balance. Meanwhile, RRSD uses a method of periodicity detection to ensure data reliability. In addition, it uses verification evaluation method to selectively remove redundant replicas to further reduce storage consumption. Extensive experiments demonstrate that RRSD has superior performance regarding load balance, data reliability and storage consumption and can deliver an improvement of 10% for load balance and reduce storage consumption by 60% while meeting data reliability requirement compared with other similar methods.