A novel replication scheme based on prediction technology in virtual P2P storage platform. (2019)

• Record Type: Journal Article
• Authors: Xiao, Peng; Zhang, Tienan
• Is Part Of: International journal of networking and virtual organisations. Volume 20:Number 1(2019)
• Page Start: 90
• Page End: 103
• Publication Date: 2019
• Subjects: peer-to-peer -- P2P -- data availability -- quality of service -- QoS -- replication scheme; Computer networks -- Periodicals; Information networks -- Periodicals; Virtual corporations -- Periodicals; Electronic commerce -- Periodicals; Management information systems -- Periodicals; Management -- Data processing -- Periodicals; 004.65
• Languages: English
• ISSNs: 1470-9503

Abstract

Recently, peer-to-peer (P2P) platforms have emerged as promising virtual storage platforms in many areas. However, most of P2P storage platforms are facing the challenging that how to maintain the data availability in such a volunteer-participating environments. In this paper, we present a novel technique which can figure out the probability of peer's failure during a given period. Based on this technique, we are enabled to evaluate the failure probability of any groups of peers so as to estimate the optimal number of replicas that can achieve better trade-offs between performance and data availability in P2P platforms. In this way, the storage platform can improve the resource utilisation without losing data availability. Extensive experiments in a real-world P2P platform indicate that the proposed replication scheme is effective to improving the data availability. In addition, it also exhibits better adaptive when the P2P platform is working in volunteering computing paradigm.