A Cloud Storage Model Based On P2P

Cloud storage is a concept arising with the development of cloud computing. It is a system to provide data storage service and outside access to business by allowing a lot of different types of network storage devices working together through applications software adopting the cluster, grid or distributed file system and other functions. Cloud storage focuses on providing Internet-based, highly reliable online storage services, which meets the large data storage requirements for the emerging web2.0applications.However, designing a high-performance cloud storage system faces many challenges, such as heterogeneity, distribution, scalability, transparency, concurrency, reliability, efficiency and so on. At present, some of the main cloud storage systems, such as Google’s GFS, Amazon’s Elastic Cloud, including the open source HDFS, have adopted a similar centralized architecture, where exists a central server to manage data storage nodes in the system and all the requests including storage and retrieval go through a central server. In this kind of system framework, there is a great risk of single point of failure (SPOF), and the central server can easily become the bottleneck of the system and could cause the entire cloud storage service not available once it crashes. Although systems in practice have used some technique as backup recovery to avoid the probable disaster, the problems have not been resolved essentially.Many characteristics of P2P such as decentralization, scalability, robustness, performance and load balancing meet the design requirements of cloud storage. The P2P technology can be used in the cloud storage system design, exploiting the advantages of P2P technology in the cloud storage system. In this paper, a new cloud storage Model based on P2P is proposed. This Model applies Chord arithmetic in managing service nodes and handing out clients’request, which solves the problems from the centralized structured architecture such as SPOF, performance bottleneck and so on, and realizes load balancing. Also, an optimization policy and an index model based on DHT are applied in the process of the query in the Chord network to improve the efficiency of query.