Research On Redundant Data Storage Based On DHT

With the development of network and application, the information explosioncontributes to the new scale of Internet data, which increases the storage pressure oflarge-scale data-centers greatly. DHT-based Peer-to-Peer storage system has receivedconsiderable research interests with its high scalability and fast data-access capabilities.However, DHT-based data storage is dynamic and heterogeneous, which causes datamay be lost at any time. Therefore, the high availability is one of key issues forDHT-based storage system.This paper mainly focuses on the redundant storage strategy and erasure codebased on Cauchy matrix. Combining Erasure Coding and the single-replication, a hybrid1+Nredundant storage strategy is given. It may be briefly summed up as follows:1. The design of Distributed Hash Table and several common DHT protocols areilluminated, the two traditional redundant storage strategies of Complete Replicationand Erasure code are explained in detail, and the usage of RS code in DHT storagesystem is studied.2. RS erasure code based on Cauchy matrix is studied in detail. A constructionalgorithm of improved Cauchy matrix is given by using the idea of linear transformation.Experiment results show that the RS code based on the improved Cauchy matrix canmake the speed of encoding and decoding process increased10%.3. A new hybrid data redundant strategy is given, which can increase thesystem availability with the same redundancy efficiently by combining the Erasurecoding and single-replication. The new strategy presented in this paper can improve theaccess efficiency and reduce the number of nodes accessed in routing as well as thenetwork maintenance costs. The results of experiment indicate the effectiveness of theproposed redundant data storage strategies.