Research On Continuous Data Protection System Based On Data Partitioning Strategy

With the advent of the "cloud era", more and more users upload data to the cloud,when cloud service providers provide services, they need handle users request at any time.But when the disaster happens, the traditional disaster recovery system cannot providemore granular RTO(Recovery Time in Objective) and RPO(Recovery Point in Objective),Therefore, CDP(Continuous Data Protection) technology for cloud services can solve thisproblem.Bring up a kind of CDP service architecture based on data partitioning which cansolve the existing continuous data protection system based on the group to deposit a copyof the policy brought a copy of migration as well as reliability issues. The system uses the"server+members of the group "storage architecture. Servers save the user’s onecomplete copy and the group saves multiple copies. The system applies data partitioningstrategy based on the time point and fix-length. Copy is divided into a number ofsub-blocks stored in the crew group. Multiple-copy enhances the reliability of the system,at the same time data chunked avoid the copy of the original system migration networkoverhead and improve the utilization of storage space in the group. Heterogeneous groupmembers in the group priority division, high-priority group members get prioritysub-block data, and priority service scheduling strategy, shorten the user synchronizationtime, and improve system performance.This paper presents the architecture and partitioning strategies were tested, the testresults show that this paper, as compared to the original system has greatly improved thesynchronization time and service efficiency.