A Distributed Multi-Objective Replication Strategy For Massive Data Disaster Recovery

Increasing number of industries such as IT industries,electricity and energy,transportation and logistics are utilizing information systems and with the help of which,discover the value of importance in the data.Data Protection is a nominant level for key resources.The expanding volume and velocity of big data makes it expensive and lasting to backup and recover data in the case of data protection.In addition,the need of data backup and protection is no longer one of basic requirements of enterprises,but medium and small businesses as well,ensuring data and services can persist in the condition of data disaster recovery infrastructure.The main problems that exist are: 1.The big volume of data costs a lot to store,and a long time to recover.2.There now exists few studies focused on the replication strategy for data disaster recovery.3.Replication placement needs an efficient and fast strategy,using the computing ability of distributed system.The work of this thesis are:1.Studies on the replication technology for achieving high reliability and high availability and studies on the optimizaiton algorithm for replication creation,number adjustment,selection and placement,and also the technologies to balance the system's workload.2.A Data replication strategy for disaster recovery(DR2S)is proposed,which is based on discrete Multi-Objective Particle Swarm Optimization and improves the management cost of backup data and access time in case of data recovery.3.A realization of DR2 S in the Map Reduce parallel frame is given,speeding up the execution of DR2S4.Design of a module of Backup Data Management on Distributed Disaster Recovery Management Platform,in which the DR2 S is imbedded and utilized on it.