A MOPSO-based Data Replication Strategy for Disaster Recovery

Data disaster recovery of information system is essential for guaranteeing the data availability in a variety of areas and industries. Traditional efforts are made to decrease the time cost for data transfer, while data management cost becomes contrarily high. In this paper, focused on optimizing the data recovery system's data management cost and data transfer time, a Data Replication Strategy for Data Recovery (DR2S) is proposed. DR2S can allocate data replicas to data nodes according to the bandwidth and storage cost. On the premise that guaranteeing data availability and data transfer time, it can further decrease data management cost and balances the data access load of the cluster. Experimental results proves that DR2S can provide a low-cost, high-performance data disaster recovery solution, compared to traditional methods, such as the rack-aware replica strategy.