Design And Implementation Of Off-site Disaster Recovery Mechanism For Data Center Networks

As an infrastructure for big data,cloud computing and other applications,the data center networks are vulnerable to the destruction of natural disasters,causing massive data loss and incalculable losses.Therefore,the off-site disaster recovery mechanism of data center networks has been widely studied and applied.The off-site disaster recovery mechanism means that the data center that is about to be damaged will back up the stored data to a remote,secure data center from the warning time of the detection to the disaster to the disaster.Based on the in-depth study of data center network and disaster recovery mechanism,this paper fully considers the characteristics of data flow and backup nodes,and designs and implements an off-site disaster recovery mechanism for data center networks.The specific work is as follows:First of all,this paper designs the off-site disaster recovery mechanism of data center network in the early warning time,which mainly includes two parts.The first part is to design the optimal matching mechanism for the data to be backed up and backup nodes by linear programming,and determine the optimal backup node for the data based on the amount of data to be backed up,link state and storage capacity of the backup node,thereby improving the utilization of links and backup nodes.The second part is to design an efficient and reliable end-to-end transmission mechanism.The transmission mecha-nism defines the privileged functions of each flow according to the data and backup node priorities,and adopts the ECN mechanism to reflect the degree of network congestion.The sender performs different degrees of congestion control on the flows according to the privilege function and network congestion.The switch determines the order of data loss and transmission according to the data and backup node priorities,and preferentially al-locates bandwidth for the flow which destination node and data priorities are high,reduce the average transmission time so that the data in the affected data center can be transmitted to the secure data center as much as possible during the warning time.Secondly,this paper establishes the solution module based on the simplex method to determine the upper limit of the backup data volume of the backup node,and then writes the solution matching backup module to determine the backup node's specific data through dynamic programming,the best matching of the backup data and the backup nodes is realized by these two modules.Then this paper implements the transmission mechanism through NS2 simulation,so as to finally realize the disaster recovery mecha-nism in the data center network.Finally,this paper is divided into two parts to verify the off-site disaster recovery mechanism.The first section verifies that the matching of data to be backed up and backup node can determine different backup nodes and the data that nodes back up according to the constraints and objective functions at different warning times.The second part verifies that the transmission mechanism can preempt bandwidth for data with high priority of destination nodes and data,and the performance of the transmission mechanism is better than that of TCP,D2TCP,and DCTCP under the condition that the warning time and the amount of data to be backed up are equal respectively.This article combines the contents of these two parts and proves that the off-site disaster recovery mechanism can back up more data in a shorter warning time,thereby minimizing losses.