Virtual Machine Migration Stradtegy Research In Cloud Data Center Based On SDN

In recent years, with the rapid development of cloud computing technology, live migration of the virtual machine as important technology have been widely used and developed in cloud data center, which enables flexibility in data center load balance, disaster tolerance, system maintenance ,resource optimization and etc. Despite its usefulness, the live migration still introduces performance degradation over across WAN,which can not meet the demands of enterprises and users very well, so how to improve the migration performance over WAN has become a hot spot in the field of cloud computing.Software Defined Networks(SDN), as a new Network architecture,decouples the control plane from the network device, so as to realize the centralized control of the network. Network virtualization? logic centralized control and network open interface can well fit for the centralized network management, flexible network path forward, the demand of VM deployment and migration in data center, So it will have great advantage for introducing SDN to cloud data center. In this paper,based on SDN cloud data center architecture, from the perspective of network resource optimization we improve the network bandwidth for online migration to optimize the performance of migration.In this paper, we propose based on dynamic routing and bandwidth-provisioning strategy(DRBP), which is dynamic and self-adaptive in SDN network architecture, firstly we discuss the relative factors affecting the performance of migration according to the mathematical model of online migration, and propose an optimal bandwidth allocation strategy, the overall time?downtime?migration resource consumption are taken into account in our strategy. In addition, dynamic routing mechanism is adopt to make routing rules and meet the demand of the bandwidth of path for migration.Through this paper, introducing SDN in the cloud datacenter network,DRBP strategy can improve the network transmission bandwidth for live migration across data center, and reduce the migration time, cut down the consumption of network resources and improve the performance of migration.