High Availability Research And Design Of Distributed SDN Cluster Controller

Software Defined Networking(SDN),as the main research direction of the future network,is also becoming more and more popular.SDN can realize the rapid deployment of services through centralized control,and make the whole network resource utilization more reasonable.With the continuous increase of business scale and the complication of application scenario,the centralized control not only brings the advantage of controllability,security and rationality,but also it has many scalability,reliability and high usability.In this paper,a high availability architecture of distributed SDN cluster controller is proposed.The fault detection for each layer and the election algorithm of host controller based on the high availability are studied in this architecture.Firstly,this paper designs a high availability architecture of distributed SDN cluster controller,which includes root controller management system,root controller,local controller and distributed data management system.Secondly,basing on the characteristics of each layer,three different methods of fault detection are proposed.For the management system of root controller,the main and backup detection method is adopted.For the detection of the root controller failure,the method of "over half detection" is proposed.For fault detection of local controller,a "loop detection" method is proposed.The main experimental results show that the proposed method can detect the fault quickly and improve the detection accuracy.In addition,this paper proposes the algorithm of master controller election,and the election algorithm of master controller based on high availability.The results of experimental show that this method can effectively select high availability controller.