Reliability-based SDN Controller Deployment

In traditional networks,the control plane and forwarding plane of network devices are tightly coupled,but this also leads to many challenges in terms of manageability,flexibility,cost,and scalability.The tightness of traditional network equipment also limits the flexibility and scalability of the network.As the scale of network development continues to expand,the management capabilities of the network are gradually weakening,and the network operation and maintenance costs of operators are gradually increasing.Therefore,Software Defined Networking(SDN)emerges as a new type of network architecture.SDN separates the control logic from the data forwarding device and provides a unified programming interface for the upper layer application,which significantly improves the flexibility of the network,manageability and programmability have become the research hotspots in the future network field.As a technology that has emerged in recent years,SDN has not developed sufficiently well,and the controller plane that plays its central role has become particularly important.Traditional single controller management can easily lead to performance bottlenecks in the control plane,resulting in a single point of failure.At present,the main means to solve this problem is to deploy multiple controllers,but as the scale of the SDN network continues to expand,its reliability problem is still very serious.One is due to the reliability controller deployment problem caused by the separation of the SDN data plane and the control plane,and the other is that for the controller failure scenario,a network recovery scheme in which the switch is re-distributed quickly is required.In view of the above problems,this paper proposes a controller deployment scheme based on high reliability,and a switch neighbor mutual assistance mechanism for controller failure.The main contributions of the paper are as follows.(1)A multi-controller deployment model with reliability as the optimization goal was established for the performance requirements of large-scale network scenarios.The model firstly combines the need for reliable information exchange between the controller and the switch,considers the network QoS(quality of service),models the average flow rule delivery delay,controller load variance and reliability,and proposes that based on the high-reliability controller deployment model,the particle swarm optimization algorithm is combined with the simulated annealing algorithm to solve the optimal controller deployment scheme,and the high-reliability controller deployment that guarantees the smaller average flow rule delivery delay and load variance is realized.The experimental results show that for a given network topology,the reliability of the whole network obtained by the algorithm proposed in this paper is higher than the other two algorithms,and the optimization effect is better.(2)For the controller failure,the switch in the failed area is re-switched quickly to ensure stable operation of the network.The average delay and the controller load variance are modeled,and the neighbor mutual assistance mechanism is proposed.The redistribution scheme of the switch for a controller failure is obtained through the simulated annealing algorithm to ensure that the load is as balanced as possible after the switch is divided,and the delay is as small as possible.