考虑控制器故障的软件定义机载网络选举算法

针对机载网络环境下控制器发生不可恢复故障的问题，提出分布式的控制节点选举（DCNE）算法. 为了保证故障恢复后控制节点的性能，综合考虑邻居节点密度、平均流请求量和控制器处理能力，计算选举权值. 通过邻居节点信息的建立和维护实现选举消息的更新，提出依据节点状态传递和处理选举消息的算法. 从故障恢复后的控制域性能角度对DCNE算法进行仿真验证. 结果表明，与现有算法对比，DCNE算法在最大时延和流部署开销指标上具有更好的性能，能够较好地适用于机载网络环境.

Software defined airborne network election algorithm considering controller failure

A distributed control node election (DCNE) algorithm was proposed aiming at the problem of unrecoverable failure of the controller in the airborne network environment. The election weight was calculated by considering the neighbor node density, the average flow request amount and the controller processing capability in order to ensure the performance of the control node after the fault recovery. Then the election message was updated through the establishment and maintenance of the neighbor node information. A method of transmitting and processing election messages according to the node status was proposed. The proposed algorithm was simulated and verified from the perspective of control domain performance after fault recovery. The experimental results show that the DCNE method has better performance in terms of maximum delay and flow deployment overhead compared with the existing methods. The DCNE method can be better applied to airborne networks.