基于蝙蝠算法的SDN多控制器部署

对于大型SDN网络,多控制器的部署和应用需求迫切。提出了一种基于蝙蝠算法的多控制器部署方法,同时优化了3个指标:最小化平均控制时延、最小化控制器负载差异度和去除孤立节点。通过在迭代时不断优化达到平均控制时延最小化;限制控制器负载利用率保证控制器间负载均衡,利用标签传递算法去除孤立节点保证域内通信。仿真结果表明该方法可以保证SDN网络在无孤立节点的情况下,获得最小时延以及负载均衡的多控制器部署方案。

A bat inspired controller placement algorithm in software defined network

For large SDN networks, the placement and application of multiple controllers is in urgent need. A multi-controller placement approach based on bat algorithm was proposed, and three indexes were optimized, which minimized the average control delay and the controller''s load difference, and remove the isolated nodes. The minimization of average control delay was realized when the average control was continuously optimized during one iteration, and the load balance was guaranteed by limiting the load utilization of the controller. At the same time, the tag delivery algorithm was used to remove the isolated nodes to ensure the intra-domain communication. The simulation results show that this approach can guarantee that SDN network achieves small delay and load balanced multi-controller placement without isolated nodes.