摘要
针对软件定义网络(Software Definition Network,SDN)多控制器负载不均衡问题,提出基于Q-learning算法区分交换机等级的SDN控制器负载均衡算法。依据交换机所影响网络资源划分等级,在进行控制器负载均衡的同时避免高等级交换机连接至同一控制器,提升全网络生存性。仿真结果表明:该算法可以有效地实现控制器负载均衡功能,同时可以有效根据控制器等级进行迁移分配,限制了单个控制器故障对全网资源的影响,提升了网络生存性。
Aiming at the load imbalance problem of SDN multi-controller,we propose a load balancing algorithm for SDN controllers based on Q-learning algorithm to distinguish switch levels.According to the classification of network resources affected by switches,it balanced the load of controllers,and avoided the connection of high-level switches to the same controller,so as to improve the survivability of the whole network.The simulation results show that our algorithm can effectively achieve the load balancing function of the controller.And it can effectively migrate and distribute according to the level of the controller,which limits the impact of a single controller fault on the whole network resources and improves the network survivability.
作者
蒋雯丽
梁思远
赵芳利
赵峰
王聪一
Jiang Wenli;Liang Siyuan;Zhao Fangli;Zhao Feng;Wang Congyi(College of Communication and Information Engineering,Xi’an University of Post and Telecommunications,Xi’an 710121,Shaanxi,China)
出处
《计算机应用与软件》
北大核心
2020年第6期142-147,153,共7页
Computer Applications and Software
基金
国家自然科学基金项目(61875164)
陕西省教育厅2017年专项科学研究计划项目(17JK0702)。