您现在的位置:首页 > 学术研究 > 论文发表 > Controller Placements for Optimizing Switch-to-Controller and Inter-controller Communication Latency in Software Defined Networks
Controller Placements for Optimizing Switch-to-Controller and Inter-controller Communication Latency in Software Defined Networks
[发布时间:2021-06-27  阅读次数: 1416]

作者:Bo Wei, Yuqi Fan, Lunfei Wang, Tao Ouyang,  Lei Shi

发表刊物:WASA 2021

年份:June 2021

摘要:The logically centralized control plane in large-scale SDN networks consists of multiple controllers, and the controllers communicate with each other to keep a consistent view of the network status. Inconsistent controllers or controllers failing to maintain the network state in time may severely degrade the network performance. However, most of the existing research focuses on the switch-to-controller communication latency by ignoring the communication latency between the controllers. In this paper, we formulate a novel multi-objective SDN controller placement problem with the objectives to minimize both the switch-to-controller and the inter-controller communication latency. We propose an efficient Multi-Objective Controller Placement (MOCP) algorithm. Algorithm MOCP generates the new controller placement solutions with crossover and mutation operations. The switches are assigned to the controllers with the greedy strategy initially, and then mapped to the other controllers via the local search strategy. Our simulation results show that algorithm MOCP can effectively reduce the latency between the controllers and from the switches to controllers simultaneously.

参考文献拷贝字段:Bo Wei, Yuqi Fan, Lunfei Wang, Tao Ouyang, Lei Shi. Controller Placements for Optimizing Switch-to-Controller and Inter-controller Communication Latency in Software Defined Networks [C]. The 16th International Conference on Wireless Algorithms, Systems, and Applications (WASA), Nanjing, China, June 25-27, 2021: 535-546


相关下载:
    Controller Placements for Optimizing Switch-to-Controller and Inter-controller Communication Latency in Software Defined Networks