基于群集运动智能的SDN业务传输优化技术  

SDN service transmission optimization technology based on swarming motility intelligence

作　　者：周宁 张建辉[1] 胡宇翔[1] 刘媛妮[2] 汪斌强[1] ZHOU Ning;ZHANG Jianhui;HU Yuxiang;LIU Yuanni;WANG Binqiang(National Digital Switching System Engineering and Technology Research Center,Zhengzhou 450002, P.R. China;Future Network Research Center,Chongqing University of Posts and Telecommunications,Chongqing 400065,P.R. China)

机构地区：[1]国家数字交换系统工程技术研究中心,郑州450002 [2]重庆邮电大学未来网络研究中心,重庆400065

出　　处：《重庆邮电大学学报（自然科学版）》2019年第3期323-330,共8页Journal of Chongqing University of Posts and Telecommunications(Natural Science Edition)

基　　金：国家重点研发计划课题(2017YFB0803204);国家自然科学基金(61521003,61502530,61802429,61872382);广东省重点领域研发计划(2018B010113001)~~

摘　　要：针对软件定义网络(software defined network,SDN)负载全局优化问题,受生物界群集运动的启发,以实现当前分布式全网状态感知、全局优化目标、区域自主协同为目标,研究了基于群集运动智能的SDN业务传输优化技术。为满足上层业务对网络传输性能的需求,根据传输时延、网络吞吐量等性能指标,在多目标优化约束条件下,通过业务聚类定义针对各业务的平稳流,并将平稳流内共同协作的节点相结合定义为平稳流协作场。协作场内的节点通过收集网络状态信息并在协作场内共享,实现网络状态的快速感知及决策。协作场之间各个平稳流以不同的带宽、时延等网络传输需求运行,使得平稳流获得了确定的资源抢占成功率,从而以稳定速率传输。针对基于平稳流的业务传输优化技术,给出了基于群集运动协同方式的流量工程处理体系和业务优化方案,并对该体系下群智节点的感知认知、业务聚类分析、平稳流协作场构建、协同智能决策优化方法、优化流程、节点架构等进行分析阐述。Aiming at the problem of SDN load global optimization,inspired by the swarm movement of the biological community to realize the network state awareness,global optimization target and regional autonomous coordination of the current distributed network,the SDN service transmission optimization technology based on swarm motility intelligence is studied.In order to guarantee the requirements of upper layer services to the network transmission performance,according to multiple performance indicators such as transmission delay and network throughput,the stationary flow for each service is defined by under the multi-objective optimization constraint,and the in same stationary flow is defined as a stationary flow collaboration field. Each node in the collaboration field can collect network state information and share it with other nodes,which can accelerate the network state convergence and decision. Each stationary flow works under different network transmission performance constraints such as delay,bandwidth,which can provide a steady transmission rate for an upper-layer service.Concerned to the stationary flow-based service transmission optimization technology,the traffic engineering processing system and service optimization scheme based on the swarming-coordination motility are proposed,and the crowd sensing and cognition,service swarming,collaborative intelligent decision optimization method,optimization process,and node architecture are also discussed.

关 键 词：软件定义网络 业务聚类 平稳流 传输优化 

分 类 号：TP393.02[自动化与计算机技术—计算机应用技术]