1.西南石油大学 计算机科学学院;2.四川大学 计算机学院
国家自然科学基金资助项目(61373091);四川省南充市科技资助项目(19SXHZ0012)。
1.School of Computer Science,Southwest Petroleum University;2.College of Computer Science,Sichuan University
针对分布式SDN网络中流量管理调度不均衡的流量工程问题,提出一种基于负载均衡的多控制域流量路由优化的解决方案。首先分析控制消息流量的组成、域内通信与域间通信规则；然后基于四种控制消息定义控制链路流量的构成,明确链路承载流量分为控制消息流量和业务流量,建立平衡控制器负载和最小化最大链路利用率的优化模型；再次基于域内通信和域间通信提出两层路由算法(TLR),为提高模型求解精度,再提出改进离散萤火虫算法求解最优路由。结合ABILENE网络和GEANT网络,分析控制消息流量、控制器负载、链路负载等评价指标。实验结果表明,优化模型能有效实现控制器和链路负载均衡,控制消息流量是流量工程重要组成部分。
Aiming at the traffic engineering problem of unbalanced traffic management and scheduling in distributed SDN, we propose a solution of traffic routing optimization in multi-control domain based on load balancing. Firstly, we analyze the composition of message traffic, intra-domain communication and inter-domain communication rules. Then, we define the composition of control link traffic by four kinds of control messages, and divide the link traffic into control message traffic and service traffic, and establish an optimization model to balance the controller load and minimize the maximum link utilization. Secondly, we propose a two-layer routing algorithm (TLR) based on intra-domain communication and inter-domain communication. In order to improve the accuracy of the model solution, we propose an improved discrete firefly algorithm to solve the optimal routing. Combined with ABILENE network and GEANT network, we analyze the model by the control message traffic, controller load, link load and other indicators The experimental results show that the optimization model can effectively balance the load between controllers and links, and controlling the message traffic is an important part of traffic engineering.
