基于改进ACO算法的SDN网络路由规划与均衡负载研究

Research on Route Planning and Load Balancing of SDN Networks Based on Improved ACO Algorithm

针对现有SDN网络路由规划算法存在的丢包率高,吞吐量过小等问题,设计了一种基于改进ACO算法的SDN网络路由规划方案。SDN网络在整体结构上包含了数据层、控制层和应用层,并通过执行OpenFlow协议进行节点之间的通信。ACO算法模拟蚂蚁个体觅食过程中遗留的信息素浓度,判断和选择最合理的路径,但该种方案容易陷入局部最优解且迭代效率过低;为此,利用POA算法对经典ACO的寻优过程优化,提升其全局能力;在通信链路的均衡负载层面,通过评估改进ACO算法的链路流量方差,同时调整目标链路的负载率,以达到均衡SDN网络链路负载的目标。仿真结果显示,改进ACO算法可以在全局范围内优选出最优路径,且在最大数据流条件下丢包率仅为1.28%。

Aiming at the problems such as high packet loss rate and low throughput of existing SDN routing algorithms,an improved ACO algorithm based on SDN routing planning scheme is designed.The SDN network contains three layers of structure,including data layer,control layer and application layer,and implements the communication between nodes through the implementation of OpenFlow protocol.ACO algorithm simulates the remaining pheromone concentration of individual ant in the process of foraging,judges and selects the most reasonable path,but this scheme is easy to fall into the local optimal solution and the iteration efficiency is too low.Therefore,POA algorithm is used to optimize the optimization process of classical ACO to improve its global capability.At the balanced load level of communication links,the link traffic variance of the improved ACO algorithm is evaluated and the load ratio of the target link is adjusted to achieve the goal of balancing the link load of SDN network.Simulation results show that the improved ACO algorithm can optimize the optimal path in the global scope,and the packet loss rate is only 1.28%under the condition of maximum data flow.