DCN中基于前馈神经网络的动态多路径负载均衡方法

针对数据中心网络(DCN)中因大象流而引起的网络负载不均衡问题,提出一种基于前馈神经网络的动态多路径负载均衡方法。在拓扑感知和流量信息监控的基础上对大象流进行标记,将收集到的网络流量信息输入前馈神经网络以预估每段链路的负载,并结合优化蚁群算法为大象流寻找最优路径,使大象流根据链路的实时状态完成路径选择。仿真结果表明,该方法能够有效降低网络传输时延,提高链路利用率和网络吞吐量。

Dynamic Multi-Path Load Balancing Method Based on Feedforward Neural Network in DCN

In order to solve the problem of unbalanced network load caused by elephant flows in Data Center Network(DCN), a dynamic multi-path load balancing method based on feedforward neural network(FNN) is proposed.In this method, topological perception and flow information monitoring are carried out first, and the elephant flows are marked.The collected network traffic information is then used as inputs to estimate the load of each link through the FNN.Finally, the optimal paths are found for the elephant flows by combining with the optimized ant colony algorithm, so that the elephant flows complete the paths selection according to the real-time state of the links.Simulation results show that the proposed method can effectively reduce network transmission delay and improve link utilization and network throughput.