端到端时延下网络多路径负载均衡算法仿真  

Simulation of Network Multipath Load Balancing Algorithm with End to End Delay

作　　者：苏富林 冯桂莲[2] SU Fu-lin;FENG Gui-lian(Department of Computer Science,Gansu Normal College for Nationalities,Hezuo Gansu 747000,China;School of Physics and Electronic Information Engineering,Qinghai Nationalities University,Xining Qinghai 810007,China)

机构地区：[1]甘肃民族师范学院计算机科学系,甘肃合作747000 [2]青海民族大学物理与电子信息工程学院,青海西宁810007

出　　处：《计算机仿真》2023年第2期429-433,共5页Computer Simulation

基　　金：国家民委2021年度高等教育教学改革研究项目(21102)。

摘　　要：为了提高网络的通讯性能,降低网络能量消耗,提出基于蚁群优化的网络多路径负载均衡算法。通过路由模块确定转发路径,在大象流检测模块中检测网络中存在的大象流,构建大象流调度目标函数,利用网络监听模块监听各节点在网络中的实时信息。根据获取的信息采用蚁群优化算法调度网络中存在的大象流,通过路由模块、大象流检测模块、网络监听模块和计算决策模块构成网络多路径负载均衡机制,完成网络多路径的负载均衡。仿真结果表明,所提算法的端到端时延短、网络能量消耗小、分组丢失率低和网络生命周期长。In order to improve the communication performance of the network and reduce energy consumption,a multi-path load balancing algorithm for the network based on ant colony optimization was proposed.Firstly,the forwarding path was determined by a routing module.And then,the elephant flow in the network was detected by the elephant flow module.Meanwhile,the elephant flow scheduling objective function was constructed,and the network sniffing module was used to monitor the real-time information of each node in the network.According to the information,the ant colony optimization algorithm was adopted to schedule the elephant flow existing in the network.Finally,the multi-path load balancing mechanism was formed through the routing module,elephant flow detection module,network sniffing module and calculation decision module.Thus,the multi-path load balancing of the network was completed.Simulation results show that the proposed algorithm has short end-to-end delay,low network energy consumption,low packet loss rate and long network life cycle.

关 键 词：蚁群优化算法 大象流检测 网络多路径传输 网络监听 负载均衡 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]