一种优化的互联网用户流量定向调度  

A Targeted Scheduling Method for Internet User Traffic

作　　者：王宏杰 杨波 徐胜超 毛明扬 蒋金陵 WANG Hongjie;YANG Bo;XU Shengchao;MAO Minyang;JIANG Jinling(School of Data Science,Guangzhou Huashang College,Guangzhou 511300,China)

机构地区：[1]广州华商学院数据科学学院,广东广州511300

出　　处：《云南师范大学学报（自然科学版）》2023年第6期39-42,共4页Journal of Yunnan Normal University:Natural Sciences Edition

基　　金：国家自然科学基金资助项目(61772221);广州华商学院校内导师制科研资助项目(2023HSDS06)。

摘　　要：利用软件定义网络架构,建立互联网用户流量分析预测—在线调度机制,通过互联网用户流量分析预测、异常流量识别以及在线调度等步骤,实现互联网用户流量定向调度;同时,利用SRv6服务链技术优化互联网用户流量的路由选择机制,从而有效避免链路中出现流量堵塞,并采用基于半监督学习的流量聚类方法进行互联网用户流量聚类,以缩短流量调度时间.实验结果表明,该方法的最大链路利用率低、吞吐率高、调度时间短,能够使调度后的网络负载更均衡,提高节点活性和调度效率.Using software defined network architecture,an online scheduling mechanism for analyzing and predicting internet user traffic was established.Through steps such as analyzing and predicting internet user traffic,identifying abnormal traffic,and online scheduling,targeted scheduling of internet user traffic was achieved.At the same time,the routing mechanism of internet user traffic was optimized by utilizing SRv6 service chain technology to effectively avoid traffic congestion in the link,and a semi supervised learning based traffic clustering method for internet user traffic clustering was adopted to reduce traffic scheduling time.The experimental results show that the maximum link utilization rate of this method is low,the throughput rate is high,and the scheduling time is short.It can make the network load of internet user traffic scheduling more balanced,improve node activity and scheduling efficiency.

关 键 词：互联网用户流量 软件定义网络 SRv6服务链技术 流量聚类 

分 类 号：TP393.4[自动化与计算机技术—计算机应用技术]