基于FRFT的网络流量异常数据快速捕获方法  

Fast Capture Method of Network Traffic Abnormal Data Based on FRFT

作　　者：钟妮 王剑[2] ZHONG Ni;WANG Jian(Sichuan Institute of Industrial Technology,Mianzhu Sichuan 618200,China;Wuhan University of Technology,Wuhan Hubei 430070,China)

机构地区：[1]四川工业科技学院,四川绵竹618200 [2]武汉理工大学,湖北武汉430070

出　　处：《计算机仿真》2023年第4期413-416,443,共5页Computer Simulation

基　　金：2020年高校优秀青年人才支持计划项目(gxyq2020267)。

摘　　要：为了降低异常数据捕获耗时,准确捕获各类攻击形式,保证网络正常运行,以傅里叶变换为构建基础,提出分数阶傅里叶变换下的网络流量异常数据快速捕获手段。结合经典傅里叶变换与角度参数,得到分数阶傅里叶变换及其相关参数,根据网络流量自相似性与异常数据间的相关性,获取变换域序列,利用能量谱参数估计法联立序列能量谱与赫斯特参数间关系。将赫斯特参数作为自相似阶段,依据离散时间序列与正常自相似范围,明确赫斯特参数的自适应阈值范围,基于此实现网络流量异常数据捕获。仿真阶段,选取企业网VLAN1、VLAN2,采用开源网络攻击模拟工具,测试所提方法在不同攻击形式下的异常数据捕获效果,结果表明所提方法具有良好的捕获精准度,且能够满足实际应用的实时性需求。In order to reduce the time consumption of abnormal data capture,accurately capture various attack forms,and ensure the normal operation of the network,based on the Fourier transform,a rapid capture method of network traffic abnormal data under the fractional Fourier transform is proposed.Combine classical Fourier transform and angle parameters to obtain fractional Fourier transform and its related parameters.According to the correlation between network trafic self-similarity and abnormal data,the transform domain sequence is obtained,and the energy spectrum parameter estimation method is used to combine the relationship between the sequence energy spectrum and the Hurst parameter.Taking the Hurst parameter as the self-similar stage,based on the discrete time series and the normal self-similar range,the adaptive threshold range of the Hurst parameter is clarified,and the network traffic abnormal data capture is realized based on this.In the simulation phase,the enterprise network VLAN 1 and VLAN 2 are selected,and the open source network attack simulation tool is used to test the abnormal data capture effect of the proposed method under dfferent ttack forms.The results show that the proposed method has good capture accuracy,and can meet the real-time requirements of practical application.

关 键 词：分数阶傅里叶变换 傅里叶变换 网络流量 快速捕获 异常数据 赫斯特参数 

分 类 号：TP399[自动化与计算机技术—计算机应用技术]