摘要
弹性光网络流量传输受到时间波动导致异常,为了提高网络传输稳定性,提出基于孤立森林算法的弹性光网络异常流量自动识别算法。根据流量的异常分布特征和正常数据的差异性进行波谱密度检测,构建弹性光网络流量的谱特征提取模型,通过低通滤波器卷积向量重组,实现对异常流量的谱特征筛选,采用孤立森林算法实现对网络流量异常检测的自适应寻优控制,结合多维空间结构重组方法实现对弹性光网络异常流量检测和识别。结果表明,漏检率及误检率较低,分别为3.16%,1.03%。检测用时较少,仅用16秒。在进行检测时,外部入侵率未超过1%,抗扰性较强。
In order to improve the stability of network transmission,an automatic identification algorithm for abnormal traffic in elastic optical networks based on isolated forest algorithm is proposed.Perform spectral density detection based on the abnormal distribution characteristics of traffic and the differences in normal data,construct a spectral feature extraction model for elastic optical network traffic,implement spectral feature filtering for abnormal traffic through low-pass filter convolution vector reorganization,adopt isolated forest algorithm to achieve adaptive optimization control for network traffic anomaly detection,and combine multi-dimensional spatial structure reorganization method to achieve detection and recognition of abnormal traffic in elastic optical network.The results showed that the missed detection rate and the false detection rate were relatively low,3.16%and 1.03%,respectively.The detection takes less time,only 16 seconds.During detection,the external intrusion rate does not exceed 1%,and the immunity is strong.
作者
李橙
何孙秦
卫星
张国华
LI Cheng;HE Sunqin;WEI Xing;ZHANG Guohua(NanJing Normal University Taizhou collage,Taizhou Jiangsu 225300,China;Dapu Primary School,Taizhou Jiangsu 225300,China)
出处
《激光杂志》
CAS
北大核心
2024年第1期179-183,共5页
Laser Journal
基金
江苏省高校自然科学研究面上项目(No.19KJD520008)
江苏省高等学校大学生创新创业训练计划项目(No.201913843018Y)
泰州市科技支撑(社发)项目(No.SSF20190072)。
关键词
孤立森林算法
弹性光网络
异常流量
谱特征提取
isolated forest algorithm
elastic optical network
abnormal flow
spectral feature extraction
