一种基于相关峰统计特征的光纤以太网设备指纹识别系统  

Fingerprint Identification System of Optical Fiber Ethernet Equipment based on Statistical Characteristics of Correlation Peaks

作　　者：陈超 彭林宁 张广凯 CHEN Chao;PENG Lin-ning;ZHANG Guang-kai(Schoolof Cyber Science and Engineering,Southeast University,Nanjing Jiangsu 211189,China;PurpleMountain Laboratories,Nanjing Jiangsu 211111,China;CRRC Information Technology Co.,Ltd.,Beijing 100084,China)

机构地区：[1]东南大学网络空间安全学院,江苏南京211189 [2]网络通信与安全紫金山实验室,江苏南京211111 [3]中车信息技术有限公司,北京100084

出　　处：《通信技术》2020年第2期284-292,共9页Communications Technology

基　　金：国家自然科学基金（No.61571110,No.61601114,No.61602113）;江苏省自然科学基金（No.BK20160692）~~

摘　　要：针对光纤以太网设备在访问认证中存在安全隐患的问题,首次提出了一种基于相关峰的光纤以太网设备物理指纹提取方法,设计了一个由24个光纤以太网器件组成的强度调制/直接检测(Intensity Modulation/Direct Detection,IM/DD)实验系统。采用一种适应环境条件的混合自适应分类方案,通过进行大量的实验来优化参数设置,并与基于经典统计特征的方法进行了比较。实验结果表明,在0 dB和20 dB的信噪比下,设计的系统分类精度可以达到85.76%和91.11%,显著优于基于经典统计特征的方法。Aiming at the problem of hidden security risks in access authentication of optical fiber Ethernet devices,a physical fingerprint extraction method of optical fiber Ethernet devices based on correlation peaks is proposed for the first time.At the same time,an IM/DD(Intensity Modulation/Direct Detection)experimental system composed of 24 fiber-optic Ethernet devices is designed.A hybrid adaptive classification scheme adapted to environmental conditions is used to optimize parameter settings through a large number of experiments,and compared with methods based on classic statistical characteristics.Experimental results indicate that the designed system classification accuracy can reach 85.76%and 91.11%under the signal-to-noise ratio of 0 dB and 20 dB,which is significantly better than the method based on classical statistical features.

关 键 词：光纤以太网设备 相关峰 物理指纹 统计特征 

分 类 号：TP309.1[自动化与计算机技术—计算机系统结构]