Development of Fingerprint Identification Based on Device Flow in Industrial Control System

Jun Tao,Xin Yuan, Shengze Zhang,Yifan Xu

APPLIED SCIENCES-BASEL(2023)

引用 0|浏览10
暂无评分
摘要
With the rapid development of industrial automation technology, a large number of industrial control devices have emerged in cyberspace, but the security of open cyberspace is difficult to guarantee. Attacks on industrial control devices can directly endanger the environment and even life safety. Therefore, how to monitor the industrial control system in real time has become the primary problem, and device identification is the basic guarantee of safety monitoring. There are limitations in building device identification model based on IP address or machine learning. The paper aim at the development of a device traffic fingerprint model and identify the device based on the periodicity of device traffic. The model generates device fingerprints based on pattern sequences abstracted from the traffic and suffix array algorithm. In the process of recognition, the exact pattern matching algorithm is used for preliminary judgment. If the exact pattern matching fails to hit, the final judgment is made by combination fuzzy pattern matching. This paper also proposes a diagonal jump algorithm to optimize the updating of the distance matrix, which saves on the computational cost of fuzzy pattern matching. Simulation results show that compared with SVM, random forest, and LSTM model, the device traffic fingerprint model has good performance advantages in accuracy, recall and precision.
更多
查看译文
关键词
pattern matching,industrial control system,device traffic fingerprint,device identification
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要