Intelligent Device Identification Method Based on Network Packet Fingerprint

Cyberspace intelligent devices refer to the general term for devices that can connect to the Internet and have data computing and processing capabilities. With the development of the Internet, the Internet of Things, and communication protocols, intelligent devices have been widely used in various fields, but they have also brought many cyberspace security issues. The implementation of cyberspace asset detection has a comprehensive grasp of the types, versions, and manufacturers of intelligent devices, which is of great significance to cyberspace security management. In order to automatically and accurately discover cyberspace intelligent devices, this paper proposes an identification method based on network packet fingerprint. This method uses a public device dataset, improves the fingerprint extraction method of existing work, and expands the data for the problem of unbalanced distribution of the dataset. Finally we use machine learning algorithms to classify devices. Through experimental evaluation, it is verified that our method can effectively identify a variety of cyberspace intelligent devices, and has an accuracy of 77.1% and a recall of 68.2%. Compared with the previous results, our fingerprint extraction method has improved the accuracy by 4.8%, and through data expansion, the recall has been improved by 8.3%.