Identify IoT Devices through Web Interface Characteristics

Abstract

With the continuous expansion of the business scale of the Internet of Things, the concept of the Internet of Everything has been deeply rooted in the hearts of the people. At the same time, various security issues have been exposed, which has greatly affected the trust of enterprises and users. Discovering and identifying IoT devices as an important part of IoT security should be taken seriously. In this paper, we propose a new method for classifying devices on the Internet of Things, which uses the data characteristics of the main page of the device's web interface to classify the vendor name, even the product name and other information of the device. We randomly selected IP address ranges on the Internet for scanning, and obtained a training set containing 80,000 entries and covering more than 1,000 devices from 74 vendors. Using the random forest algorithm and proceeding 10-fold cross-validation, our method classified IoT devices with an accuracy rate of over 97.5%.