DevPF: Device identification through passive fingerprints in IoT

In the realm of Industrial IoT, devices are characterized by their diversity and complexity. Upon connection, these devices are exposed to a myriad of security threats. Nevertheless, the adoption of unique device fingerprints facilitates the secure identification and authentication of these devices during access. In this study, we propose DevPF, a novel framework for IoT device identification based on passive traffic analysis. Unlike traditional active detection methods, DevPF analyzes network traffic characteristics without interacting with devices, ensuring minimal network load and no interference with device operations. The framework utilizes machine learning algorithms to classify devices based on extracted traffic features. Experimental results on a public IoT dataset demonstrate that DevPF achieves high classification accuracy, making it a promising solution for enhancing the security of IoT networks.