Traditional techniques to detect malware infections were not meant to be used by the end-user and current malware removal tools and security software cannot handle the heterogeneity of IoT devices. In this paper, we design, develop and evaluate a tool, called NURSE, to fill this information gap, i.e., enabling end-users to detect IoT-malware infections in their home networks. NURSE follows a modular approach to analyze IoT traffic as captured by means of an ARP spoofing technique which does not require any network modification or specific hardware. Thus, NURSE provides zero-configuration IoT traffic analysis within everybody's reach. After testing NURSE in 83 different IoT network scenarios with a wide variety of IoT device types, results show that NURSE identifies malware-infected IoT devices with high accuracy (86.7%) using device network behavior and contacted destinations.
@article{arxiv.2203.04776,
title = {NURSE: eNd-UseR IoT malware detection tool for Smart homEs},
author = {Antoine d'Estalenx and Carlos H. Gañán},
journal= {arXiv preprint arXiv:2203.04776},
year = {2022}
}