Edge Machine Learning to Detect Malicious Activity in IoT Devices through System Calls and Traffic Analysis

With the rise of Internet of Things (IoT) came the explosion of smart devices and applications, as well as the security concerns to protect such devices from malicious activity. Although such devices are still with limited compute power and resources, they are capable of processing and running pre-trained machine learning models. That is where Edge Machine Learning (Edge ML) comes into play. Edge ML is a technique by which Smart Devices can process data locally (at the device-level) using machine learning algorithms. This reduces the reliance on Cloud computing making decisions faster and on the spot an advantage for security objectives. In this paper, we leverage the Edge ML approach in order to perform traffic analysis as well as system call tracing and monitoring in order to detect the existence of any malicious activity at the edge devices in IoT networks.