A Lightweight Monitoring and Anomaly Detection Framework for IoT Devices
Jie Yin, Yutaka Ishikawa, Atsuko Takefusa · 2025
The always-online nature and the lack of sufficient built-in protection make Internet of Things (IoT) devices highly susceptible to various cyberthreats. An efficient and effective anomaly detection system is an essential need for IoT device security. However, it’s challenging to apply advanced host-based anomaly detection techniques from conventional systems to IoT devices due to the device resource constraints. This paper introduces a monitoring and anomaly detection framework based on thread-level system call streams for IoT devices. It leverages the execution pattern of IoT applications and detects anomalies by analyzing system call arguments and associated I/O attributes, in addition to the invocation sequence in real-time. The evaluation results highlight the feasibility of the proposed approach in terms of both performance and security.