Towards Semantic Sensitive Feature Profiling of IoT Devices

Billions of Internet of Things (IoT) devices are being adopted in our daily life as personal wearables, home automation agents, medical appliances, etc. Many domains of their use nowadays rely on the privacy and security of these devices-critical infrastructure, healthcare, logistics, manufacturing. In this paper, we aim to establish a standardized framework, which does not require access to physical devices yet allows to profile security and privacy-sensitive functionality in both existing and upcoming IoT products, based on semantic analysis of discovered technical information. We develop a software tool for automatic feature profiling of IoT devices and present case studies on two real-world IoT devices-a fitness tracker, Garmin Forerunner 230, and a voice-controlled home assistant, Amazon Echo Dot second generation and further provide comparative results analysis.