Iotspy: Uncovering Human Privacy Leakage In Iot Networks Via Mining Wireless Context

Internet of Things (IoT) has been emerging as one of the most significant technologies that may change the world. The development and deployment of IoT systems significantly improve the efficiency of work, advance the development of productive force, and affect the human society's modes of production, working, and life deeply. The number of deployed IoT devices and applications has had explosive growth in the past years. While people enjoy the benefits brought by IoT, the addressing of its security and privacy issues does not keep pace with the development of IoT technologies. We find that encrypted wireless IoT traffic can still leak information about user privacy. We can infer what the user is doing at home by deploying a wireless sniffer outside the user's house.In this paper, we propose a method to eavesdrop different kinds of user privacy via analyzing the wireless context. First, we extract the packet sequence features to fingerprint and detect the IoT events. Then we analyze the detected IoT events and infer the user's activities and even the user's moods. Furthermore, by analyzing the wireless context, which is discovered by mining the IoT event dependencies, we can infer the user's living habits, routines, and even installed IoT applications. The leakage of such information not only exposes the user's privacy but also extends the attack surface that an attacker can utilize to control the smart home. We implement our eavesdropping approach and conduct extensive experiments on the SmartThings platform. The evaluation demonstrates the feasibility and effectiveness of our eavesdropping approach.