Key points are not available for this paper at this time.
The Internet of Things (IoT) becomes a novel paradigm as more and more devices are connected to the Internet, enabling several innovative applications such as smart home, industrial automation, and connected health. However, the cyber-attack to these applications is a big issue and countermeasures are in dire need to provide system security and user privacy. In this paper, we address the traffic analysis attack to smart homes, where adversaries intercept the Internet traffic from/to the smart home gateway and profile residents' behaviors through digital traces. Traditional cryptographic tools may not work well due to the effectiveness of adversaries' machine learning algorithms in classifying encrypted traffic, so here we propose a privacy-preserving traffic obfuscation framework to achieve the goal. To be specific, we leverage the smart community network of wirelessly connected smart homes and intentionally direct each smart home's traffic to another home gateway before entering the Internet. The design jointly considers the network energy consumption and the resource constraints in IoT devices, while achieving strong differential privacy guarantee so that adversaries cannot link any traffic flow to a specific smart home. Besides, we consider a hostile smart community network and develop secure multihop routing protocols to guarantee the source/destination unlinkability and satisfy each user's personalized privacy requirement. To evaluate the effectiveness of our framework in protecting privacy and reducing network energy consumption, extensive simulations are conducted and the results demonstrate that our design outperforms other differential privacy mechanism in preserving privacy and minimizing network utility cost.
Liu et al. (Thu,) studied this question.