A Dynamic Method to Protect User Privacy Against Traffic-based Attacks on Smart Home

Abstract

The Internet of Things (IoT) has revolutionized how people interact with their living spaces. However, attackers can perform traffic-based attacks to reveal the behavior of legitimate users, seriously compromising their privacy. Studies have proposed to obfuscate network traffic to avoid these attacks. However, there is still the challenge of ensuring a trade-off between privacy and network overhead. This work introduces the MITRA method to protect user privacy in smart homes while keeping IoT network overhead low. The method relies on the dummy traffic injection following different levels of obfuscation. The different levels mask network traffic, improving privacy without harming network performance unnecessarily. Results show that the obfuscation of IoT device traffic reduces the traffic identification accuracy by up to 42%.