PWR: Path Weighted Random Scheme in Multipath Traffic Splitting Against Website Fingerprinting

With the continuous evolution of network attack methods, traditional security mechanisms based solely on data content encryption are no longer sufficient to ensure data security and privacy. Traffic analysis techniques, such as website fingerprinting (WFP) attacks, can effectively classify and extract information without cracking encrypted content, which enables the analysis of user behavior. The splitting of traffic in heterogeneous network resources has emerged as a new defense mechanism to disrupt traffic characteristics, however current research has not achieved a balance between security and efficiency. To address this challenge, we propose a path-weighted random (PWR) scheduling mechanism based on path quality to construct a time-varying path selection probability distribution function, that disrupts traffic characteristics and achieves resistance to traffic analysis. Furthermore, we incorporate order scheduling based on data arrival estimation to balance the transmission efficiency in heterogeneous networks. Our proposed solution was experimentally validated in real-world settings, demonstrating superior performance in terms of both security and efficiency.