TorKameleon: Improving Tor's Censorship Resistance With K-anonymization and Media-based Covert Channels

Anonymity networks like Tor greatly improve online privacy but are susceptible to correlation attacks from state-level adversaries and Internet censors. To enhance privacy, covert channels encapsulated in media protocols, particularly WebRTC-based encapsulation, have shown promise against passive traffic correlation attacks. However, their effectiveness against active correlation attacks has not been explored, and compatibility with Tor remains limited. This paper introduces TorKameleon, a censorship evasion solution that protects Tor users from passive and active correlation attacks. It incorporates K-anonymization techniques to fragment and reroute traffic through multiple paths formed by multiple proxies and uses covert WebRTC-based channels or TLS tunnels to encapsulate user traffic. The developed prototype has undergone extensive validation for performance and resilience against correlation attacks, showcasing its effectiveness.