Anti-misbehavior System for Tor Network

Tor is the second generation onion router, supporting the anonymous transport of TCP streams over the Internet. Tor has become the most successful public anonymity communication service in the Internet, and has more than one thousand relay nodes and thousands of users. In this paper, we investigate Tor current exit policies, and found some insufficiencies. Based on the investigation we propose anti-misbehavior system. The system includes two blacklists: global blacklist and local blacklist, and three protocols: reporting misbehavior protocol, building global blacklist protocol and blocking misbehavior users protocol. Reporting misbehavior protocol describes how to report misbehavior from exit node to entry node and how to build local blacklist. Building global blacklist protocol describes how to build global blacklist and how to distribute global blacklist to every Tor node. Blocking misbehavior users protocol describes how to block misbehavior users for the entry node. In addition we also present an evaluation to the system in terms of user experience, performance and anonymity. Through our evaluation, we think anti-misbehavior system can provide better user experience to both Tor users and Tor node administrators; better anti-misbehavior performance; the same transmission performance level with exit policies; and the same anonymity level with exit policies to legitimate Tor users.