Pollution and whitewashing attacks in a P2P live streaming system: Analysis and counter-attack

Abstract

P2P live streaming are increasingly popular nowadays. Due to their popularity, these systems may be a target of attacks and opportunistic user behavior. In this paper, we address the pollution attacks in such systems. We present a pollution damage model and also analyze a reputation system as a tool to fight attacks in P2P live streaming systems. The model we propose evidences that attacks are harmful even in a system with a small number of polluters. In this case, peers must have more than 3 times network bandwidth than they should have in a system without polluters. Our experimental results on PlanetLab show that just check data integrity is not an effective protection. In this case, we observe a very high data loss rate. Finally, the reputation system is effective against pollution attack. When peers do not whitewash their identities, the reputation system quickly identifies polluters. In this case, the overhead and loss rate can be negligible. During a whitewashing, the new approach presents less than 20% overhead and 3% of loss.