Stochastic analysis of a randomized detection algorithm for pollution attack in P2P live streaming systems

Pollution attack is known to have a disastrous effect on existing P2P infrastructures: it can reduce the number of legitimate P2P users by as much as 85%, and it generates abundant bogus data which may deplete the communication bandwidth. We propose a distributed defense and detection mechanism to resolve pollution attacks. The mechanism is composed of a set of “randomized” and “fully distributed” algorithms that can be executed by any legitimate peer. We present the analytical framework to quantify (a) the probability of false negative, (b) the probability of false positive, and (c) the distribution of time needed for detection. In our detection algorithm and analysis, we consider the case of (1) single attacker within the neighborhood, (2) multiple attackers within the neighborhood. Furthermore, we show how to “optimize” the system parameters so as to quickly discover and eliminate malicious peers from the system.