Cache enhanced anonymity systems against probabilistic attacks

Abstract

The technique of data caching is being increasingly recognized for its ability to increase the amount of anonymity provided by an anonymous communication network, and is expected to be widely adopted in the next generation of these systems. We present a method to measure the degree of anonymity remaining in a cache enhanced anonymity system after a probabilistic attack has been carried out on it. Our method determines the probability distribution induced by the attack on all possible communication patterns of being the true one, from which a system-wide anonymity metric is developed. The scope of our metric, in terms of the attacks it is applicable to, is far wider than that of existing methods.