Using linkability information to attack mix-based anonymity services

Abstract

There exist well established models for anonymity focusing on traffic analysis, i. e., analysing properties of single messages as, e. g., timing. However there is only little work done that use linkability information, that is information about the probability that two messages have been sent by the same sender. In this paper we model information about linkability between messages as a weighted graph. We show lower and upper bounds with regards to the usefulness of linkability information for matching messages to senders. In addition to that we present simulation results, showing to which extent a matching of messages to senders is possible by using linkability information with different grades of noise. © 2009 Springer Berlin Heidelberg.