An evidence theory based anonymity metrics

For privacy protection, several anonymous communication systems have been presented in the past a few years, such as Crowds, Onion routing, TOR etc. In order to measure the anonymity these systems provided, an evidence theory based approach is proposed in this paper, which presents Evidence Collection Model (ECM) to collect evidence, i.e. leaked attributes that can be obtained by attackers from userspsila actions; presents Anonymity State Model (ASM) to describe the dynamic transition of anonymity states within a given time period. According to these two models, a new method to quantify the anonymity and the unlinkability of actions is introduced. Compared with the traditional Shannon information theory, our new approach is shown to be more general and useful for evaluating the anonymity.