Slotted packet counting attacks on anonymity protocols

Abstract

In this paper we present a slotted packet counting attack against anonymity protocols. Common packet counting attacks make strong assumptions on the setup and can easily lead to wrong conclusions, as we will show in our work. To overcome these limitations, we account for the variation of traffic load over time. We use correlation to express the relation between sender and receiver nodes. Our attack is applicable to many anonymity protocols. It assumes a passive attacker and works with partial knowledge of the network traffic.