2003 Symposium on Security and Privacy, 2003. (2003)
May 11, 2003 to May 14, 2003
Dakshi Agrawal , IBM Watson Research Center
Dogan Kesdogan , Aachen University of Technology
Stefan Penz , Aachen University of Technology
The goal of anonymity providing techniques is to preserve the privacy of users, who has communicated with whom, for how long, and from which location, by hiding traffic information. This is accomplished by organizing additional traffic to conceal particular communication relationships and by embedding the sender and receiver of a message in their respective anonymity sets. If the number of overall participants is greater than the size of the anonymity set and if the anonymity set changes with time due to unsynchronized participants, then the anonymity technique becomes prone to traffic analysis attacks. In this paper, we are interested in the statistical properties of the disclosure attack, a newly suggested traffic analysis attack on the MIXes. Our goal is to provide analytical estimates of the number of observations required by the disclosure attack and to identify fundamental (but avoidable) ?weak operational modes? of the MIXes and thus to protect users against a traffic analysis by the disclosure attack.
D. Agrawal, S. Penz and D. Kesdogan, "Probabilistic Treatment of MIXes to Hamper Traffic Analysis," 2003 Symposium on Security and Privacy, 2003.(SP), Berkeley, CA, 2003, pp. 16.