info-theoretic model of anonymity

[Adam Back]

Just read this paper published in PET02 "Towards an Information
Theoretic Metric for Anonymity" [1]:

	http://www.cl.cam.ac.uk/~gd216/set.pdf
or	http://www.cl.cam.ac.uk/~gd216/set.ps

it uses a Shannon like entropy model for the anonymity provided by a
system uses this model to analyse the effect of different parameters
one can tune with mixmaster (POOLSIZE, RATE, in mixmaster.conf).

The "anonymity entropy" measurement can be interpreted as how many
bits of information the attacker needs to identify a user and is
computed from probabilities.

Would be interesting to try estimate the entropy provided by the
current mixmaster network.  A number of nodes publish their parameter
choices, and traffic volume over time (in hourly increments).

Adam
--
http://www.cypherspace.org/adam/

[1]

@inproceedings{Serjantov:02:info-theoretic-anon,
  author = "Andrei Serjantov and George Danezis",
  title = "Towards an Information Theoretic Metric for Anonymity",
  booktitle = "Proceedings of the Workshop on Privacy Enhancing Technologies",
  year = "2002",
  note = "Also available as \url{http://www.cl.cam.ac.uk/~aas23/papers_aas/set.ps}"
}