Quantitative Analysis of Information Leakage in Probabilistic and Nondeterministic Systems

This thesis addresses the foundational aspects of formal methods for applications in security and in particular in anonymity. More concretely, we develop frameworks for the specification of anonymity properties and propose algorithms for their verification. Since in practice anonymity protocols always leak some information, we focus on quantitative properties, which capture the amount of information leaked by a protocol. The main contribution of this thesis is cpCTL, the first temporal logic that allows for the specification and verification of conditional probabilities (which are the key ingredient of most anonymity properties). In addition, we have considered several prominent definitions of information-leakage and developed the first algorithms allowing us to compute (and even approximate) the information leakage of anonymity protocols according to these definitions. We have also studied a well-known problem in the specification and analysis of distributed anonymity protocols, namely full-information scheduling. To overcome this problem, we have proposed an alternative notion of scheduling and adjusted accordingly several anonymity properties from the literature. Our last major contribution is a debugging technique that helps on the detection of flaws in security protocols.

💡 Research Summary

The thesis tackles the long‑standing challenge of formally analysing anonymity protocols, which inevitably leak some information, by providing a comprehensive quantitative framework. Its central contribution is the introduction of cpCTL (conditional probability Computation Tree Logic), the first temporal logic capable of expressing and verifying properties that involve conditional probabilities—an essential ingredient for most anonymity specifications. In cpCTL, state and path formulas can be annotated with operators such as P≥p