Quantitative Information Flow as Safety and Liveness Hyperproperties
This work addresses verification challenges in quantitative information flow for security applications, representing an incremental advancement by building on existing hyperproperty frameworks.
The paper tackles the problem of verifying quantitative information flow by using hyperproperties to classify verification issues, resulting in a unified and extended understanding of hardness in checking and inferring such flows, with the identification of a subclass called 'k-observable hyperproperties' that enables checking via self composition relative to a reachability oracle.
We employ Clarkson and Schneider's "hyperproperties" to classify various verification problems of quantitative information flow. The results of this paper unify and extend the previous results on the hardness of checking and inferring quantitative information flow. In particular, we identify a subclass of liveness hyperproperties, which we call "k-observable hyperproperties", that can be checked relative to a reachability oracle via self composition.