Intransitive Non-Interference by Unfolding

Non-interference characterizes the absence of undesired information flows in a computing system, by requiring that activities involving actions with higher level of confidentiality does not cause any observable effect at the lower level. Recently, a causal characterisation of non-interference on Petri nets has been given in terms of the unfolding semantics, a classical true concurrent semantics, enabling quite efficient checks of the property. Concretely, the non-intereference requirement is too strict and it is necessary to offer downgradings, namely the possibility of declassifying information: confidential actions have been performed prior to a downgrading become visible without concern about information leakage. Here we the unfolding-based characterisation can be extended to trhe BINI property (Bisimilarity-based Intransitive NonInterference), a formalization of non-interference in presence of downgrading transitions. This generalisation includes an algorithm for checking BINI for safe Petri nets which relies on the construction of suitable complete prefixes of the unfolding. A further gain of efficiency arises from an orthogonal extension, namely the possibility of dealing with Petri nets with read arcs.