Supporting privacy preferences in credential-based interactions

Users can today enjoy the many benefits brought by the development and widespread adoption of Internet and related services conveniently accessing digital resources. Servers offering such resources typically require users to release information about them, which servers can then use for enforcing possible access policies on the offered services. A major problem in this context relates to providing users with the ability of determining which information to release to satisfy the server requests during their electronic interactions. In this paper, we provide an approach for empowering the user in the release of her digital portfolio based on simple sensitivity labels expressing how much the user values different properties, credentials or combinations thereof, as well as on additional constraints that the user might impose on information disclosure. We provide a generic modeling of the problem and illustrate its translation in terms of a Weighted MaxSat problem, which can be conveniently and efficiently managed by off the shelf SAT solvers, thus resulting efficient and scalable.