Adversary Gain vs. Defender Loss in Quantified Information Flow

ABSTRACT. Metrics for quantifying information leakage assume that an adversary’s gain is the defender’s loss. We demonstrate that this assumption does not always hold via a class of scenarios. We describe how to extend quantification to account for a defender with goals distinct from adversary failure. We implement the extension and experimentally explore the impact on the measured information leakage of the motivating scenario.