Science of Quantitative Information Flow

This book presents a comprehensive mathematical theory that explains precisely what information flow is, and how it can be assessed quantitatively (bringing precise meaning to the intuition that certain information leaks are small enough to be tolerated), and how systems can be constructed that achieve rigorous, quantitative information-flow guarantees in those terms. This theory addresses a fundamental challenge: functional and practical requirements frequently conflict with the goal of preserving confidentiality, making perfect security unattainable. The authors include: a systematic presentation of how unwanted information flow, i.e. "leaks", can be quantified in operationally significant ways and then bounded, both with respect to estimated benefit for an attacking adversary and by comparisons between alternative implementations; a detailed study of capacity, refinement, and Dalenius leakage, supporting robust leakage assessments; a unification of information-theoretic channels and information-leaking sequential programs within the same framework; and a collection of case studies, showing how the theory can be applied to interesting realistic scenarios. The text is a unified, self-contained, and comprehensive presentation, accessible to students and researchers with knowledge of discrete probability and some mathematical maturity, with numerous exercises to facilitate use as a course textbook.