TWC: Small: Functional Reactive Cryptography

University Of California-San Diego, La Jolla CA

Investigators

Abstract

Cryptographic protocols are a fundamental tool to secure distributed computer applications, but also notoriously hard to design and analyze. With modern computer applications becoming increasingly complex, interconnected, distributed and interactive, there is a pressing need to improve the researchers' ability to design and analyze protocols that go well beyond the traditional problem of secure message transmission. This research investigates frameworks to improve the way cryptographic protocols are designed and analyzed, by investigating new mathematical models of computation specifically targeted to security analysis. These methods will allow security analysis of more complex applications, and also to improve public confidence in the analysis itself. This project explores computation models where time is only treated implicitly, through logical dependencies. Avoiding the explicit modeling of time greatly simplifies the analysis of cryptographic protocols, still capturing timing constraints that are relevant to security properties. Different communication and security models are considered, and compared to each other via reductions, with respect to security guarantees, their ability to capture realistic attacks, and expressiveness in describing protocols of cryptographic interest. Models are evaluated by exemplifying their use in the analysis of a set of representative case studies, including Oblivious Transfer protocols, and protocols for secure multiparty computation. These tools and techniques will help developers to more easily reason about the security of a wider range of applications, and lead to more secure and trustworthy software.