Enhancing Privacy in PUF-Cash through Multiple Trusted Third Parties and Reinforcement Learning
Georgios Fragkos,Cyrus Minwalla,Eirini Eleni Tsiropoulou,Jim Plusquellic
DOI: https://doi.org/10.1145/3441139
IF: 2.013
2022-01-31
ACM Journal on Emerging Technologies in Computing Systems
Abstract:Electronic cash ( e-Cash ) is a digital alternative to physical currency such as coins and bank notes. Suitably constructed, e-Cash has the ability to offer an anonymous offline experience much akin to cash, and in direct contrast to traditional forms of payment such as credit and debit cards. Implementing security and privacy within e-Cash, i.e., preserving user anonymity while preventing counterfeiting, fraud, and double spending, is a non-trivial challenge. In this article, we propose major improvements to an e-Cash protocol, termed PUF-Cash, based on physical unclonable functions ( PUFs ). PUF-Cash was created as an offline-first, secure e-Cash scheme that preserved user anonymity in payments. In addition, PUF-Cash supports remote payments; an improvement over traditional currency. In this work, a novel multi-trusted-third-party exchange scheme is introduced, which is responsible for “blinding” Alice’s e-Cash tokens; a feature at the heart of preserving her anonymity. The exchange operations are governed by machine learning techniques which are uniquely applied to optimize user privacy, while remaining resistant to identity-revealing attacks by adversaries and trusted authorities. Federation of the single trusted third party into multiple entities distributes the workload, thereby improving performance and resiliency within the e-Cash system architecture. Experimental results indicate that improvements to PUF-Cash enhance user privacy and scalability.
engineering, electrical & electronic,computer science, hardware & architecture,nanoscience & nanotechnology