Implementing anonymous credentials over bilinear groups: A practical approach to privacy-preserving multi-company access services

Details

Supervisors

Peters, Thomas

Faculty

Ecole polytechnique de Louvain

Degree label

Master [120] en cybersécurité, à finalité spécialisée: stratégies en entreprise

Abstract

In the domain of digital identity management, traditional centralised systems have demonstrated vulnerabilities, necessitating the development of more secure and privacy-preserving solutions. This thesis examines the potential use of different cryptographic building blocks in a collaborative anonymous credentials setting. Anonymous credentials enable individuals to assert specific attributes without disclosing their full identity, leveraging cryptographic techniques to safeguard personal information exchanges. The primary objective of this research is to develop a collaborative Common Reference String (CRS) framework for a multi-signer model. This extends the existing single-signer model of the work "Practical Signatures with Efficient Protocols from Simple Assumptions" by Libert, Mouhartem, Peters, and Yung, by distributing the CRS computation among all the entities participating in the collaborative system. This enhances security through decentralised trust. The theoretical foundation is based on the well-known SXDH cryptographic assumption, with a focus on enhancing the security of both the user and the entity, given that adversarial actors may participate in the system. Furthermore, this study integrates the collaborative CRS scheme with established techniques such as range proofs to form the anonymous credentials system, with the objective of facilitating efficient and secure attribute validation. The system addresses as an example the critical challenges in identity management within a collaborative transportation application by allowing users to selectively disclose verifiable information while preserving their anonymity. This theoretical objectives supported by the implementation of the proposed framework using Python and the MIRACL library, with particular emphasis on the framework’s feasibility as a provable application for anonymous credentials.