Fully Dynamic Attribute-Based Signatures for Circuits from Codes
Publication Name
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Abstract
Attribute-Based Signature (ABS), introduced by Maji et al. (CT-RSA’11), is an advanced privacy-preserving signature primitive that has gained a lot of attention. Research on ABS can be categorized into three main themes: expanding the expressiveness of signing policies, enabling new functionalities, and providing more diversity in terms of computational assumptions. We contribute to the development of ABS in all three dimensions, by providing a fully dynamic ABS scheme for arbitrary circuits from codes. The scheme is the first ABS from code-based assumptions and also the first ABS system offering the full dynamicity functionality (i.e., attributes can be enrolled and revoked simultaneously). Moreover, the scheme features much shorter signature size than a lattice-based counterpart proposed by El Kaafarani and Katsumata (PKC’18). In the construction process, we put forward a new theoretical abstraction of Stern-like zero-knowledge (ZK) protocols, which are the major tools for privacy-preserving cryptography from codes. Our main insight here actually lies in the questions we ask about the fundamental principles of Stern-like protocols that have remained unchallenged since their conception by Stern at CRYPTO’93. We demonstrate that these long-established principles are not essential, and then provide a refined framework generalizing existing Stern-like techniques and enabling enhanced constructions.
Open Access Status
This publication is not available as open access
Volume
14601 LNCS
First Page
37
Last Page
73
Funding Number
2022YFA1004900
Funding Sponsor
Ministry of Education - Singapore