TY - GEN
T1 - Fully Adaptive Decentralized Multi-Authority ABE
AU - Datta, Pratish
AU - Komargodski, Ilan
AU - Waters, Brent
N1 - Publisher Copyright: © 2023, International Association for Cryptologic Research.
PY - 2023
Y1 - 2023
N2 - Decentralized multi-authority attribute-based encryption (MA- ABE ) is a distributed generalization of standard (ciphertext-policy) attribute-based encryption where there is no trusted central authority: any party can become an authority and issue private keys, and there is no requirement for any global coordination other than the creation of an initial set of common reference parameters. We present the first multi-authority attribute-based encryption schemes that are provably fully-adaptively secure. Namely, our construction is secure against an attacker that may corrupt some of the authorities as well as perform key queries adaptively throughout the life-time of the system. Our main construction relies on a prime order bilinear group where the k-linear assumption holds as well as on a random oracle. Along the way, we present a conceptually simpler construction relying on a composite order bilinear group with standard subgroup decision assumptions as well as on a random oracle. Prior to this work, there was no construction that could resist adaptive corruptions of authorities, no matter the assumptions used. In fact, we point out that even standard complexity leveraging style arguments do not work in the multi-authority setting.
AB - Decentralized multi-authority attribute-based encryption (MA- ABE ) is a distributed generalization of standard (ciphertext-policy) attribute-based encryption where there is no trusted central authority: any party can become an authority and issue private keys, and there is no requirement for any global coordination other than the creation of an initial set of common reference parameters. We present the first multi-authority attribute-based encryption schemes that are provably fully-adaptively secure. Namely, our construction is secure against an attacker that may corrupt some of the authorities as well as perform key queries adaptively throughout the life-time of the system. Our main construction relies on a prime order bilinear group where the k-linear assumption holds as well as on a random oracle. Along the way, we present a conceptually simpler construction relying on a composite order bilinear group with standard subgroup decision assumptions as well as on a random oracle. Prior to this work, there was no construction that could resist adaptive corruptions of authorities, no matter the assumptions used. In fact, we point out that even standard complexity leveraging style arguments do not work in the multi-authority setting.
UR - http://www.scopus.com/inward/record.url?scp=85161645196&partnerID=8YFLogxK
U2 - https://doi.org/10.1007/978-3-031-30620-4_15
DO - https://doi.org/10.1007/978-3-031-30620-4_15
M3 - Conference contribution
SN - 9783031306198
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 447
EP - 478
BT - Advances in Cryptology – EUROCRYPT 2023 - 42nd Annual International Conference on the Theory and Applications of Cryptographic Techniques, 2023, Proceedings
A2 - Hazay, Carmit
A2 - Stam, Martijn
PB - Springer Science and Business Media Deutschland GmbH
T2 - 42nd Annual International Conference on Theory and Applications of Cryptographic Techniques, EUROCRYPT 2023
Y2 - 23 April 2023 through 27 April 2023
ER -