Paper 2025/1487

Towards Building Efficient SCALES Protocols

Anasuya Acharya, Aarhus University
Carmit Hazay, Bar-Ilan University
Vladimir Kolesnikov, Georgia Institute of Technology
Manoj Prabhakaran, Indian Institute of Technology Bombay
Abstract

SCALES (Small Clients And Larger Ephemeral Servers) (Acharya et al., TCC 2022, CRYPTO 2024) is a recently proposed model for MPC with several attractive features, including resilience to adaptive corruption. Known SCALES constructions, while offering reasonable asymptotics for large-scale MPC, incur high concrete costs both in computation and communication. As our primary contribution, we dramatically improve both asymptotic and concrete costs of SCALES for permutation branching programs (PBP), a well-motivated practical model of computation. We achieve linear cost in program length, input size, and the security parameter. Our instantiations of the building blocks may be of independent interest. Further, we present generic transformations to extend any semi-honestly secure SCALES protocol to achieve (1) guaranteed output delivery in the presence of mixed adversaries (that corrupt servers maliciously and clients semi-honestly) in the all-but-one corruption setting; and (2) protocols for computing general functionalities where each server's computation scales sub-linearly in the function~size.

Metadata
Available format(s)
PDF
Category
Cryptographic protocols
Publication info
A major revision of an IACR publication in ASIACRYPT 2025
Keywords
SCALESMPC with Ephemeral ServersSemi-Honest SecurityGuaranteed Output DeliveryRerandomizable Garbled Circuits
Contact author(s)
acharya @ cs au dk
carmit hazay @ biu ac il
kolesnikov @ gatech edu
mp @ cse iitb ac in
History
2025-08-20: approved
2025-08-17: received
See all versions
Short URL
https://ia.cr/2025/1487
License
Creative Commons Attribution
CC BY

BibTeX

@misc{cryptoeprint:2025/1487,
      author = {Anasuya Acharya and Carmit Hazay and Vladimir Kolesnikov and Manoj Prabhakaran},
      title = {Towards Building Efficient {SCALES} Protocols},
      howpublished = {Cryptology {ePrint} Archive, Paper 2025/1487},
      year = {2025},
      url = {https://eprint.iacr.org/2025/1487}
}
Note: In order to protect the privacy of readers, eprint.iacr.org does not use cookies or embedded third party content.