Span Programs and General Secure Multi-Party Computation
AbstractThe contributions of this paper are three-fold. First, as an abstraction of previously proposed cryptographic protocols we propose two cryptographic primitives: homomorphic
shared commitments and linear secret sharing schemes with an additional multiplication property. We describe new constructions for general secure multi-party computation protocols, both in the cryptographic and the information-theoretic (or secure
channels) setting, based on any realizations of these primitives.
Second, span programs, a model of computation introduced by Karchmer and Wigderson, are used as the basis for constructing new linear secret sharing schemes, from which the two above-mentioned primitives as well as a novel verifiable secret sharing scheme can efficiently be realized. Third, note that linear secret sharing schemes can have arbitrary (as opposed to
threshold) access structures. If used in our construction, this yields multi-party protocols secure against general sets of active adversaries, as long as in the cryptographic (information-theoretic) model no two (no three) of these potentially misbehaving player sets cover the full player set. This is a strict generalization of the threshold-type adversaries and results previously considered in the literature. While this result is new for the cryptographic model, the result for the information-theoretic model was previously proved by Hirt and Maurer. However, in addition to providing an independent proof, our protocols are not recursive and have the potential of being more efficient.
How to Cite
Cramer, R., Damgård, I., & Maurer, U. (1997). Span Programs and General Secure Multi-Party Computation. BRICS Report Series, 4(28). https://doi.org/10.7146/brics.v4i28.18954
Articles published in DAIMI PB are licensed under a Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Unported License.