On unconditionally binding code-based commitment schemes

TLDR: This work constructs a dual version of statistically binding commitment scheme by Jain et al. (Asiacrypt 2012) with shorter commitment size under hardness of syndrome decoding, and observes that perfect binding requires exact knowledge of minimal distance of the underlying code.

Abstract: In this work, we construct a dual version of statistically binding commitment scheme by Jain et al. (Asiacrypt 2012) with shorter commitment size under hardness of syndrome decoding. Then, we point out that perfectly binding variants of the above schemes follow directly from the Randomized McEliece and Niederreiter public key encryption schemes, assuming indistinguishability of permuted Goppa codes, as well as hardness of the exact learning parity with noise (xLPN) problem (for the McEliece scheme) and hardness of syndrome decoding (for the Niederreiter scheme). Our key observation here is that perfect binding (as opposed to statistical binding) requires exact knowledge of minimal distance of the underlying code. Finally, we provide security evaluation of our proposals, and compare their performance with that of existing schemes.