Uniform First-Order Threshold Implementations

Abstract

Most masking schemes used as a countermeasure against side-channel analysis attacks require an extensive amount of fresh random bits on the fly. This is burdensome especially for lightweight cryptosystems. Threshold implementations (TIs) that are secure against first-order attacks have the advantage that fresh randomness is not required if the sharing of the underlying function is uniform. However, finding uniform realizations of nonlinear functions that also satisfy other TI properties can be a challenging task. In this paper, we discuss several methods that advance the search for uniformly shared functions for TIs. We focus especially on three-share implementations of quadratic functions due to their low area footprint. Our methods have low computational complexity even for 8-bit Boolean functions.