Algebraic aspects of solving Ring-LWE, including ring-based improvements in the Blum-Kalai-Wasserman algorithm
We provide several reductions of Ring-LWE problems to smaller Ring-LWE problems in the presence of samples of a restricted form (i.e. (a,b) such that a is restricted to a subring, or multiplicative coset of a subfield of one CRT factor). To create and exploit such restricted samples, we propose Ring-BKW, a version of the Blum-Kalai-Wasserman algorithm which respects the ring structure. It has several key advantages based on the ring structure, including smaller tables, reduced or eliminated back-substitution, and a new opportunity for parallelization. We focus on two-power cyclotomic Ring-LWE with parameters proposed for practical use, with the exception that many splitting types are considered. The orthogonality of the lattice for two-power cyclotomics is exploited. In general, higher residue degree is an advantage to attacks.
READ FULL TEXT