Secure and Efficient Masking of Lightweight Ciphers in Software and Hardware.

The article focuses on improving the efficiency of masked implementations against side-channel attacks for bitslice lightweight cryptographic algorithms RECTANGLE, TANGRAM, and KNOT, both in software and hardware. For software masking, it presents optimized higher-order Ishai–Sahai–Wagner (ISW) masking implementations on 32-bit ARM microprocessors using a Common Shares technique that reduces the number of nonlinear cross-products, achieving a 13%-34% speed-up in parallel ISW-based secure logical AND operations. For hardware masking, the paper develops decomposition methods of 4-bit cubic Sboxes—used in the three ciphers and several NIST Lightweight Cryptography (LWC) third-round candidates—into compositions of quadratic permutations and affine transformations, enabling efficient threshold implementations (TI) with 3 and 4 shares that minimize area and latency. The authors provide exhaustive search algorithms and detailed decompositions for various cubic Sbox classes, demonstrating the fastest and smallest hardware implementations for TI, and discuss open questions regarding decomposition lengths and efficient Sbox design within equivalence classes.