|
PQ-WB-KEM: White-Box ML-KEM-768 1.0
PQ-WB-KEM: Toward a White-Box Construction of ML-KEM-768 with Arithmetic Masking for M2M Communications - Mathematics (MDPI)
|
White-box implementation overhead calculations.
Collaboration diagram for Overhead Analysis:White-box implementation overhead calculations.
Analyzes the performance overhead introduced by white-box protections compared to baseline implementation.
| Component | Contribution | Basis |
|---|---|---|
| Protected primitive (native C, mask-generation RNG included) | 4.30x | measured, x86-64 (2.85x aarch64/QEMU) |
| Embedded cache factor | ~11x | mid-estimate of a measured 1.9-63x bracket |
| Total (Construction B, decapsulation) | ~47x | deployment projection: 4.30x x 11x |
| Labeled lower bound | ~17x | RNG-excluded masking floor 1.55x x 11x |
The two factors are not strictly independent – the cache factor models the same table lookup already counted inside the protected-primitive term – and neither figure is an end-to-end white-box measurement. The 1.55x floor term substitutes a deterministic stand-in for the per-share mask-generation randomness, which is why the deployment projection anchors on the RNG-included 4.30x primitive instead. Against the 0.213 ms x86-64 host decapsulation baseline the ~47x projection is ~10 ms (7-14 ms with the +/-35% band); no ARM-device wall-clock is claimed (overhead_calculation.json, cache_effect.json, primitive_bench_x86.json).