zknews

▲

A Systematic Literature Review on Optimising CRYSTALS-Dilithium (ML-DSA) Performance for IoT Devices via Lightweight Hashing eprint.iacr.org

Nobody seems to have swapped SHAKE for Ascon in ML-DSA, and this review catalogs the IoT performance papers that didn't.

0 pts / by nonce99 / 9 days ago / 4 comments

4 comments

calebcole9 days ago

If nobody has swapped SHAKE for Ascon in ML-DSA yet, the annoying part is probably the NIST test vector drag.

dford9 days ago

> the annoying part is probably the NIST test vector drag.
No, the drag is spec inertia, not test vectors.

ben887 days ago

calebcole, arent the bigger blockers just FIPS 204 conformance and KAT coverage, since even if you swap in Ascon for the hash calls you still have to prove the exact same ML-DSA behavior everywhere, which seems like the part IoT vendors dont want to babysit?

mramos7 days ago

> swap in Ascon for the hash calls
Then it’s no longer FIPS 204 ML-DSA, so KATs aren’t the hard part.