Minutes of the December 3rd, 2024 Meeting

[johngray-dev]

Attended by: Felipe, Christopher Abel, Mike Ounsworth, Tomofumi, Petr Muzik, ALexand Ralien, Varun, John Gray, Dan VanGeest, Jan Kalubner, Julient Prat, Jean-Pierre Fiset, Markku,

Round table Discussion:

• Varum - seventhsense.ai - Face derived keys - composite signatures

• JP Fiset - Cyrpto4a - Working on composite signatures

• Petr Muzikant - Experimenting with Post quantum of e-government services. Interoperability of system components.

• Julien Pratt - Testing interoperability, since last hackathon haven't submitted anything new - just pushed test vectors for pure and pre-hash mode

• Daniel VanGeest - Interested in looking at private key interop testing

• Markku - Hashing fast, signing

• Christopher - Kyber and Dilithium in a satelite communication protocol. SP-800-227 - Guidance on KEM and combining two KEMS. SP-800-227 - they have scheduled a workshop in February and we will see it before the workshop

• Alexander - Test suite for CMP - internal product on GitHub soon. Wrapper for the python cryptography library. Interested in making a python wrapper. Rust can be used to generate wrappers for Python, Swift, etc. It is possible to generate python binding cryptographic primitives.

• John G - Working on updated IETF composite signatures and KEM. Working on removing ASN.1 wrappings arounds composite keys, signatures and ciphertexts.

• Abel - Added artifacts recently for interop testing.

• External Mu - HSM cares - software library has to care. Crypto4a has those use-cases, where mu is computed ahead.

• Participate in the Hackathon - compatibility testing, for composite great.

• Abel - would like participation through e-mail

Open quantum safe

• about to release version 0.12 with ML-DSA context Support - This has been released since these notes were originally written.

External Mu discussion

• Mike went over the status of the external mu (FIPS 204 Algorithm 7 step 6).
• This use case is in hardware for example if only SHA2 is available. You need to have a sha3 module. Only matters if a protocol that doesn't sign digest at the protocol layer.
• HashML-DSA - it has no value, should use external mu. Why did NISTinclude HashML-DSA? It may have been for CNSA for the test vector. The internal ones are XOF's,
• HashML-DSA - Purpose is to solve the crypto agility issue. It is really a solution for agility without modifying the interface. This is the debate we need to have in public.

ML-DSA Private Key format

• The Private key format should include words about how to adjust the API for the seeds. Pull request.