On Fri, Sep 01, 2023 at 06:58:04PM +0300, Ilari Liusvaara wrote:
> On Thu, Aug 31, 2023 at 02:12:00PM +0000, John Mattsson wrote:
> > Hi all,
> > 
> > --------------------------------------------------------------
> > This group is chartered to work on the following deliverables:
> > 
> > - Document registering cryptographic algorithm identifiers that fully
> >   specify the cryptographic operations to be performed.
> > - Document describing the use of the NIST algorithm ML-KEM in JOSE.
> > - Document describing the use of the NIST algorithm ML-DSA in JOSE.
> > - Document describing the use of the NIST algorithm SLH-DSA in JOSE.
> > - Document describing the use of the NIST algorithm NL-DSA in JOSE.
> > - One or more documents describing the proper use of algorithms.
> >   These algorithms must meet the requirements outlined above.
> > --------------------------------------------------------------
> 
> I think there should also be work item for sound signature pre-
> hashing. Doing pre-hashing in naive way is cryptographically
> *unsound*.

Oh, and there should be general framework for using KEMs in JOSE,
and ML-KEM stuff should be instance of that.

Such framework can be done by taking ECDH-ES, doing minimal
modifications required for things to work (basically using a new header
parameter instead of "eph" and changing some wording from ECDH to KEM)
and then rethinking the algorithms a bit (I think SHAKE256[1] and
SHAKE256+AES256KW[2]).

If one wants a test KEM for it, I think X25519+Kyber768 using KEM
combiner would make a good one. That stuff has effectively been
vetted by CFRG.


(Using HPKE is not a good idea because HPKE and JWE do not mix.)



[1] E.g., one needs SHAKE256 (or something extremely close) anyway
for ML-KEM.

[2] Post-quantum!




-Ilari

_______________________________________________
jose mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/jose

Reply via email to