On Mon, Sep 11, 2023 at 09:13:53PM +0900, AJITOMI Daisuke wrote:
> Of course, as an implementer, I support the a la carte (polymorphic)
> approach.
> 
> However, I respect the outcome of this consensus call and I'm ready to
> contribute to moving this spec forward.

Same.


Since the consensus seems to be ciphersuites, what I think would
address my most major concerns:

- Make it easier to define new ciphersuites that work the same way
  as existing ones than ones that somehow work differently (most
  probably via naming of things).
- Write guidance for selection of ciphersuites.
- Register all ciphersuites according to the guidance.
- Get the compact curve stuff to HPKE KEM registry somehow (since a
  ciphersuite that warrants one-byte identifier uses one of those
  KEMs).
- Register crv(OKP) values for the X25519/Kyber hybrid and the
  compact curve stuff (with trivial mapping to HPKE key formats,
  like x25519 and x448 have currently).


> > *Von:* COSE <[email protected]> *Im Auftrag von *Ivaylo Petrov
> > *Gesendet:* Dienstag, 5. September 2023 19:31
> > *An:* cose <[email protected]>
> > *Betreff:* [COSE] Consensus call on COSE HPKE algorithm representations
> >
> >
> >
> > Dear all,
> >
> > We had long discussions on the mailing list (see [COSE] COSE-HPKE and the
> > Single Algorithm Discussion (ietf.org)
> > <https://mailarchive.ietf.org/arch/msg/cose/_GQXXpltAgXNozXIFC9z1yY1CCM/>)
> > and a summary presentation at the last IETF meeting (see
> > slides-117-code-updated-hike-summary-notes-from-hannes (ietf.org)
> > <https://datatracker.ietf.org/meeting/117/materials/slides-117-cose-updated-hpke-summary-notes-from-hannes>
> > and COSE IETF 117 - HedgeDoc <https://notes.ietf.org/notes-ietf-117-cose>)
> > about the question whether algorithm combinations should be presented in an
> > “a la carte” style or as a “ciphersuite” in draft-ietf-cose-hpke (“A la
> > carte” corresponds to “polymorphic” in
> > https://www.ietf.org/archive/id/draft-jones-jose-fully-specified-algorithms-01.html
> > and “ciphersuite” corresponds to “fully specified”). On the mailing list
> > and during the face to face meeting there was a strong consensus to use the
> > ciphersuite (fully specified) approach.
> >
> > The chairs would like to confirm this decision. You have 2 weeks till
> > Tuesday, the 19th of September to provide your final feedback. Please
> > reply-all indicating your preference for either:
> >
> >    - ciphersuite (fully specified) approach
> >    - A la carte (polymorphic) approach
> >
> > The authors will be then instructed to update the COSE-HPKE working group
> > draft accordingly.
> >
> > Best wishes,
> > Ivaylo on behalf of the COSE chairs




-Ilari

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

Reply via email to