On Mon, Nov 28, 2022 at 10:39:19AM -0800, Laurence Lundblade wrote:
> All I think that is important right now is that we know how we’re
> going to label mode_auth separate from mode_base.
> 
> I think registering two COSE algorithm IDs, HPKE_BASE and HPKE_AUTH
> is workable and probably preferable to an HPKE mode parameter. I
> think this is what Ilari suggested, but I’m not sure.

Right.


> I think the general case for mode_auth is quite strong independent
> of a use case like firmware encryption because there are so many
> use cases for signing+encryption in COSE, but we can work on that
> in another draft.

Yes, do it in another draft if you are to do it at all.

(The security considerations stuff is very much nontrivial.)


> One question below.
> 
> > On Nov 28, 2022, at 10:28 AM, Hannes Tschofenig <[email protected]> 
> > wrote:
> > 
> > Laurence,
> > 
> > I would prefer not to design the HPKE auth mode without a scenario to 
> > motivate it.
> > 
> > Ciao
> > Hannes
> > 
> > -----Original Message-----
> > From: COSE <[email protected]> On Behalf Of Ilari Liusvaara 
> > 
> > AFAICT, Neither PR9 and PR10 preclude reusing the HSI codepoint
> > for this sort of extension in another draft.
> 
> What’s an HSI code point?

HPKE Sender Info. That is, the algorithm parameter all the HPKE stuff
goes into.



-Ilari

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

Reply via email to