On Tue, 29 Jul 2025 at 00:11, Neil Madden <[email protected]> wrote:

> On 28 Jul 2025, at 19:07, Ilari Liusvaara <[email protected]>
> wrote:
> >
> > On Mon, Jul 28, 2025 at 06:13:51PM +0200, Filip Skokan wrote:
> >>
> >> As far as the other questions from the presentation to the working
> group go
> >> I believe this is fair summary
> >>
> >>   - Why not Use AKP Slide - Use of AKP doesn't introduce ambiguity, it
> >>   removes it. Not being able to use the same JWK for both modes when
> using
> >>   AKP is a problem that doesn't need solving.
> >
> > It is a operational problem that needs solving.
>
> I really don’t see why this is a problem. If you really need the same key
> to be used for both algorithm variants then simply publish it twice in the
> key set, once with each “alg”.


This duplication increases operational complexity:

   -

   Key rotation and lifecycle management (e.g., revocation) become more
   complex when the same key material is duplicated
   -

   "AKP" would require different "kid" values for each duplicated key

Let's weigh these trade-offs before switching to "AKP".

-Tiru


>
>
> > One way to solve it
> > would be to remove the Direct Key Agreement algorithms, since at least
> > Key Agreement with Key Wrapping is at worst a size penalty, instead of
> > not working at all.
>
> I have some sympathy for removing Direct Key Agreement. (That said, if we
> ever want to bring back HPKE’s AuthKEMs, this seems easier to get right in
> direct mode).
>
> > Then one could convert the algorithm into Key Encryption algorithm so
> > that compact encoding does not need to double-base64url the already
> > large KEM ciphertext in compact encoding.
> >
> > - 1088 bytes double encoded -> 1935 bytes.
> > - 1088+40 bytes encoded -> 1504 bytes.
>
> — Neil
> _______________________________________________
> jose mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
jose mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to