Re: [TLS] Inclusion of OCB mode in TLS 1.3

[Nikos Mavrogiannopoulos <[e-mail]>]

On Mon, 2015-01-19 at 15:06 +0100, Aaron Zauner wrote:

> >>> I think it would make sense to also include pure PSK (no DH) suite for very
> >>> constrained devices that can't afford the asymetric crypto.
> >> Makes sense. So your suggestion is to swap or add these
> >> ciphersuites? I have no real-world data on the usage of TLS with
> >> pre-shared keys, nor it's deployment. I can only guess, would be
> >> interesting to hear the opinion of embedded development people on
> >> that subject.
> > I don't believe you can second guess here. There can be many valid
> > use-cases unknown to anyone in the WG, or IETF, as PSK is often used in
> > protocols which are non-standardized or standardized outside IETF.
> I'm not trying to second guess anybody/anything. I'd like to have some
> common ground. If I take the lead from RFC6655 (AES-CCM) that'd be:
> CipherSuite TLS_PSK_WITH_AES_128_OCB
> CipherSuite TLS_PSK_WITH_AES_256_OCB
> CipherSuite TLS_DHE_PSK_WITH_AES_128_OCB
> CipherSuite TLS_DHE_PSK_WITH_AES_256_OCB
> Plus:
> >>>> TLS_ECDHE_PSK_WITH_AES_128_OCB
> >>>> TLS_ECDHE_PSK_WITH_AES_256_OCB
> Does this make sense for you?

Yes that looks reasonable.

regards,
Nikos