On Sun, Nov 27, 2022 at 2:03 PM Paul Wouters <[email protected]> wrote:

> On Wed, 23 Nov 2022, Tero Kivinen wrote:
>
> > I.e., the main reason being that group 2 was only MUST algorithm
> > before, and moving it from MUST to MUST NOT while we do not have any
> > oher algorithms as MUST was considered bad. Also the group is formed
> > inin a deterministic way which should not make it possible that the
> > group is created to be weak from the beginning.
>
> Right, so if we were to update 8247 (post ikev1 historicness), we should
> do:
>
> * AES_GCM_16 from SHOULD to MUST
> * AES_CBC from MUST to SHOULD
> * 3DES from MAY to MUST NOT
>
> * PRF_HMAC_SHA1 from MUST- to SHOULD
>
> * AUTH_HMAC_SHA1_96 from MUST- to SHOULD
>
> It is tempting to speed it up to SHOULD NOT though SHA1 may not be a big
issue there.

> * 1024-bit MODP Group from SHOULD NOT to MUST NOT
> * 1536-bit MODP Group from SHOULD NOT to MUST NOT
>
> Arguably, the SHA1 entries could go to MUST NOT because no one should
> have ever had a need for those for IKEv2.
>
> Paul
>
> _______________________________________________
> IPsec mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/ipsec
>


-- 
Daniel Migault
Ericsson
_______________________________________________
IPsec mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/ipsec

Reply via email to