Reyk Floeter <[email protected]> writes:

> 2. draft-ietf-ipsecme-safecurves-00 "Curve25519 and Curve448 for IKEv2
> Key Agreement": 
>
> iked supports Curve25519 since August 2014.  I compared the draft with
> our implementation and it matches the described application.  We're
> currently using an xform type 4 id from the private space, but we hope
> to switch to an official id once it is assigned by IANA.
>
> Curve25519: We hope that the addition of Curve25519 will be finalized
> in an RFC.  We also hope that it will be added as SHOULD or MUST in a
> future update of RFC 4307.

Thanks for feedback!

> Curve448: We discussed Curve448 internally and we are not going to add
> the curve at this point.  Compared to Curve25519 it had much less
> cryptographer review, provides a less proven and more complex
> reference implementation, and is basically "too new".  We also think
> that Curve25519 provides sufficient security at present.  Afaik, there
> are no plans to add Curve448 in OpenSSH or LibreSSL either.  The draft
> should clarify Curve448 as an optional extension for potentially
> higher security levels or omit it completely.
>
> We're not fully convinced of the meaningfulness of the security levels
> in this regard.  After all, Curve25519 is a solid choice.

I can understand these concerns for an implementer.  I think there is
little harm in the draft specifying how to use Curve448 for potential
future use though.  I agree normative language or mandatory-to-implement
recommendations for Curve448 may be premature, at least until other
implementers speak up in favor of implementing it.

/Simon

Attachment: signature.asc
Description: PGP signature

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

Reply via email to