Re: [IPsec] I-D on Using the ECC Brainpool Curves for IKEv2 Key Exchange

[Johannes Merkle <[e-mail]>]

Andrey,

>
> For any new format / protocol / feature, RFC 6090 + http://tools.ietf.org/html/draft-jivsov-ecc-compact should be a more
> natural path.

Well, RFC 6090 is not always a good reference, at least it is very unfortunate that the signature scheme KT-I is not
fully compatible with ECDSA (Errata 2777).

> I agree that the statement was unclear. I was saying that because the certificates contain only one public key, the
> space saving is only size-of-the-curve+1 byte per certificate. Nonetheless, given that certificates need to be
> pre-processed and validated, and these results are often cached, the overhead of calculation of the 'y' is probably not
> material.

For X.509 certificates, the ANSI format is already well established, so the alternative method is hardly relevant there.
As you said, it is more a suggestion to consider for new specifications.


-- 
Johannes