Re: [TLS] AD comments on draft-ietf-tls-rfc3546bis

Wright, Timothy, VF-Group wrote:
> No hollering forthcoming from me!

No objection from me, either. I think it's unnecessary to change the
spec from RFC 3546.

It's worth pointing out that:

  - ciphersuite-specific extensions are useful only if the TLS *client*
    has extension options that depend on the ciphersuite. If the server
    has such options, no additional mechanism or conventions are needed,
    since the server can choose the ciphersuite and its extension replies
    together.

    Note that the issue of hardware that only supports certain ECC curves
    *mainly* affects servers, not clients.

  - not discussing ciphersuite-specific extensions in RFC3546bis will
    not prevent ECC (or anything else) from using a ciphersuite-specific
    extension. It just defers the issue of how to define such extensions
    to the documents that actually use them.

(While I remember, although it's not very important: Simon's proposal in
Message ID <05d801c58d6d$36c3f210$3200a8c0@simon> would have had a minor
incompatibility with RFC 3546 -- it depended on sending multiple extensions
of the same type, and section 2.3 of RFC 3546 explicitly disallows that.
A spec that wanted to define ciphersuite-specific extensions would have
to encode them all into a single extension to avoid this. I don't think
that's a problem, though.)

-- 
David Hopwood <david.nospam.hopwood@blueyonder.co.uk>

_______________________________________________
TLS mailing list
TLS@lists.ietf.org
https://www1.ietf.org/mailman/listinfo/tls