Re: [TLS] Ideas for TLS 1.3: Server key continuity and certificate reporting

Hi Kai,

Have you seen TACK?  It does something similar but with self-managed
signing keys, and has good procedures (IMO) for handling key and cert
changes and pin lifetimes.

http://tools.ietf.org/html/draft-perrin-tls-tack-02
http://tack.io/

Paul is correct there's more discussion about pinning on websec.  But
I agree with you this approach should be applied to SMTP, IMAP, etc.

I'll also note that the bottleneck for ideas like this isn't IETF
acceptance, but rather browser willingness to spend effort on
experimental proposals (there are sites interested).

Most browser experimentation in this area is going into CT at the
moment, it seems to me.

Trevor

On Fri, Jan 10, 2014 at 12:28 PM, Kai Engert <kaie@kuix.de> wrote:
> On Fr, 2014-01-10 at 12:12 -0800, Paul Hoffman wrote:
>> These ideas (and the followups from Alyssa) are interesting, but I
>> suspect that they belong in the websec WG instead of here, unless you
>> truly mean them to apply to every application that uses TLS. I don't
>> see how they would work, for instance, with SMTP servers that do TLS.
>
> I'm motiviated to find ways to makes these ideas work with any TLS
> connection, completely independent of the encapsulated application
> protocol.
>
> I think a SMTP/TLS client could equally cache the previous cert and
> potentially report it to the TLS server on a subsequent connection.
>
> Where do you see differences that would cause it not to work with
> protocols other than http?
>
> Regards
> Kai
>
>
> _______________________________________________
> TLS mailing list
> TLS@ietf.org
> https://www.ietf.org/mailman/listinfo/tls