Re: [Trans] DNSSEC also needs CT

[Ben Laurie <[e-mail]>]

On 13 May 2014 07:34, Nico Williams <[e-mail]> wrote:

> On Tue, May 13, 2014 at 12:22 AM, Joseph Bonneau <[e-mail]>
> wrote:
> >> Is CT intended to be run all the way from the root to the CAs furthest
> >> from the root?  I didn't think it was, and if it is, please tell me.
> >
> > Yes, it is. The goal of CT is that browsers will eventually reject any
> > end-entity TLS certificate that doesn't have an SCT. I believe this is
> true
> > regardless of the number of intermediate CAs in the cert's path to a
> trusted
> > root. There's an exception for trust anchors manually added to the
> browser
> > to accommodate private CAs, but essentially all certificates that
> standard
> > browsers will accept out of the box must be logged.
>
> Ah, yes, and actually we should want the same for DNSSEC.  The problem
> then becomes privacy.  But I think we can achieve that by not logging
> names, just hashes of the relevant RRsets, no?  Since public keys will
> generally be part of the relevant RRsets this won't help zone
> enumerators.
>

Is it necessary to log anything other than keys? My base assumption was no:
if the keys are as expected, then all records signed by those keys can be
trusted. If someone wants to publish RRsets that are other than the one the
true domain owner wants to publish, they necessarily have to inject a key
they control, which becomes apparent from the logs.

-- 
Certificate Transparency is hiring! Let me know if you're interested.