Re: [TLS] [therightkey] Fwd: Improving EV Certificate Security

[Ben Laurie <[e-mail]>]

On 26 September 2013 15:44, Joseph Bonneau <[e-mail]> wrote:
>
>> >> Not sure what the question is - as the doc says, the list will be
>> >> constructed from the logs...
>> >
>> >
>> > I think I read it incorrectly as "without an embedded CT from *any*
>> > qualify
>> > logs" instead of "from all qualifying logs." Now I can see how the
>> > whitelist
>> > is created, but I'm less clear on what the intention of it is. Is the
>> > assumption that some certs will be issued with more than zero but fewer
>> > than
>> > three SCTs (proposed to the minimum acceptable in the "Qualifying
>> > Certificates" section) and you'd like to whitelist such certs during the
>> > rollout period?
>>
>> Ah. So, all existing certs do not have embedded SCTs. So, we either
>> wait until all existing certs expire before we can enforce CT, or we
>> whitelist the unexpired certs.
>
>
> I think I'm back to my original question now :-) How do all the existing
> certs get into the CT log? (at which point building the whitelist is easy)
> Is the onus on EV users then to log their old certs or face failing in
> Chrome? Or do EV-issuing CAs have sufficient records of what they've issued?

My assumption is that this is true. If it's not, then CAs need to tell us :-)

> Or is this something that we're hoping can be (at least mostly) achieved by
> network observation?

This is also likely.