On 18 May 2015, at 17:08, Ben Laurie <[email protected]> wrote:

> 
> 
> On 18 May 2015 at 15:50, Stephen Farrell <[email protected]> wrote:
> (All those issue tracker mails reminded me of a question
> I'd meant to and had forgotten to ask...)
> 
> At the acme BoF there was some talk of short lived certs.
> The thought was that if acme succeeds then it'd be more
> practical to use certs with a lifetime of a day or so. And
> that might raise a question as to how that'd affect CT or if
> there's an elegant way to support such.
> 
> I don't think this is something that has to be part of the
> current bis RFC necessarily but it'd probably be good to
> get the collective wisdom of the list on the topic.
> 
> So, what do we think of CT when faced with 1 day duration
> certs or similar?
> 
> CT is necessarily after the fact detection anyway. However, 1 day duration 
> certs would make that more painfully obvious!
> 
> That said, if you are issuing 1 day certs, then clearly you are going to be 
> replacing them every day. Presumably you could issue them in advance, so they 
> appeared in the log before they were valid.
> 
> The other thing to consider is the size of the log - I would want a different 
> log for short-duration certs, since it will grow fast, but also most of it 
> rapidly becomes irrelevant and can be ignored.

+1; it seems to me that the sensible way to improve scalability is to segregate 
the high-volume, ephemeral certs from the longer-lived ones as early as 
possible (presumably by checking if “not valid after” is only a day or so 
greater than “not valid before”).

As Ben implies, I think it’s also generally valid to assume that ephemeral 
certs would give rise to a different kind of risk from that of longer-lived 
certs, so the corresponding CT logs would benefit from different 
management/policy.

Yrs.,
Robin 



> _______________________________________________
> Trans mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/trans

Attachment: signature.asc
Description: Message signed with OpenPGP using GPGMail

_______________________________________________
Trans mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/trans

Reply via email to