Re: [Ntp] Splitting the Roughtime draft?

On Sun, Jan 31, 2021 at 1:06 AM Hal Murray <hmurray@megapathdsl.net> wrote:
>
>
> marcus@dansarie.se said:
> > I think we need to be very clear about the fact that all trust in Roughtime
> > is rooted in the long-term keys and that they are expected to be valid for a
> > very long time indeed.
>
> How long is "very long"?

CT logs are active for one year of certificate expiry, but that's due
to log growth causing issues. Roughtime keys could be used for
multiple years, although shorter lifetimes would force updates on
clients. I'm open to suggestions.

>
> I've been trying to figure out how to use Roughtime to get NTS off the ground
> when the time isn't known/trusted yet.  If it needs long term keys, is there
> any advantage to long-term Roughtime keys as compared to trusted certificates
> with a long lifetime?

You shouldn't bake in ecosystem.json and expect this to work securely.
The benefit of roughtime is that malfeasance is detected, hence
updates can remove bad servers without much trouble, and it readily
provides enough accuracy for X509 validation. Set and forget with a
limited number of servers doesn't enable those ecosystem benefits.

What we observe in the browser world is that even 24 hour windows of
validity are problematic for clients due to inaccuracy, but a few
seconds of accuracy is just fine for X509 validation.

Sincerely,
Watson Ladd

>
> --
> These are my opinions.  I hate spam.
>
>
>
> _______________________________________________
> ntp mailing list
> ntp@ietf.org
> https://www.ietf.org/mailman/listinfo/ntp

--
Astra mortemque praestare gradatim