On Fri, Aug 7, 2020 at 6:28 PM Puneet Sood <[email protected]> wrote:

> On Fri, Aug 7, 2020 at 9:22 PM Rob Sayre <[email protected]> wrote:
> >
> > On Fri, Aug 7, 2020 at 6:18 PM Puneet Sood <puneets=
> [email protected]> wrote:
> >>
> >> I think this is worth doing.
> >
> >
> > I agree. The part that I worry about is the computational cost of
> reestablishing links after an outage. Is there a way to model this?
> (Perhaps this work has already been done)
>
> Outage of the authoritative or resolver? Can you explain your concern
> in more detail?
>

I meant an outage of the authoritative, or some substantial network link
connected to it.


> For a resolver an outage is similar to a cold start - the cost of
> connection setup will be amortized over all the queries done during
> the lifetime of the sessions.
>

I agree that the recursive resolver's reliability concerns don't change
much with encrypted DNS.

Assuming this traffic is encrypted, which I am in favor of, the CPU load on
the authoritative server will increase after an outage or network problem.

Is this already factored in?

thanks,
Rob
_______________________________________________
dns-privacy mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dns-privacy

Reply via email to