Re: [Ntp] NTS Pools

[David Venhoek <[e-mail]>]

Dear Miroslav,

For now, I think we just have to agree to disagree on that point. Yes,
the approach I intend to experiment with is computationally a lot more
expensive than the current pool, and it may very well be that
eventually this turns out to be a sufficiently large issue to kill the
idea. However, I think it should be doable and believe there are
sufficient advantages as setting up time servers around the world with
good quality time is also a significant challenge.

Kind regards,
David Venhoek

On Thu, Feb 29, 2024 at 10:34 AM Miroslav Lichvar <[e-mail]> wrote:
>
> On Thu, Feb 29, 2024 at 09:35:05AM +0100, David Venhoek wrote:
> > Rationale for the NTS-KE load balancer approach:
> > Given these goals, I myself find the NTS-KE load balancer approach the
> > most attractive. Requirement 1 limits the options to either Shared
> > domain name, NTS-KE generating cookies or NTS-KE load balancer.
>
> The trouble with the load balancing approach is that it's not
> practical, at least I still don't see a good use case for it.
>
> As your table shows, the computational cost of load balancing is
> higher than the cost of the load-balanced service itself. It's better
> to use all resources to serve NTS-KE and NTP directly. It's simpler,
> and provides better performance. With or without load balancing, the
> servers to which clients are making TLS connections need to share the
> hostname.
>
> --
> Miroslav Lichvar
>