[dns-operations] Limit on Name Servers & their IPs for a sub-domain

[George Michaelson]

11 servers. 11 labels. 11 named entities, each with a single IPv6 binding..

but in terms of 'distinct, autonomous, reachable entities, capable of
supplying the root zone' .. its kind of moot what is, or is not a
magic number here. What fits in the packet feels like a really bad
constraint in some ways. If we move off UDP as a core dependency
(which btw, is kind of a given if we go IPv6 only, because UDP+IPv6 is
a horror story all of its own because of host-only fragmentation and
pMTU) then why should we be subject to a proscriptive limit? What if
DOH succeeds in volume and DNS priming moves to using TCP via CDNs
instead?

I can recall somebody (John Crain? David Conrad?) saying back in a
Dallas IETF (the one with the flood) that if you did anycast properly,
you could do the root with one label.

I'm probably off in the room-of-one-person again, but I just don't
feel like this number is either that important, or a very real
constraint right now.

-G

On Fri, Jan 12, 2018 at 10:32 AM, Brett Watson <brett at the-watsons.org> wrote:
> On Jan 11, 2018, at 3:54 PM, Mark Andrews <marka at isc.org> wrote:
>
> If we ever get to the state where the roots are IPv6 only, 11 servers will
> fit.
>
>
> First I’ve seen anyone do the math on this, and wondering did anyone think
> of this and what’s the impact if in fact some day roots are v6 only? (I
> certainly never thought about it)… maybe it’s come up on various lists and I
> missed it but I don’t recall if so.
>
> -b
>
> _______________________________________________
> dns-operations mailing list
> dns-operations at lists.dns-oarc.net
> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
> dns-operations mailing list
> https://lists.dns-oarc.net/mailman/listinfo/dns-operations
>