On Fri, Jan 23, 2026 at 10:06 PM, John Levine <[email protected]> wrote:

> 1. Root servers SHOULD offer open AXFR over TCP* (perhaps updating RFC
> 7720).
>
> It appears that Wes Hardaker <[email protected]> said:
>
> Zone file over HTTP has a number of advantages, according at least the
> views of a few people that I've talked with (and they should speak up
> here). One obvious one is negotiated compression (mentioned in the
> document) and potentially better global load balancing infrastructure.
>
> The advantage of AXFR is that DNS servers know how to use it to keep a
> copy of a zone up to date.
>
> The advantage of http is that we know how to use it to distribute files at
> scale -- that's a CDN.
>
> It is not obvious to me which will be more difficult to set up in the
> short run and maintain in the long run, create AXFR CDNs, or add features
> to DNS servers to fetch updated zones via http.
>

I'll note that both the Knot (
https://knot-resolver.readthedocs.io/en/stable/modules-prefill.html)
and Unbound (
https://unbound.docs.nlnetlabs.nl/en/latest/manpages/unbound.conf.html#unbound-conf-auth-url)
example configurations use already HTTPS to perform this functionality, so
they at least already support this.


(I realize that there are ways to fake the latter, but we want something
> that doesn't need a DNS expert to set up.)
>
> ICANN has two public AXFR servers at xfr.cjr.dns.icann.org and xfr.lax.
> dns.icann.org. How about asking them what their experience has been,
> how's the load, how hard is it to manage, how have they dealt with the
> sorts of attacks that people make on public servers.
>

Wes and I have already had this chat with IANA (in the form of Kim); I'll
let him chime in here.
W


> R's,
> John
>
> _______________________________________________
> DNSOP mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to