On Wed, May 10, 2023, at 11:57, Hollenbeck, Scott wrote:
> RFC 5731 contains text that has led some domain name registrars to 
> adopt an operational practice of re-naming name server host objects so 
> that they can delete domain objects.  

There aren't any other real working methods unfortunately :-(
Except in cases of registry fine into letting the host delete go through even
if still linked to other domain objects.

> There's a risk in renaming the host object, though. If the host is 
> renamed using a domain that isn't currently registered, such as 
> ns1.randomfoo.example, it becomes possible for someone to gain DNS 
> resolution control of ns1.randomfoo.example by registering 
> randomfoo.example and creating ns1.randomfoo.example. 

If you stay in same registry, the registrar can just register once for all
a "dummy" name like `this-is-really-junkyard.example` and then it renames
all hosts below that name. As he keeps registration of the base name, no bad
things can happen. With some cost, yes, the cost of one domain only, no matter
how many nameservers renamed there.

I think the problem is really in the thinking of "let us just use some random
non existing name and hope it will work", which created problems as well as the 
TLD
level when people decided to just arbitrarily use `.dev` locally. Also anyway 
some
registries might check that the name resolves at least and hence won't allow
non existing domains.

There is also the idea of using names below `in-addr.arpa` or other valid 
suffixes
but where obviously no registration happens. Some very widely known names
do also provide wildcard behavior to make any name suddenly resolve even if 
really
no nameserver there, and the base name will never go out of existence.

> Is this a topic we should address?

Do we want to address the generic problem of how to break associations
(discussed since EPP existed or even before but as yet no solutions besides
renaming or registry forcibly broking the link), or just address
what the RFC says to help people do things differently/with more caution?

I think the second case can be done, but not really the first.
And even so, it is maybe out of IETF scope as it is more an operational matter
than really a problem in the protocol itself.

-- 
  Patrick Mevzek
  [email protected]

_______________________________________________
regext mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/regext

Reply via email to