> Petr paek wrote:
> > That means the proposed protocol would depend on non-minimal authority
> > sections. I thought the trend nowadays is the opposite, i.e. providing
> > more and more minimal answers.
> 
> Indeed! Condition 1 in 4.1 seems hard to hit, in current operational
> reality.

To bring this proposal more in line with DELEG I wonder if it should be 
changed to the following:

Suppose we have
$ORIGIN com
customer in NS ns.provider.net.

Then a resolver will issue queries for ns.provider.net./A and
ns.provider.net./AAAA

What we could do is specify that the resolver issues an addition query of
the form _dns.ns.provider.net./SVCB And any SVCB result is processed
according RFC 9461.

Advantages are:
- no change to authoritative software
- can be introduced locally without support from the root or TLDs.
- is secure if provider.net. is signed.

Disadvantages:
- does require an extra query from resolvers
- extra load on authoritative servers

The case where the delegation requires glue (i.e. customer NS ns.customer) is
explicitly excluded. Though resolvers that implement revalidation could 
try to fetch _ns.customer/SVCB during revalidation.

_______________________________________________
DNSOP mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to