Hi Philip,

Thanks a lot for this, good catch! I've committed the following change:

OLD
   The reduction should be in effect at least until the previous DS
   record set has expired from caches, that is, the period during which
   the low-TTL is applied should exceed the normal TTL value.  The
   routine re-signing of the DS RRset (usually after a few days)
   provides a convenient opportunity for resetting the TTL.  When using
   EPP, the server MAY advertise its TTL policy via the domain <info>
   command described in [RFC9803], Section 2.1.1.2.

NEW
   The reduction should be in effect at least for a couple of days and
   until the previous DS record set has expired from caches, that is,
   the period during which the low-TTL is applied typically will
   significantly exceed the normal TTL value.  When using EPP, the
   server MAY advertise its TTL policy via the domain <info> command
   described in [RFC9803], Section 2.1.1.2.

Best,
Peter


On 1/29/26 18:10, Philip Homburg wrote:
This message starts a WG
Last Call for:  draft-ietf-dnsop-ds-automation-02

I just re-read the draft and I think it is in good shape. I have one
minor issue.

In Section 2.2.2, the draft requires lowering the TTLs of changed DS RRsets.
That is good.

The purpose of this is to allow the operator of the child zone to quickly
change the DS RRset if something goes wrong. This means the short TTLs
should stay for a period long enough that the operator can detect a problem
can correct it.

The draft ties the period during which the TTL is lowered to the TTL of the
previous version of the DS RRset.

For example, suppose the normal TTL of the DS RRset is one hour. Then a strict
interpretation of this text would allow an operator to switch back to the
higher TTL after one hour. Suppose the the child's signer changes the DNSKEY
more than one hour after the DS RRset has changed. In that case, the
child has no benefit from the lower TTL (unless just publishing the new
DS RRset would immediately break something).

This is made worse by the fact that scanners run asynchronously, so the
operator of the child zone may not be aware when the new DS RRset will be
introduced.

So I think "The reduction should be in effect at least until the previous DS
record set has expired from caches, that is, the period during which the
low-TTL is applied should exceed the normal TTL value."

should be changed to something like "The reduction should be in effect at
least for a couple of days and until the previous DS record set has expired
from caches, that is, the period during which the low-TTL is applied should
exceed the normal TTL value."

I now notice that the next sentence ("The routine re-signing of the DS RRset
(usually after a few days) provides a convenient opportunity for resetting
the TTL.") seems to assume very specific signing model. For our new signer,
the signer has no clue about the policy for DS TTLs, it signs zones. But the
software that generates unsigned versions of a zone has no clue when the signer
would refresh signatures. So maybe this sentence should just be removed.


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

--
Like our community service? 💛
Please consider donating at

https://desec.io/

deSEC e.V.
Möckernstraße 74
10965 Berlin
Germany

Vorstandsvorsitz: Nils Wisiol
Registergericht: AG Berlin (Charlottenburg) VR 37525

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

Reply via email to