./DS is NOERROR NODATA.  This is RFC described behaviour to handle legacy 
servers
and to handle the “grandparent problem”.  “.” is not special here.  If you get a
query for DS at an apex of the zone you return NOERROR NODATA.  The validator 
will
then look the the cut point by using NS queries chopping off labels.


> On 22 Jul 2025, at 18:11, Petr Špaček <[email protected]> wrote:
> 
> Hi all.
> 
> I wonder how to interpret '. DS'/'. DELEG' queries and welcome opinions!
> 
> First problem:
> I did not find algorithm update for RFC 1034 5.3.3. Algorithm which would 
> clearly show that QTYPE=DS is special.
> 
> Secondly it's unclear if RFC 4035 4.3. Determining Security Status of Data 
> somehow should applies to DS responses from root. I guess it should, but how?
> 
> 
> With strict interpretation of 'DS lives at parent' I would argue '. DS' 
> should result in SERVFAIL: No parent for . can be contacted.
> 
> If that does not fly and we don't like SERVFAIL, . DS answer should probably 
> get AD=0 in answer as security status would be Indeterminate - no way to 
> validate and no trust anchor for parent-of-root?
> 
> Needless to say implementations vary in their responses.
> 
> 
> I don't think this is operationally relevant, but it is sorta related to 
> DELEG protocol design which has the same property as DS - it is authoritative 
> at the parent side.
> 
> If there are undefined corner cases for DS we might as well not repeat the 
> same mistakes for DELEG (or we can even fix it in one go ...).
> 
> Thank you for your time.
> 
> -- 
> Petr Špaček
> 
> -- 
> dd mailing list -- [email protected]
> To unsubscribe send an email to [email protected]

-- 
Mark Andrews, ISC
1 Seymour St., Dundas Valley, NSW 2117, Australia
PHONE: +61 2 9871 4742              INTERNET: [email protected]

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

Reply via email to