Mats Dufberg <[email protected]>
writes:

> Parent is not authoritative of the NS in the delegation. The same with
> any glue address records on or below the delegation point. The parent
> does not sign non-authoritative records.

The odd thing about this situation, as the above well states the
history: the parent isn't authoritative for the NS and glue, but yet it
still exists in their zone.  So it's this hybrid case where they publish
data, and seem authoritative to the DNS client asking for where to talk
to the parent's child, but it's the client's fault if the parent is
serving incorrect data (IE, the child/parent are out of (c)sync).

The history basically says "we believe that there is only one
authoritative source of this data", but the reality says "there may be
only one source of authoritative data, but multiple publishers may
distribute copies of it that may or may not be correct".

In the end, there are two sources of authority:

- who creates the records
- who is distributing the records

IMHO, it would be helpful if the distributors would stand by their
statements of publication, even if they didn't actually create the
records.  I don't think signing the distributed records would change the
behavior of whether or not clients cached them or whether clients
believed who was authoritative (IE, I don't think resolvers are making
caching decisions based on whether something was signed or not).
-- 
Wes Hardaker
USC/ISI

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

Reply via email to