[ietf-dkim] New issue: Upward query vs. wildcard publication

[Michael Thomas]

Dave Crocker wrote:
> 
> 
> Eliot Lear wrote:
>> Publish a record at the zone level.  If no other record exists, use 
>> that.  If that record doesn't exist, stop.  This is a cheap form of 
>> wildcarding.  it requires at most two queries, which while unpleasant 
>> for some purists avoids people having to deal with wildcards, which 
>> themselves are a pain.
> 
> 
> Given that zones are administrative constructs for use by operators, and 
> are not intended to be visible to client DNS activities -- and well 
> might not be visible, no matter the intent -- then how does the upward 
> tree-walk know when to stop?

In the general case, it doesn't. However, it's possible that in some
cases where the stars align that the tree walk could be made into a
two message exchange affair. It seems to me that it's worth digging
deeper on this point as it seems like it's actually the common case
that the SOA for a given subdomain is the same as the parent of that
subdomain. Sorry to be a bit vague.. these pages are still getting
swapped back into memory.

		Mike