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

[Mark Delany]

John L wrote:
>> percentages are "normal" vs. "unusual", but my cursory look a
>> long time ago suggested that it met the 80-20 rule.
> 
> You are certainly correct that most zones are pretty flat, but this
> sounds like a DOS attack waiting to happen, send out junk with long
> bogus addresses

I'm just raising this as a discussion point; what if we said that the 
SSP record must (at least) exist at the registry cut-point?

It's not particularly pretty, but you (only) need about a 1,000 entry 
database to define all the registry cut-points today. I know the size 
because we've built this sort of database for other reasons. I think 
SpamAssassin has something similar as well.

That "root" SSP record could tell us max-depth within it's balliwick, if 
that's of use.


I'm kindof a fan of the registry cut-point because that segues nicely 
into a responsible and hopefully knowable entity.


Mark.