On Mon 21/Oct/2024 17:51:00 +0200 Todd Herr wrote:
Issue is here - https://github.com/ietf-wg-dmarc/draft-ietf-dmarc-dmarcbis/issues/156


That's the DNS tree walk definition. What happens when the first identifier looked up in step 2 has psd=y? The current text says to continue walking to the next parent. However, what if the parent domain is in turn a PSD?

I report Daniel-t's comments on Github:

-----
Our scenario is a PSD, which is a child of a PSD. Our immediate parent doesn't yet publish a DMARC policy however hopefully will in the not too distant future though, I can't guarantee that they will implement psd=y when they do. There are about 250 PSDs in the publicsuffix.org list which will be in a similar situation to ourselves.

If we (a PSD) publish psd=n, we'd override all of our sub-domains DMARC policies (all controlled by separate organisations) until they update their own to include psd=n. This would be a nightmare scenario for us.

If we dont publish psd at all, we'll still override the sub-domain policies (shortest number of tags) until they add psd=n. Again a bad outcome.

Our only option with the new discovery process is psd=y, as this will cause the existing sub-domain policies to be respected as is, while providing a policy for our top level domain and the children which dont have a policy for themselves.

There needs to be a way for legitimate emails coming from a PSD (with psd=y) to be handled. The easiest way is as proposed by @toddherr above and accept that when the first query returns psd=y it is the organisational domain.
-----

Note that Section 4.10.2 excludes the case that psd=y can be found where the tree walk started:

   2.  If a valid DMARC Policy Record, other than the one for the domain
       where the tree walk started, contains the "psd" tag set to "y"
       ("psd=y"), the Organizational Domain is the domain one label
       below this one in the DNS hierarchy, and the selection process is
       complete.


Best
Ale
--




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

Reply via email to