Regarding:
2.6. Non-existent Domains
For DMARC purposes, a non-existent domain is a domain for which there is an 
NXDOMAIN or NODATA response for A, AAAA, and MX records.

Comments:
- sometimes a domain used for mailing purposes does not have a MX record - I am 
not sure if 'and' is appropriate word here,
- question: what if there is a CNAME record?,
- email receivers could/should perform reverse DNS lookup, however they do not 
- as a result, an email from (both: Envelope and Mail from), is accepted by the 
MTAs,
- today, an email ‘from’ (Envelope and Mail) NXDOMAIN is accepted by vast 
majority of MTAs, and SPF check result is “spf=neutral” (in general), same with 
non-existent sub-domains (even if there is DMARC in place a message from 
non-existent sub-domain could be successfully delivered).

There is a solution for non-existent subdomains: Wildcard SPF. Wildcard SPF 
covers sub-domains (if there is no other RR for such sub-domain), and (in 
general) it works with DMARC.
For example:
*.example.com IN TXT v=spf1 -all
together with
example.com IN TXT v=DMARC1; p=reject;
Covers each and every NX-sub-domain, and it works pretty well.

Currently proposed solution, even with the 'np' tag, may not work. It could be 
rather fatally flawed.

That being said, we should consider (for PSDs) a solution similar to a Wildcard 
SPF, if we want PSD-DMARC work as it should.

And, because a Wildcard SPF is a TXT - not A, AAAA, MX - record, and it does 
not mess with the definition (2.6. Non-Existent Domains).


Sincerely,
Mnpn.

‐‐‐‐‐‐‐ Original Message ‐‐‐‐‐‐‐
On Saturday, September 28, 2019 11:38 PM, Scott Kitterman 
<[email protected]> wrote:

> WGLC started on 2019-06-26. It ended 2019-07-17.
>
> 2019-06-26: WG Secretary called for additional discussion on three topics:
>
> 1.  What further context is needed in the introduction
> 2.  If explicit call outs to ICANN/limited operator capacity to implement are
>     needed
>
> 3.  If an np= tag is needed to allow PSD functioning for only NXDOMAINs
>
>     By the end of the WGLC period, after a slow start, there was a vigorous
>     discussion on all these issues and (with a few days post-WGLC for the np= 
> tag
>     discussion to finish) moved towards what was in my judgment a reasonably 
> solid
>     consensus. All issues that people brought up were addressed by the WG
>     (including inclusion of some recommended editorial changes made before 
> WGLC
>     that hadn't made it into the draft).
>
>     A new revision addressing the WGLC discussion (-05) and document shepherd
>     comments was uploaded 2019-07-27.
>
>     Post WGLC discussion:
>
>     2019-07-22: A question was brought up if the draft should include hard
>     requirements to forbid mail from non-existent domains. The WG concluded 
> that
>     independent of the wisdom of such a rule, it was out of scope for the WG.
>
>     2019-07-27: A set of editorial nits were provided against -05.
>
>     2019-07-31: Additional document shepherd feedback based on revisions 
> provided
>     in -05 and earlier comments.
>
>     2019-08-09: New revision (-06) published fixing editorial nits and
>     incorporating additional changes based on document shepherd feedback.
>
>     2019-08-10: Comments on -06 from Alessandro Vesely, but no suggested 
> changes.
>
>     2019-08-13: Comments on the general PSD DMARC concept from Dave Crocker.
>
>     2019-09-03: WG Chair feedback on 2019-08-13 comments, some discussion 
> ensues.
>
>     2019-09-09: Discussion concludes. It appears to me that the upshot of the
>     discussion is that the Section 2.3 Longest PSD definition needs to be 
> updated.
>     All other issues have been addressed.
>
>     2019-09-24: Alessandro Vesely reports having implemented PSD DMARC in
>     zdkimfilter for Courier MTA using an alternative data format for PSD DMARC
>     participants based on the PSL format.
>
>     Based on the post-WGLC discussion, I believe an updated draft before IETF 
> last
>     call is appropriate with three changes:
>
> 4.  Updated Longest PSD definition:
>
> > 2.3. Longest PSD
> > The longest PSD is the Organizational Domain with one label removed.
>
> 2. Addition of the PSL data format to Appendix B (it would be B.3). I
> haven't drafted text yet, but I don't expect it to be controverisial.
>
> 3. Add zdkimfilter to Appendix C (also didn't make text yet).
>
> Unless someone tells me otherwise, I plan to go ahead with those changes.
>
> Scott K
>
> dmarc mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/dmarc

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

Reply via email to