Re: [Bimi] BIMI/DMARC & PSL vs Tree-walk

Are you seeking input from PSL volunteers on this?

-Jothan

On Fri, May 12, 2023, 5:54 PM Brotman, Alex <Alex_Brotman=
40comcast.com@dmarc.ietf.org> wrote:

> Hey folks,
>
> With DMARCbis, the current draft illustrates a method by which the DMARC
> policy is discovered via a tree-walk instead of utilizing the PSL.  I'm not
> trying to hash out the decision for DMARCbis to do that, there's a
> different list for that.  I do have a couple of questions how this relates
> to BIMI.
>
> Assumptions:
> 1) Message is coming from 5322.From domain "marketing.emails.example.org".
> No DMARC or BIMI policy declared here
> 2) There exists a DMARC policy at example.org for p="reject" and a BIMI
> policy at example.org
> 3) There exists a DMARC policy at emails.example.org, p="none", no BIMI
> policy declared here
> 4) All authentication aligns/passes
> 5) 7489 would have found the DMARC policy at "example.org", DMARCbis will
> find policy at "emails.example.org"
>
> Questions:
> 1) Does it matter that the DMARC policy and BIMI policy live in different
> locations?
> 2) Do we care if the DMARC policy is different?  The apex domain is still
> p=reject.  However, by inheritance, the exact 5322.From domain is no longer
> protected by DMARC (with a policy of "none").  This wouldn't have happened
> with the PSL method.
> 3) If 1/2 are problematic, does BIMI need to consider a tree-walk method?
> That seems like it may be less ideal as it may not protect the apex domain.
> 4) Or is this just not a problem at all?
>
> Thanks
>
> --
> Alex Brotman
> Sr. Engineer, Anti-Abuse & Messaging Policy
> Comcast
>
> --
> bimi mailing list
> bimi@ietf.org
> https://www.ietf.org/mailman/listinfo/bimi
>