As noted by Al et al - the forensic reports have a PII problem so sharing
them at scale would always be problematic from Google's perspective. So
option #3 is my favorite. Technically I'm not against a proposal that
includes forensic reports as an option to aggregated reports. I'm just
saying those forensic reports would only be shared with a small number of
senders and therefor maybe does not make sense as part of DMARC reporting.

Context on PII: My impression is that a lot of senders believe that since
they sent the message, they already have the data. I don't think that is
the complete picture - the act of reporting an email as spam/phishing is
between the recipient and the mailbox provider (unsubscribe is between
recipient and sender), hence there is a PII angle.

Context on problem to scale: Sharing forensic reports between companies and
organizations that have strong relationships and trust makes sense. But if
we wanted to have a gate where forensic reports are only shared with "good
senders", identifying good senders accurately would be very costly at scale.

/E

On Thu, Mar 20, 2025 at 9:59 AM Tim Draegen <[email protected]> wrote:

> > On Mar 19, 2025, at 7:08 AM, Barry Leiba <[email protected]>
> wrote:
> >
> > 1. Continue discussing the document, complete it, and ask Andy to
> AD-sponsor it.
> >
> > 2. Abandon the document, leave failure reporting as it had been, and
> refer people to the old (Informational) DMARC spec for documentation of it.
> >
> > 3. Abandon the document and deprecate failure reporting.  That would
> involve mentioning failure reports, noting that they have been seldom used
> and problematic, and stating that their use going forward is not
> recommended.
> >
> > I recommend that we do (3), and call for objections to that path.  If
> you agree with (3), please note that here.  If you prefer (1) or (2),
> please state that and say why.  If you see another reasonable option and
> prefer it, please describe it.
>
>
> Hi, I was asked to chime in here. RUF-style reporting was born into a
> conflicted environment, with A) commercial interests seeing value in the
> reports to power anti-abuse solutions versus B) using RUF-style reporting
> to correct authentication problems. The use cases are hugely different.
>
> IMO, the commercial interests represented by "A" have dominated
> discussions with a tragedy-of-commons effect. Those interests want(ed?)
> RUF-style reporting to become a form of threat feed. If you're lucky enough
> to get access to the private RUF-style sources, you can differentiate your
> threat intel / anti-abuse business. RUF-style reporting is only useful in
> this context if it includes chunks of email content, which runs head on
> into privacy problems. The tragedy is that commercial interests have said
> "Dear Internet, please send RUFs to power our threat-feeds"...  "Sorry,
> can't and won't".
>
> I am firmly in the "B" camp - even though my own commercial interest is in
> widespread DMARC adoption. The presence of RUF-style reporting is
> incredibly helpful in identifying legitimate sources of email that need
> attention. Unfortunately, working for further adoption of RUF-style
> reporting is near impossible because "A"-style interests dominate.
>
> That said, I cannot support #3 as "seldom used" just isn't true, they're
> problematic largely in terms of "A" above, and any attempt to make them
> better would be hindered by "their use going forward is not recommended".
>
> I can't get behind #1 mainly because the expertise needed to inform the
> work isn't here. Murray wrote a bit about creating ways to debug DKIM
> signatures and seeing almost zero adoption. I read this to mean that this
> is not the venue where software developers who want to implement new email
> features live or can work within.
>
> I'm in favor of #2.. except leave open an invitation for a different body
> to refine the work. Bonus points if the decision can include a blurb about
> "this is not supposed to be used as a channel to transmit threat intel" to
> give future humans a chance at success.
>
> Take care,
> =- Tim
>
> _______________________________________________
> dmarc mailing list -- [email protected]
> To unsubscribe send an email to [email protected]
>
_______________________________________________
dmarc mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to