I tend to agree on that last Receiver bullet being unenforced.  If I had to 
choose between an organization deploying DMARC without reporting, or holding up 
on deploying DMARC because they can’t provide reporting for X,Y,Z reasons .. 
I’m choosing the former.  Does it potentially leave a hole in intelligence?  
Yes, though doesn’t leave a hole in protection.  I suppose there’s the case 
where they just say they’ve only “partially” implemented DMARC, but then what’s 
the point of the MUST.

I want to stew on some of the other bits.  I’m on the fence for the Domain 
Owner requirements.  I also feel like the document needs a better definition of 
Mediator (I didn’t see one in the document).

--
Alex Brotman
Sr. Engineer, Anti-Abuse & Messaging Policy
Comcast

From: dmarc <dmarc-boun...@ietf.org> On Behalf Of Murray S. Kucherawy
Sent: Thursday, August 19, 2021 3:16 PM
To: IETF DMARC WG <dmarc@ietf.org>
Subject: Re: [dmarc-ietf] Ticket #66 (Define What It Means to Have Implemented 
DMARC) and #62 (Reporting a MUST)

On Thu, Aug 19, 2021 at 11:24 AM Todd Herr 
<todd.herr=40valimail....@dmarc.ietf.org<mailto:40valimail....@dmarc.ietf.org>> 
wrote:



   Mail Receiver: To implement DMARC, a mail receiver MUST do the

   following:



   *  Perform DMARC validation checks on inbound mail



   *  Perform validation checks on any authentication check results

      recorded by mediators that handled the message prior to its

      reaching the Mail Receiver.



   *  Send aggregate reports to Domain Owners at least every 24 hours

      when a minimum of 100 messages with that domain in the

      RFC5322.From header field have been seen during the reporting

      period

Let's discuss...

I'm of the opinion that this last bullet can't be a MUST.  I understand that 
operators in this space really want this to be mandatory, but we are going to 
run into cases where doing this is difficult or impossible either because of 
operational difficulties (think resource-constrained environments) or policies 
("I am not willing to share any detail about what mail arrives here").  Making 
this a MUST explicitly disqualifies them.

Moreover, I would claim that not generating aggregate reports does not impede 
interoperability at all, which means use of MUST or even SHOULD here is not 
appropriate.

-MSK, participating only
_______________________________________________
dmarc mailing list
dmarc@ietf.org
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to