On Mon, Dec 7, 2020 at 4:41 PM John Levine <[email protected]> wrote:

> >I don't understand the security or GDPR references.
>
> Well this is amusing. I wondered if anyone had ever implemented some
> version of the http reporting in early DMARC drafts, so I set up a new
> domain
> with a server that will accept POST or PUT requests and added its URI to my
> DMARC records.
>

What is tht "mantra"? Running code and rough consensus? Has anyone
implemented such reporting?

>
> I didn't get any of those (the POSTs below are not to the right URI)
> but it's impressive how fast Russian bots started to probe it, within
> hours.
>

I thought it's about interoperability. Simply having a webserver running
doesn't come close to interoperability, and certainly not at scale.


>
> This is not a reason to avoid https reporting. Every web site gets
> probed like this and so long as your web server rejects unknown URIs,
> they're harmless. After all, my e-mail reporting addresses get a
> certain amount of spam, too.
>
> R's,
> John
>

My question was not intended to imply that HTTPS reporting should be
avoided. My point was that there has been basically no security discussion
or scrutiny of such an implementation. This document is now proposed
Standards Track, not Informational or Experimental. As such, proposed
changes or additions should be given. Your surprise at what happened so
quickly after your "implementation" proves my point that there has been
little if any thought of security considerations for this proposed change.
My comment about GDPR was simply meant to indicate that the world has
changed quite a bit since DMARC was made public in 2011. So too have
security considerations evolved. I have no objection in principle to HTTPS
reporting, only that such an addition be given the same consideration and
scrutiny that DMARC itself is being held to.

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

Reply via email to