> On Aug 15, 2016, at 8:39 AM, Eliot Lear <[email protected]> wrote:
> 
> Hi Steve,
> 
> 
> 
> On 8/15/16 5:30 PM, Steve Atkins wrote:
>>> On Aug 14, 2016, at 10:11 PM, Eliot Lear <[email protected]> wrote:
>>> 
>>> 
>>> 
>>> On 8/14/16 6:46 AM, Steve Atkins wrote:
>>>> If there were a protocol that said "if you receive mail signed by this
>>>> domain / this key and the recipient isn't in the To: or Cc: field,
>>>> block it", or some similar protocol that signed the envelope
>>>> recipient, that would pretty much eliminate DKIM replay as a threat in
>>>> some cases.
>>> That would be a DKIM flag, right?
>> Yes, part of the DKIM-Signature header, probably (though that's not a
>> requirement, just one obvious implementation).
> 
> Ok.
>> 
>>> And you don't want to block- you just
>>> want the signature treated as invalid.
>> That's one option, and likely the more useful one.
> 
> Ok.  Then the next question is if we do such an option would it help
> fastmail.fm, especially in conjunction with ARC?

Only if a significant fraction of mailboxes implemented it. Or, more accurately,
if a significant fraction of mailboxes that do per-domain reputation tracking
based on d= of DKIM signed mail.

If it were to simply invalidate the DKIM signature then the effect it would have
on forwarding would be minimal in the simple cases. Interaction with ill-advised
DMARC records could be more problematic - I don't think the problems would
be too serious, but it'd take a bit more thought about how it could affect some
typical delivery paths.

Whether those risks - and the significant deployment effort - are justified as a
response to DKIM replay attacks is much less clear. Having 1e-8 of traffic
define your protocol is seldom a good idea.

Cheers,
  Steve
_______________________________________________
mailop mailing list
[email protected]
https://chilli.nosignal.org/cgi-bin/mailman/listinfo/mailop

Reply via email to