On 11/3/2017 3:32 PM, Brandon Long wrote:
If you look at RFC 7960, ARC is intended mostly for the mediators
case, though it would also be for the MDA/MTA case.  It does nothing
for the MSA case, though there have been some proposals about having a
hop=0 or just falsifying hop=1 for that.

Most of the MSA issues, though, are mostly of the type "well, dmarc
p=reject/quarantine means the domain holder doesn't want you doing
that", so I'm not clear there's a solution there.


If I still understand ARC and its attempt to do a chain of trust concept, I think there would still be a need for a 3rd Party Seal authorized domain or "registration" concept, otherwise, only the entire chain validity matters. We can have policies for strong and relaxed ARC seal requirements.

Overall, the integrated DKIM policy solution begins with RFC7489.DMARC Tag Extensions:

   3.1.3.  Alignment and Extension Technologies

   If in the future DMARC is extended to include the use of other
   authentication mechanisms, the extensions will need to allow for
   domain identifier extraction so that alignment with the RFC5322.From
   domain can be verified.

We can consider an "atps=" tag and "arc=" tag as experimental concepts. One is DNS-based, no additional RFC5322 overhead, the other has additional RFC5322 modification and processing requirements. I suggest there will be a significant percentage of smaller/private domains that will not have the same requirements as the larger "public service" ESP domains.

If a DMARC record has an "atps=1" it uses rfc6541 to check the 3rd party signer DKIM.SDID domain.

If a DMARC record has an "arc=1" it uses the ARC proposal and some new "policy" proposal (TBD) to check the 3rd party seals and probably 1 or more sealer domain.

I would even consider adding to the experiment the exploration of Levine's Conditional DKIM Signer proposal, with a "cond=1" tag. This proposal is a non-DNS lookup idea to authorize an expected 3rd party signer I believe added to the original signature. The DMARC extended "cond=1" tag could say "a failed p=reject message can be promoted to pass with valid 3rd party conditional signer."


--
HLS


_______________________________________________
dmarc mailing list
dmarc@ietf.org
https://www.ietf.org/mailman/listinfo/dmarc

Reply via email to