In draft-ietf-marf-spf-reporting we changed the address construction rules to 
use a localpart from the new mechanism published in DNS and the SPF domain in 
order to limit the abilty of random third parties to point this kinds of 
reports at unrelated receivers who may not be prepared for them (and retired a 
security consideration in the process).

We did not (I now notice, thanks to Alessandro Vesely's not on the subject in 
the SPF draft, make a similar change in the DKIM draft.  I think we should.

My proposal is to drop 3.1.  Extension DKIM Signature Tag and change the 
address construction in the ra= tag to use the signing domain (d=) in the 
signature.  In this manner the reports will only go back where they came from 
(in a general sense).

Scott K
_______________________________________________
marf mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/marf

Reply via email to