Ill respond to each point  in a separate thread so it can easily be seen in the 
archives. 

Murray wrote
"We discussed this internally before the DMARC spec became public.  If any
particular technology DMARC supports can have either a "pass+align" or
"fail" result, then the cross product of DKIM and SPF means you have four
possibilities.  That's now p= through p3=.  Now suppose we adopted another
technology like TLS; now it's p= through p7=.  Obviously this won't scale.

Is the complexity this adds worth the use cases it's protecting?"

..........
Chris Lamont Mankowski wrote: 

I think a longer term cross product would be better viewed in terms of 

   The abstraction WHAT is being validated. 
   Is the authentication technology "Active" or "Passive" in terms of DMARC 
evaluation result.  

I think there are already ambiguities that complicate things for me.  This 
point is all about what email technologies (present and future) are in scope 
and out of scope for DMARC evaluation. Also how does DMARC figure that out. 

Instead of explicitly listing p1... to p7 as you describe, let's abstract the 
common feature.  Similar to how the OSI has layers to a network stack, there 
are levels of authentication integrity and privacy are layered:

Keeping with the idea that DMARC is an abstraction we have:  
 1. Several technologies that authenticate from one MTA to the next (SPF and 
TLS) that focus on the message envelope.  AKA p1= 

 2. Several technologies that bind the message body (or from field) to the 
sender (DKIM, SMIME,PGP )  AKA p2=

 3. Several technologies that ensure no modification to the message (SMIME, 
PGP, DKIM) 

 4. Several technologies that encrypt the payload from one MTA to the next 
(SMIME, PGP, not DKIM) 


The DMARC spec says something to the effect of:  if SPF records can't be found 
then DMARC policy won't apply.  But what about a domain that doesn't have ADSP 
and also signs the messages with DKIM?  I think it would be helpful to catalog 
and classify existing technology and deployments into two categories

   1. Active: meaning the policy and senders support for a technology can be 
discovered without receiving an email.  

   2. Passive; meaning the senders support for a technology can be discovered 
only by receiving an email. (PGP, SMime, TLS Auth, DKIM without ADSP) 

It would then be helpful to describe in the DMARC spec the disposition 
assumptions that should be made in each situation. (what are they)   

For example, should DMARC treat DKIM signed messages differently if ADSP exists 
versus if it is missing?   I'm pretty sure this is in the archives but not in 
the forward looking context of "active" vs "passive". 

I think such a construct (p1=) with examples and definitions of active vs 
passive models will ensure a robust DMARC implementation across 3rd parties. 

Granted right now in DMARC we are only talking about p1= and p2=    I want to 
discuss p3= and p4= at some point in the future. 

  


Sent via BlackBerry from T-Mobile

_______________________________________________
dmarc-discuss mailing list
[email protected]
http://www.dmarc.org/mailman/listinfo/dmarc-discuss

NOTE: Participating in this list means you agree to the DMARC Note Well terms 
(http://www.dmarc.org/note_well.html)

Reply via email to