On Saturday, June 7, 2014 2:25 PM, Vlatko Salaj <[email protected]> wrote:


> anyway, what i propose is quite simple, pretty trivial
> and easy to implement, actually:
> http://www.ietf.org/mail-archive/web/dmarc/current/msg00813.html


consider this example of DMARC DNS record tags:

adkim=s:author-domain,s:yahoo.com,n:gmail.com
aspf=r:author-domain,r:yahoo.com,n:gmail.com


so, what i am proposing is changing adkim and aspf DMARC tags so
they become:

a comma-separated list of "alignment-strength:domain" pairs, in which

1. domains in colon pairs r 1st/3rd party domains author-domain
wants to align to or not,

2. default is "r:author-domain", similar to way it is now [relaxed],
and can be left out completely, or partly, on any colon side,

3. alignment-strength can be n, r or s [none, relaxed or strict],
which beside the usual relaxed and strict alignment policy, adds
possibility [n=none] to withheld alignment from a domain
[as a countermeasure against a newly discovered abuse].


the above example would read:

1. request strict dkim alignment for author-domain,

2. request strict dkim alignment for yahoo.com dkim signatures,

3. request dkim alignment failure for gmail.com dkim signature
[not rly needed if gmail.com isn't used for autor-domain's email,
but a useful thing to combat new abuse and remove trust from
previously trusted domain],

4. request relaxed spf alignment for author-domain and yahoo.com,

5. and finally, request spf alignment failure for gmail.com.
[same as 3.]


also, since relaxed and author-domain r defaults, we could do:
aspf=yahoo.com,n:gmail.com
instead of previous example, with same results.

even things like
adkim=yahoo.com,ymail.com,gmail.com,mybrosdomain.com
would be legitimate and would specify relaxed alignment policy
for all listed domains.


also, since i'm proposing a change on DMARCv1 tags, any such
change would essentially require DMARC advancing to v2,
but i consider this appropriate action, since it introduces
principal differences in how basic operations [alignment]
work in DMARC.


On Saturday, June 7, 2014 4:08 PM, Stephen J. Turnbull 
<[email protected]> wrote:

> Unfortunately there's no protocol in there, you leave it implicit. :-(

look up.


> Unfortunately, AFAICS it doesn't address my needs (ie, MLMs), so I
> doubt I will be able to find time to work through it and figure out
> what you're suggesting in concrete terms.

on small scales, like 1-15 ML-domains, it can address those needs too.
be free to spend ur time on it, if u find it interesting enough,
and forget our hostility from yesterday. :D

sure, it won't fix yahoo's "p=reject", since yahoo will not publish
hundreds of ML-domains in their DMARC DNS record alignment list, but
it will fix other, smaller domains that use "p=reject".


>> to domain-owner's control, where it should be.
> I disagree with "where it should be".  The receiver has ultimate say
> on what messages it will accept, whether for relay to third parties or
> local delivery (or even "silent discard").

in respect to DMARC policy, author-domain should have control
over who posts email on its behalf. that was my point. receivers
should have nothing to do with that, no guesswork, in respect
to DMARC, but they r forced to do it now, going even as far to process
"p=reject" as "p=quarantine".

i'm not talking about receivers' anti-spam and other policies,
we r not in that mailing list. however, i do see DMARC used in
anti-spam filters too, in the future... with troubling results.


>> what u r proposing, Stephen, instead, is not at all trivial or easy
>> to implement,
> Define "trivial" and "easy." Asking the ESP to handle all
> of SPF, DKIM, and DMARC for mail purporting to be from your
> domain is an obvious solution

yeah, well i don't define trivial and easy like that. and i doubt
any ESP will introduce something like that. it essentially changes
the way they work now. i see no incentives for them to adapt to
these requests.

sure, i support them, be free to add my name to ur petition, but
i'm sure it's a waste of our efforts.


>> nor does it solve more than just my special use case.
> So what?

so, solution fragmentation. there r better solutions that cover
more stuff at once, without any 3rd party involvement, which is
preferential. it is also my principal viewpoint on life.


-- 
Vlatko Salaj aka goodone
http://goodone.tk


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

Reply via email to