[ietf-dkim] Is anyone using ADSP? - bit more data from the receiving side
gmail.sant9442 at winserver.com
Wed Oct 14 07:06:26 PDT 2009
Ian Eiloart wrote:
> --On 13 October 2009 09:32:20 -0700 "Murray S. Kucherawy"
> <msk at cloudmark.com> wrote:
>>> -----Original Message-----
>> Another data point: Google Mail won't use ADSP because they will not
>> discard someone's mail outright without a written agreement from the
>> sending domain agreeing to same, absolving them of responsibility for
>> mail that never arrives.
> You mean that they won't publish ADSP records? Or that they won't respect
> any ADSP records? Or that they won't discard "discardable" messages?
> Logically, none of these things follow. Publishing ADSP records doesn't
> mean that Google will discard anything, though it does grant permission for
> others to do so. They have lots of other things that they can do as a
> result of ADSP fails. Presumably, they'd be more aggressive with
> quarantining mail if there's an ADSP record that renders a specific email
> discardable. Heck, they could even argue that publication of
> "dkim=discardable" does absolve them.
+1 and that is a very critical point for product engineers especially
when there new legal terms like "domain responsibility" peppered
throughout the documents. This is just asking for trouble one way or
another. That alone can scare people away (raises the barrier to
POLICY provides indemnification for receivers with a clear DOMAIN
publication for its expectation for signatures.
ADSP also provides a newly IETF sanctioned and officially authorized
protocol mechanism for silent discarding of accepted mail without
notifications. And its not a coincidence, the new 2009 RFC 5321 for
SMTP did finally recognize the backscatter industry problem and added
a provision that allows for reasonable discarding of mail due to
abuse. ADSP provides that reasonable classification.
I can understand a RECEIVER not supporting RFC 5617. The idea here
there is a final deposition. It is not pushing mail back into the mail
stream. Passthru systems are traditionally not expected not to alter
mail (there is legal case history here) with the exception of adding
network control lines (trace headers).
Regardless of what ADSP is suppose to mean, the issue is forwarders,
relays, intermediary (re)signer ignorance for ADSP domains,
effectively pushing DKIM/ADSP violating transactions back into the
network mail stream and thus causing down links consequences.
In effect, ADSP (LEVINE) is saying:
This is possible useful for MDA to use.
But MTA (intermediary signers) can ignore it.
I don't think that is sound engineering.
More information about the ietf-dkim