[Top] [All Lists]

Re: [ietf-smtp] SMTP server "RFC2821 Violation" for EHLO ip-literal.

2019-12-15 15:18:21
On Dec 15, 2019, at 3:35 PM, Valdis Klētnieks 
<valdis(_dot_)kletnieks(_at_)vt(_dot_)edu> wrote:

In a walled garden it makes sense to accept mail from anyone who is
able to connect.  To do so globally has proven to be unaffordable.
Anti-spam techniques vary from dropping messages at random to
assessing senders reputation, each has its FPs.

If we write a standard that says address literals aren't permitted, then the
writers of MTAs will be totally in their rights to refuse to accept mail with
an address literal.

As John Klensin observed, I did not suggest (nor did anyone else in this thread)
that address literals should be prohibited.  They are fine to use in private
networks, and "mostly" work even on the public Internet, but (hearsay) over the
years I've seen repeated posts on e.g. the Postfix users list, where sysops 
using regexp HELO name filters that would block address literals.

I do not encourage such filters, they probably don't generally catch junk to
warrant their use, but it is my (admittedly unscientific) impression that
such filters are not uncommon.  It would be good to have real data on
this, but getting it would be a non-trivial exercise.

So the question to AMSL would be whether the rules were adopted second-hand,
and perhaps don't specifically address abusive traffic directed at

As to the standards landscape, address literals should remain part of the
SMTP grammar, but prudent relays should avoid using them outbound.  That's
almost what the document says already.  The only thing I'd consider changing
is adding operational guidance that if you do use them, you may encounter
difficulties delivering to outside organizations across the public Internet.

Now that advice is (admittedly) based on impressions formed by reading various
posts from email administrators over the years, and could be heavily biased.
It is weakly supported by the fact that even AMSL saw fit to implement such
filters, but perhaps they are one of the exceptions that prove the rule[1].


[1] Here I'm going with the archaic meaning of "prove" meaning "test".

[...] The alternative origin given is that the word "prove" is used in the
archaic sense of "test". In this sense, the phrase does not mean that an
exception demonstrates a rule to be true or to exist, but that it tests the
rule, thereby proving its value. There is little evidence of the phrase being
used in this second way.

[ Recursion warning: My use of the idiom is apparently an exception. :-) ]
ietf-smtp mailing list

<Prev in Thread] Current Thread [Next in Thread>