ietf-mxcomp
[Top] [All Lists]

RE: Obstacles between us and the finish line

2004-07-01 08:08:07

****   DMP   ****
1) solid I-D:  Yes

There appears to be a security problem caused by the way a failed MAIL
FROM check may fall back to an EHLO check, allowing 
unintended forgery.
Alternatively this is a source of troublesome 
interoperability problems
if some servers fall back to EHLO and some don't.

Would be eliminated if the EHLO fallback were replaced with checking
Resent-From / SUBMITTER.  A domain that wants to forward mail should start
using these, which is the case with marid-core anyway.

That alone wouldn't fix null reverse paths unless EHLO were used exclusively
for those, or the extra information in Resent-From / SUBMITTER were provided.

5) gratuitous incompatibilities:  None.

Wildcard MX records.

Use an identical wildcard DMP record alongside it, such as:

$ORIGIN example.com.
@       IN      MX      mailhost1
*       IN      MX      mailhost1
*       IN      TXT     "dmp="

That would at least prevent forgeries of undefined subdomains until you had
an actual host you wanted to send mail as (like 
user(_at_)dummy(_dot_)example(_dot_)com), at
which point you'd create records for it or have the sending server provide
records via dynamic DNS update[1].

Doesn't everything tabled so far break when faced with a wildcard MX record
anyway?

[1] Someone told me providing records dynamically was a problem but never
provided an explanation.

-- 
PGP key (0x0AFA039E): 
<http://www.pan-am.ca/consulting(_at_)pan-am(_dot_)ca(_dot_)asc>
Sometimes it's hard to tell where the game ends and where reality bites,
er, begins. <http://vmyths.com/resource.cfm?id=50&page=1>