ietf-smtp
[Top] [All Lists]

Re: Comments (II) on draft-church-dns-mail-sender-02.txt

2002-08-29 00:30:15

     As I said before in private mail, my goal is not to avoid modification
of the headers en route (though I agree that's not a bad idea), but to
ensure that the original sender doesn't give a false sender address.

http://www.ietf.org/internet-drafts/draft-church-dns-mail-sender-02.txt

In section 3.2: 
  "which the sending MTA would encrypt with its private key;"

In section 3.4: 
  "as long as the relay MTA can properly respond to an MS challenge for 
   the domain, the sender address will be accepted." 

May I know, how a proper relay MTA (e.g. mailing list, bounce, forward)
is able answer a MS challenge without having access to the private key?

     It can't, thus it will need to rewrite the envelope sender.  (As Keith
Moore pointed out earlier, mailing lists are expected to do this anyway,
and IMO this is proper behavior for forward/bounce as well.)  This is
discussed more thoroughly in revision 03 (currently being edited):

    "In cases where a message is forwarded through one or more MTAs before
     reaching its final destination [...], the MTA at the message's
     destination will typically reject the message when the forwarding MTA
     fails to respond properly to an MS challenge for the sender's domain.
     Thus, MTAs which forward messages should modify the sender address of
     such messages to one which they can respond to an MS challenge for.
     However, depending on circumstances, it may be appropriate for the MTA
     to record elsewhere in the message the original sender address."

  --Andrew Church
    achurch(_at_)achurch(_dot_)org
    http://achurch.org/