It hath once been written:
In section 5.3, this specification assumes all SPF checking will be done
at SMTP time which is not realistic. Some examples: some
implementations may retest a message later to either verify Sender-ID
results, because the load was too high when the message was received,
because the SPF checking was done after-SMTP (SpamAssassin).
...
Sorry, but I do think this should be the case. The SPF checks are not
excessively intensive in almost all cases.
Now, given, I did say in _almost_ all cases. Of course, excessive loads may be
an issue to have to consider. But, considering the SPF checks are RFC 2821
checks, it would be much better for all SPF checks to be done during the
envelope exchange.
If you are having a load issue, would it not be better to defer the connection
with a 4XX response? Having the remote MTA try the connection again at a later
time would allow for an abatement to the the heavy load the receiving MTA may be
experiencing.
This seems much preferable as it would alleviate the need for any DATA to be
transmitted during the session.
Cheers!
================================================================
Steven G. Willis sgwillis(_at_)deepskytech(_dot_)com 772.794.9494
Deep Sky Technologies, Inc. http://www.deepskytech.com/
http://www.badchickens.com/ http://www.store-secure.com/
AIM-iChat: dstisgwillis
================================================================
A: Yes.
Q: Are you sure?
A: Because it reverses the logical flow of conversation.
Q: Why is top posting frowned upon?
================================================================