On Mon, 15 Jan 2007, Jeremy Harris wrote:
Perhaps it's time for forwarding to move away from being done
store-and-forward, towards being done synchronously?
You mean the stunt of having an MTA begin the outgoing mail transaction
while the incoming mail transaction is still in progress, so that the
forwarder doesn't acknowledge the sender's DATA until the recipient
acknowledges the forwarder's?
Aside from the brittleness of such a scheme (it breaks if the sender MTA
has short timeouts and the recipient is sluggish, even if both are within
RFC limits), it only saves you from backscatter if you *don't* use SRS.
Under "traditional forwarding", if the recipient MX decides it can't
deliver the mail post-transaction, it has responsibility for the bounce.
If they actually do bounce, it's their IP that gets a blackened reputation
(now that RBLs punish backscatter). This isn't the forwarders problem.
Under SRS, if the recipient backscatters, the bounce goes to the
SRS-transformed address, and then the forwarder has to relay it. As far
as his IP reputation goes, it will be just as if he was the backscatter
emitter. In short, he takes the reputation hit for the recipient's late
rejections.
---- Michael Deutschmann <michael(_at_)talamasca(_dot_)ocis(_dot_)net>
-------
Sender Policy Framework: http://www.openspf.org/
Archives at http://archives.listbox.com/spf-discuss/current/
To unsubscribe, change your address, or temporarily deactivate your
subscription,
please go to http://v2.listbox.com/member/?list_id=735