Jon Kyme wrote:
OK, we're required not to use source routes nowadays - in "contemporary
clients" - so some equivalent is called for. But we can use the source
routing notation if you like.
I like soure routing notation because using existing syntax to express
new needs seems to be a good thing.
I now see that session verification implies expression of forwarding
between mailboxes. Maybe extending the host relay path syntax (RFC821
"@ONE,@TWO:JOE(_at_)THREE") to be a mailbox forward path syntax
("mb1(_at_)ONE,mb2(_at_)TWO:JOE(_at_)THREE") can help in some cases.
A simple example:
1 - Sender to Forwarder:
mail from:<x(_at_)a>
rcpt to:<y(_at_)b>
(Forwarder decide that there is a forwarding rule from 'y(_at_)b' to 'z(_at_)c')
2 - Forwarder to Receiver:
mail from:<y(_at_)b:x(_at_)a>
rcpt to:<z(_at_)c>
(Receiver can verify that the SMTP client IP is authorized by 'b'.)
3 - Receiver back to Forwarder:
mail from:<z(_at_)c:> (bounce)
rcpt to:<y(_at_)b:x(_at_)a>
(Forwarder decide that there is a forwarding rule from 'y(_at_)b' to 'z(_at_)c'.
Otherwise it is an unauthorized forward attempt.)
4 - Forwarder back to Sender:
mail from:<y(_at_)b:> (bounce variation 1)
mail from:<> (bounce variation 2)
rcpt to:<x(_at_)a>
What do You mean?
z2
_______________________________________________
Asrg mailing list
Asrg(_at_)ietf(_dot_)org
https://www1.ietf.org/mailman/listinfo/asrg