The original message was received at Mon, 18 Oct 1999 08:07:28 -0700 (PDT) from danmcd@localhost ----- The following addresses had permanent fatal errors ----- kent@bbn.com ipsec@lists.tislabs.com ----- Transcript of session follows ----- ... while talking to opal.eng.sun.com.: >>> RCPT To:<ipsec@lists.tislabs.com> <<< 550 5.7.1 <ipsec@lists.tislabs.com>... Relaying denied. IP name possibly forged [2000::56:a00:20ff:fe8f:cf02] 550 ipsec@lists.tislabs.com... User unknown >>> RCPT To:<kent@bbn.com> <<< 550 5.7.1 <kent@bbn.com>... Relaying denied. IP name possibly forged [2000::56:a00:20ff:fe8f:cf02] 550 kent@bbn.com... User unknown
message/delivery-status
-- BEGIN included message
- To: kent@bbn.com
- Subject: Re: Outbound interface as a selector
- From: Dan McDonald <danmcd>
- Date: Mon, 18 Oct 1999 08:07:27 -0700 (PDT)
- Cc: ipsec@lists.tislabs.com
> Absent the use of IPsec, how would a user have selected one interface > vs. another via the usual OS calls (or why would he care)? I can select an interface for multicast traffic using sockets today. IPv6's advanced API also has additional facilities for this. If I want to send a packet to ff02::1 or 224.0.0.1, I need to know which interface I'm sending it out. If I'm multihomed, I'd like to have two SAs for destination 224.0.0.1 or ff02::1, where each SA is for an individual link. Dan
-- END included message