On 08/04/2009 03:26:03 PM, Paolo Lucente wrote: > Things get tricky for the output interface instead - as that > can change dynamically, given that multiple interfaces are > available. Two approaches come to my mind at these regards:
> Anybody interested willing to comment? Any third way i might > have missed? To fling ideas around with callus disregard: You could cache the datagram and watch for matches on outbound interfaces in order to capture the outbound MAC address. Then a table lookup would suffice to get the outbound interface name. You should know ahead of time which datagrams will route out, so need cache only those. With tcp you don't need the whole datagram to match, just the source/dest quad and the sequence number. This approach sounds like good enough of an idea to be a bad idea, but maybe it'll spark a good idea in someone. Karl <k...@meme.com> Free Software: "You don't pay back, you pay forward." -- Robert A. Heinlein _______________________________________________ pmacct-discussion mailing list http://www.pmacct.net/#mailinglists