>But AFAIK at the moment practically only if there is no NAT involved >(or if you do not want to classify de-nated traffic). >Any news regarding this problem? Suppose we have simple router with upstream interface connected to internet (eth0) and downstream interface connected to lan (eth1). Lan uses private addressing so there is NAT rule used for traffic leaving eth0. You can redirect lan->internet traffic from ingress qdisc of eth1 to ifb0. Traffic on ifb0 will be in "before-nat" state, so private address based shaping will be possible. So no need for classifying de-nated traffic. Some people here on lartc list shared opinions that shaping in ingress is not effective. It worked for me well on routers with hundreds of clients. Cheers, Marek Kierdelewicz _______________________________________________ LARTC mailing list LARTC@xxxxxxxxxxxxxxx http://mailman.ds9a.nl/cgi-bin/mailman/listinfo/lartc
