Date: Fri, 21 May 2010 08:38:40 +0200 From: "Spenst, Aleksej" <Aleksej.Spenst@harman.com> To: "'freebsd-pf@freebsd.org'" <freebsd-pf@freebsd.org> Subject: Re: Ingress traffic shaping Message-ID: <20290C577F743240B5256C89EFA753810C3CC9FE54@HIKAWSEX01.ad.harman.com>
next in thread | raw e-mail | index | archive | help
=20 -----Urspr=FCngliche Nachricht----- Von: Raymond Gesendet: Donnerstag, 20. Mai 2010 16:29 An: Spenst, Aleksej Betreff: Re: Ingress traffic shaping On 5/20/2010 04:18, Spenst, Aleksej wrote: > If I understand it correctly, ingress traffic shaping is not possible wit= h pf/altq. > Are there any tricks to do it? > =20 Yes. Inbound traffic can be queued as you discovered by setting up a queue= on the internal NIC. You can split off a separate full speed queue for tr= affic originating from the local host if you don't want local network traff= ic to be slowed as well. The problem is that such throttling is largely fruitless. The traffic has = already been sent by the remote end, and dropping the packet simply means t= he remote end will have to send it again, consuming even more bandwidth. Y= ou are simply hoping the remote end follows proper procedure and throttles = itself as it starts getting dropped packets.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20290C577F743240B5256C89EFA753810C3CC9FE54>