From: Stef Coene Date: Sun, 28 Mar 2004 15:33:39 +0000 Subject: Re: [LARTC] CBQ weight Prioritizing Message-Id: <200403281733.39280.stef.coene@docum.org> List-Id: References: <406594FE.7090405@hardcampa.mine.nu> In-Reply-To: <406594FE.7090405@hardcampa.mine.nu> MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable To: lartc@vger.kernel.org On Saturday 27 March 2004 15:51, lartc wrote: > Hello people, > > I'm curious about the weighting in CBQ, if i have these classes in my CBQ > script, will it work as I expect it to, or are there drawbacks? I am > already using HTB for the moment so don't tell me to use that, this is > simply something I want to know. =3D) > > tc class add dev $DEV parent 1:1 classid 1:10 cbq bandwidth 10Mbit rate > 10Mbit weight 0.3Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 tc > class add dev $DEV parent 1:1 classid 1:11 cbq bandwidth 10Mbit rate 10Mb= it > weight 0.3Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 tc class a= dd > dev $DEV parent 1:1 classid 1:12 cbq bandwidth 10Mbit rate 10Mbit weight > 0.2Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 tc class add dev > $DEV parent 1:1 classid 1:13 cbq bandwidth 10Mbit rate 10Mbit weight > 0.1Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 tc class add dev > $DEV parent 1:1 classid 1:14 cbq bandwidth 10Mbit rate 10Mbit weight > 0.05Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 tc class add dev > $DEV parent 1:1 classid 1:15 cbq bandwidth 10Mbit rate 10Mbit weight > 0.05Mbit prio 1 allot 1514 cell 8 maxburst 20 avpkt 1000 > > I'm allowing all my CBQ classes to use full speed (10Mbit) if they can, > controlled only by my weight ratios, note that the priority is the same f= or > all, which in effect is what I hope will make this work. So the first ent= ry > will thus be given a 3 mbit "minimum" speed as I give it weight ratio 0.3. > I'm trying to get CBQ to behave somewhat more like HTB with this script > with its rate and ceil commands. > > Well I'm curious will this actually work as I intend it to? I don't know. I did some cbq tests myself (the results can be found on=20 docum.org). I only know that Devik told me that weight is used to calculat= e=20 some sort of quantum. That means that weight is only used AFTER each class= =20 is able to send his rate. So in your case, that means weight will never be= =20 used. Stef --=20 stef.coene@docum.org =A0"Using Linux as bandwidth manager" =A0 =A0 =A0http://www.docum.org/ =A0 =A0 =A0#lartc @ irc.openprojects.net _______________________________________________ LARTC mailing list / LARTC@mailman.ds9a.nl http://mailman.ds9a.nl/mailman/listinfo/lartc HOWTO: http://lartc.org/