Hopefully someone will correct me if I am wrong, QoS is my weakest subject in my studies.
My understanding, QoS when configured takes into account line speed. You can adjust bandwidth limits, through policing or shaping, but QoS sends at the rate of your hardware line.
I believe that QoS, all it does, is hold data for a little bit and then sends it at the tip toppest speed it can on the physical line. It just basically does this in bursts when limiting stuff. So you really only see QoS come into play when there is "saturation" of classes you are trying to limit the speed on. So if you have HTTP limited at 50 percent of the line, it will shape it, but when it does send data it still sends it no slower on Layer 1 then it would normally. It just does it in intervals.
After I posted this I finally managed to find some information. Apparently an interfaces bandwidth is used as a reference for QOS. So the same thing a routing protocol would use.
Comments
My understanding, QoS when configured takes into account line speed. You can adjust bandwidth limits, through policing or shaping, but QoS sends at the rate of your hardware line.
I believe that QoS, all it does, is hold data for a little bit and then sends it at the tip toppest speed it can on the physical line. It just basically does this in bursts when limiting stuff. So you really only see QoS come into play when there is "saturation" of classes you are trying to limit the speed on. So if you have HTTP limited at 50 percent of the line, it will shape it, but when it does send data it still sends it no slower on Layer 1 then it would normally. It just does it in intervals.