Could somebody please help me understand QoS a little better...
1/ Hardware Queue / TX Ring
My understanding is that if the hardware queue/tx ring on an interface is not full, then outgoing packets are put into this queue ready for transmission out of the interface.
Whatever QoS you've configured will not get looked at because no scheduling needs to be done when the hardware queue is not congested.
This is based on what I got from reading this QoS cheat sheet.
2/ If the WAN link has been rate limited between A and B to be 100M but the interfaces connecting these two end points are GigE, do I have to use hierarchical QoS policy with shaping on the egress to enforce my QoS policy because the hardware queue on a GigE card will never be exhausted.
A (ISP1 GigE) ---> (ISP2 Rate Limit 100M) ---> B (ISP1 GigE)
3/ What happens to packets that exceed the policing rate which I've elected to have remarked. Do these packets get remark and then buffered?
That's it for now...