is this also true for wrr in the hardware queues of a 3560?

Hi Mike,

It is not the same.

In the 3560 there is srr queuing, which means shared round-robin and shaped round-robin, depending on the queuing mode.

In general, both shared round-robin and shaped round-robin queuing limit the bandwidth of the priority queue to prevent bandwidth starvation of non-priority queues.

But in shared round-robin mode, if no packets are in the non-priority queues, then the priority queue can be serviced at up to the line rate.

Also, in shaped round-robin mode, the excess traffic in the priority queue is queued and shaped rather than dropped.

Here's the link to the QoS config guide of the 3560 switches.

http://www.cisco.com/en/US/docs/switches/lan/catalyst3560/software/release/12.2_50_se/configuration/guide/swqos.html#wp1284809

Cheers:

Istvan

i understand that, but does shared round robin take effect even if the port is not congested, say its a 100mb port and only 30mb used or in proper terminology, the tranmit ring isn't full, does the priority queue still get serviced first.

Hi Mike,

I have a supposition as to how it works, but without a clear Cisco reference I do not want to tell you anything.

I have been looking for a Cisco documentation that tells about how it works, but I still haven't found anything.

I keep on searching....

Istvan