We have an ethernet backbone/backhaul with Cisco (mostly 2960). Our links providers operate fiber SDH network and they give us the links with 100/1000 Ethernet interfaces.
This creates a problem because while the interfaces are 100/1000 the links are not. They start from 2Mbps and only very few get to 100Mbps.
Beginning of this year we implemented priority queueing which I believe doesn't really have a problem weather the link is 2 or 100Mbps since priority queue is always served first.
But now we want to go for more classes and have a problem needing help solving.
2960 use SRR (shaped/shared round robin) queueing. With shared round robin minimum bandwidth is guaranteed to a queue in case of congestion of the link.
I extend this question not just for switches but also for routers.
English is not my native language I hope my question and my intentions were made clear.
Thank you all in advance
On a router ethernet interface you use the the bandwidth command to specify the true throughput.
Sent from Cisco Technical Support iPad App
Thanks very much for your post. the "bandwidth" command unfortunatelly does not limit bandwidth. Its is a mostly "informational" command for routing protocols and some QoS funcionalities use it as reference. Would be amazing to use it unfortunatelly no QoS command that use it is available on 2960 and 3560 switches.
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
The solution to your problem is to first "shape" for available bandwidth, then prioritize within the shaped rate.
L3 switches, particularly LAN L3 switches, often have the fewest WAN like features. For a 2960, you might "shape" using the interface command "srr-queue bandwidth limit". You can use PQ and shared SRR to prioritize between the 4 hardware egress queues.
Thanks very much for your post. I completely ignored this command because the lowest I could get was 10Mbps (10% of interface bandwidth is the lowest allowed) but than I read something this morning about changing port speed to 10Mbps which could get it to as low as 1Mbps.
I'll do some tests and let everyone know.
Just did some quick tests and that is just what expect.
SRR shape weights (defined by srr-queue bandwidth shape w1 w2 w3 w4 command on interface configuration) are called absolute and the weight divides the port bandwidth, meaning it depends on the speed of the port, not the srr bandwidth limit.
SRR share weights (defined by srr-queue bandwidth share w1 w2 w3 w4 command on interface configuration) are relative and the final reserved bandwidth is influenced by the srr-queue bandwidth limit command of the interface.
For all this mls qos must be enabled ("mls qos" on config mode
Normally, this case should be solved by the priority based flow control (PFC).
This mechanism is an improvment of the PAUSE, but it takes into account the priority. It's described in 802.1Qbb-2011.
I don't know if there is a Cisco tool that includes PFC.
That's very new to me but a quick look showed me that at least for Cisco it is only implemented on Nexus switches. Also since the switches are connecting to radios and fiber equipment those would also have to support it.
For anyone interested here's a Cisco whitepaper on Priority Flow Control:
> For anyone interested here's a Cisco whitepaper on Priority Flow Control:
Thanks for the link, it is interesting.
However, the factors described to set up the receiver threshold, and to send the PFC frame, are not independant. They overlap. For example, the response time of sender includes the MTU factor.