06-17-2010 10:22 AM - edited 03-06-2019 11:37 AM
According to some Cisco 360 R&S lab documentation I'm reading the command 'srr-queue bandwidth shape 4 4 0 0' will result in queue's 1 and 2 getting 25% of the shaped bandwidth each. I could understand this if the last two 0's were 4 but they're not, so how does the weighting allocation work?
The full syntax is as follows, although I don't think the share command has anything to do with the weighting in the shape command?
srr-queue bandwidth shape 4 4 0 0
srr-queue bandwidth share 1 1 7 3
Any help appreciated...
06-22-2010 03:16 PM
Hi Richard,
the way the shape command works is that the inverse of the specified weight (i.e. 1/4) multiplied by the interface bandwidth will be what is allocated and guaranteed to the queue. So in this case it could be 1/4 * 100Mbps (assuming an FE interface), which is 25Mbps and 25%.of 100Mbps.
If a weight of 0 is specified for a queue, the queue does not participate in shaping and will share the remaining bandwidth with any other queues that also have been assigned a weight of 0.
In the example configuration you gave, the remaining bandwidth for queues 3 and 4 would be 7/(7+3) * 100Mbps and 3/(7+3)*100Mbps respectively.
HTH
Marcel
06-23-2010 03:24 AM
Hi,
Just a minor correction to Marcel's reply, the remaining bandwidth for queues' 3 & 4 would be 7/(7+3) * (100Mbps - 25Mbps - 25 Mbps) and 3/(7+3)*(100Mbps - 25Mbps - 25 Mbps) respectively.
Moath,
06-23-2010 03:06 PM
Hi Moath,
in the case when both shaped queues are full, you're absolutely right. That would be a worst-case scenario for the shared queues. When there is no traffic in the shaped queues however, the shared queues should be able to go up to interface speed.
Sorry if I wasn't clear on that.
Cheers,
Marcel
06-24-2010 08:18 AM
Thanks for your responses - I get it now
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide