Hi all,
I'm trying to get my head around how buffers and queues function. Am I correct in saying that the buffers shown by the "show buffers" command are used only by the hardware queues (tx and rx rings)?
If this is correct then can I assume that buffers for mqc based queues are held elsewhere in memory and we only really use the "show policy-map interface" command to work out what's going on?
Next I'm trying to work out the difference between the "hold-queue" command on an interface vs in a service policy. If I configure the following:
policy-map SCHEDULE
class FTP
bandwidth 200
hold-queue 100
class HTTP
bandwidth 200
hold-queue 100
policy-map TEST
class class-default
shape ave 512000
hold-queue 200
service-policy SCHEDULE
interface f0/0
hold-queue 200
service-policy output TEST
Am I correct in saying that the hold-queue 200 on the interface allocates enough memory to hold 200 packets for all layer 3 queues meaning that the hold-queue command under the class-default class of TEST should never be higher than 200?
In addition if I want to ensure that classes under the SCHEDULE map are guaranteed to get a buffer when required then the aggregate of hold-queues specified under classes in SCHEDULE should never exceed the 200 specified for the hold-queue for class-default of TEST?
many thanks for any insights here.
cheers,
Ben.
Blogs and organic groups at http://www.ccie.net
Received on Mon Feb 27 2012 - 09:35:08 ART
This archive was generated by hypermail 2.2.0 : Thu Mar 01 2012 - 11:46:56 ART