Disclaimer
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.
Liability Disclaimer
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.
Posting
Classes, excluding LLQ, are generally serviced relative to their weights relative to other active classes.
Given:
policy sample1
class A
bandwidth percent 25
class B
bandwidth percent 25
class C
bandwidth percent 50
If all classes were active, classes A and B should each get 25% and class C 50%, as expected.
However what if only classes A and B were active, or just classes B and C?
Classes A and B, if only ones active, would each get 50%.
Classes B and C, if only ones active, class B would get 1/3 and class C would get 2/3 (except on ASRs, where I understand "excess" bandwidth is proportioned differently).
What if given:
policy sample2
class A
bandwidth percent 10
class B
bandwidth percent 10
class C
bandwidth percent 20
My understanding is, policy sample2 would work like sample1 (except on ASRs).
What if given:
policy sample3
class A
bandwidth percent 25
class B
bandwidth percent 25
class class-default
fair-queue
If the above is pre-HQF, multiple flow queues will compete for bandwidth with the class queues for classes A and B. Servicing is a bit complex, I recall someone had posted an article on the internet how queue weights are determined. Also, BTW, part of the complexity, class-default FQ computes IP Precedence in weight determination.
policy sample4
class A
bandwidth percent 25
class B
bandwidth percent 25
class C
bandwidth percent 50
fair-queue
In HQF, any class can use FQ. The aggregate of all the class flow queues is supposed to be "honor" the class bandwidth. FQ in this case does not use IP Precedence.
Cisco doesn't always well/clearly document how CBWFQ services all its flows, but its operation does differ a bit on IOS version and/or platform. (Personally, I find Cisco allowing different CBWFQ servicing, between pre-HQF and HQF CBWFQ, for the same exact syntax, disconcerning.)
