Information Technology Reference
In-Depth Information
Congestion Management: Queuing
Queuing is a congestion-management mechanism that can ensure that during (transient) periods of congestion, traffic is appropriately
buffered, prioritized, and reordered before subsequent transmission. Queuing mechanisms are active only when there is congestion. If
there is no congestion, queuing mechanisms are not used.
Several queuing mechanisms are available on Cisco routers, including first-in, first-out (FIFO) queuing, weighted fair queuing
(WFQ), priority queuing (PQ), custom queuing (CQ), IP RTP priority queuing, class-based weighted fair queuing (CBWFQ), and
low-latency queuing (LLQ).
Older queuing mechanisms include PQ, CQ, and IP RTP priority queuing. These older mechanisms have been superseded, and to
support voice together with data applications, CB-WFQ and LLQ are recommended.
Both CB-WFQ and LLQ provide class-based classification of traffic. CB-WFQ provides up to 256 classes (and queues), and LLQ
provides a priority queue with CB-WFQ. Both CB-WFQ and LLQ can provide a bandwidth guarantee, but only LLQ can provide the
latency guarantee that voice media traffic requires.
The following example contains a sample configuration of CB-WFQ:
! <class maps omitted for brevity>
!
policy-map
cbwfq-policy
class
critical-apps
bandwidth percent
30
class
other-apps
bandwidth percent
20
class class-default
fair-queue
!
interface serial 2/0
ip address
192.168.2.1 255.255.255.0
service-policy output
cbwfq-policy
The CB-WFQ policy in the example (
cbwfq-policy
) reserves a minimum of 30 percent of link bandwidth (
bandwidth percent
percentage
) for critical apps and a minimum of 20 percent of link bandwidth for other apps, and then implements a fair queuing
strategy (
fair-queue
) for all other traffic. Finally, the policy map is attached to interface serial 2/0.
