25-8
Catalyst2950 Desktop Switch Software Configuration Guide
78-11380-05
Chapter25 Co nf iguring QoS
Understanding QoS
Port Priority
Frames received from users in the administratively-defined VLANs are classified or tagged for
transmission to other devices. Based on rules that you define, a unique identifier (the tag) is inserted in
each frame header before it is forwarded. The tag is examined and under stood b y each devi ce befo re an y
broadcasts or transmissions to other switches, routers, or end stations. When t he f rame rea ches th e l ast
switch or router, the tag is removed before the frame is sent to the target end station. VLANs that are
assigned on trunk or access ports without identification or a ta g ar e ca lled native or untagged frames.
For IEEE 802.1Q frames with tag information, the priority value from the header frame is used. For native
frames, the default priority of the input port is used.
Port Scheduling
Each port on the switch has a single receive queue buffer (the ingress port) for incoming traffic. When
an untagged frame arrives, it is assigned the value of the port as its port default priority. You assign this
value by using the CLI or CMS. A tagged frame continues to use its assigned CoS value when it passes
through the ingress port.
CoS configures each transmit port (the egress port) with a normal-priority transmit queue and a
high-priority transmit queue, depending on the frame tag or the port information. Frames in the
normal-priority queue are forwarded only after frames in the high-priority queue are forwarded.
The switch (802.1P user priority) has four priority queues. The frames are forwarded to appropriate
queues based on the priority-to-queue mapping that you defined.
CoS and WRR
The switch supports four CoS queues for each egress port. For each queue, you can specify t hes e typ es
of scheduling:
•Strict priority scheduling
Strict priority scheduling is based on the priority of queues. Queues ca n have priorities from 0 to 7,
7 being the highest. Packets in the high-priority queue always tra nsm it first, an d packet s in th e
low-priority queue do not transmit until all the high-priority queues become empty.
•Weighted round-robin (WRR) scheduling
WRR scheduling requires you to specify a number that indicates the importance (weight) of the
queue relative to the other CoS queues. WRR scheduling prevents the low-priority queues from
being completely neglected during periods of high-priority traffic. The WRR sc hedu ler tra nsm its
some packets from each queue in turn. The number of packets it send s corr e spon ds to the r elat ive
importance of the queue. For example, if one queue has a weight of 3 and another has a weight of
4, three packets are sent from the first queue for every four that are sent from the second queue. By
using this scheduling, low-priority queues have the opportunity to send packets even though the
high-priority queues are not empty.