Filter
Top Products
User’s Manual of SGSD-1022 / SGSD-1022P
SGSW-2840 / SGSW-2840P
620
Command Usage
• The precedence for priority mapping is IP Port, IP Precedence or IP DSCP, and default switchport priority.
• The default priority applies for an untagged frame received on a port set to accept all frame types (i.e, receives both
untagged and tagged frames). This priority does not apply to IEEE 802.1Q VLAN tagged frames. If the incoming
frame is an IEEE 802.1Q VLAN tagged frame, the IEEE 802.1p User Priority bits will be used.
• This switch provides eight priority queues for each port. It is configured to use Weighted Round Robin, which can be
viewed with the show queue bandwidth command. Inbound frames that do not have VLAN tags are tagged with
the input port’s default ingress user priority, and then placed in the appropriate priority queue at the output port. The
default priority for all ingress ports is zero. Therefore, any inbound frames that do not have priority tags will be placed
in queue 0 of the output port. (Note that if the output port is an untagged member of the associated VLAN, these
frames are stripped of all VLAN tags prior to transmission.)
Example
The following example shows how to set a default priority on port 3 to 5:
Console(config)#interface ethernet 1/3
Console(config-if)#switchport priority default 5
Console(config-if)#
Related Commands
show interfaces switchport
queue bandwidth
This command assigns weighted round-robin (WRR) weights to the four class of service (CoS) priority queues. Use the no form
to restore the default weights.
Syntax
queue bandwidth weight1...weight4
no queue bandwidth
weight1...weight4 -The ratio of weights for queues 0-3 determines the weights used by the WRR scheduler. (Range: 1-15)
Default Setting
Weights 1, 2, 4, 8 are assigned to queues 0-3 respectively.
Command Mode
Global Configuration
Command Usage
WRR controls bandwidth sharing at the egress port by defining scheduling weights for allocated service priorities.
Queue weights must be configured in ascendant manner, assigning more weight to each higher numbered queue (that is,
Q0 ≤Q1 ≤Q2 ≤ Q3).