DGS-3024 Layer 2 Switch CLI Reference Manual
13
Q
O
S C
OMMANDS
The DGS-3024 switch supports 802.1p priority queuing. The Switch has 4 priority classes of service. These priority classes of
service are numbered from 3 (Class 3) — the highest priority class of service — to 0 (Class 0) — the lowest priority class of
service. The eight priority queues specified in IEEE 802.1p (p0 to p7) are mapped to the Switch’s priority classes of service as
follows:
•
Priority 0 is assigned to the Switch’s Q1 class.
•
Priority 1 is assigned to the Switch’s Q0 class.
•
Priority 2 is assigned to the Switch’s Q0 class.
•
Priority 3 is assigned to the Switch’s Q1 class.
•
Priority 4 is assigned to the Switch’s Q2 class.
•
Priority 5 is assigned to the Switch’s Q2 class.
•
Priority 6 is assigned to the Switch’s Q3 class.
•
Priority 7 is assigned to the Switch’s Q3 class.
Priority scheduling is implemented using two types of methods, strict priority and round-robin priority. If no changes are made
to the QoS priority scheduling settings the method used is strict priority.
For strict priority-based scheduling, packets residing in the higher priority classes of service are transmitted first. Only when
these classes of service are empty, are packets of lower classes of service allowed to be transmitted. Higher priority packets
always receive preference regardless of the amount of lower priority packets in the buffer and regardless of the time elapsed
since any lower priority packets have been transmitted. By default, the Switch is configured to empty the buffer using strict
priority.
NOTICE:
The default QoS scheduling arrangement is a strict
priority schedule. To customize scheduling to set up round-robin
queue clearing, the MAX. Latency and MAX. Packets values need
to be changed using the config scheduling command. See
config
scheduling
below.
To use implement round-robin (weighted) priority, the Switch’s four priority classes of service can be configured to reduce the
buffer in a round-robin fashion - beginning with the highest priority class of service, and proceeding to the lowest priority class
of service before returning to the highest priority classes of service.
The weighted-priority based scheduling alleviates the main disadvantage of strict priority-based scheduling
−
in that lower
priority class of service get starved of bandwidth
−
by providing a minimum bandwidth to all classes of service for transmission.
This is accomplished by configuring the maximum number of packets allowed to be transmitted from a given priority class of
service and the maximum amount of time a given priority class of service will have to wait before being allowed to transmit its
accumulated packets. This establishes a Class of Service (CoS) for each of the Switch’s four hardware priority classes of
service.
The possible range for maximum packets is: 0 to 15 packets.
80
