Configuration Guide User guide

FastIron Configuration Guide 585
53-1002494-02
Dynamic buffer allocation for QoS priorities for FastIron X Series devices
• Total Transmit Queue Depth Limit – The total maximum number of transmit buffers allocated
for all outbound packets on a port. Packets are added to the port's outbound queue as long as
the number of buffers currently in use is less than the total transmit queue depth limit. When
this limit is reached, any new packets attempting to enter the port’s transmit queue will be
dropped until at least one buffer is freed.
• Transmit Queue Depth Limit for a Given Traffic Class – The maximum number of transmit
buffers allocated for packets with a given traffic class (0 through 7) on a port. Packets with the
specified traffic class are added to the port’s outbound queue as long as the number of
buffers currently in use for that traffic class is less than the transmit queue depth limit for the
traffic class. When this limit is reached, any new packets with the specified traffic class
attempting to enter the port’s transmit queue will be dropped.
Except for the third-generation Interface modules, you can increase or decrease both of these
queue depth limits per port. On the SX-FI48GPP, SX-FI-24GPP, SX-FI-24HF, SX-FI-2XG or SX-FI-8XG
modules, to increase or decrease the queue depth limits for a port, you configure a buffer profile
and then apply it to the port. A buffer profile can be tied to one or more ports. Therefore, if you
change the configuration on one port, it will change the configuration on all ports associated with
the same buffer profile.
Dynamic buffer allocation for QoS is useful in situations where applications have intermittent
bursts of oversubscription. For example, by increasing the buffers on the egress port, the Brocade
device will be able to forward oversubscribed packets instead of dropping them.
Default queue depth limits for FastIron X Series devices
Table 104 defines the default maximum queue depth values per port, per traffic class. The
Brocade device drops the packets that cause the port to exceed these limits.
NOTE
The SX-FI48GPP Interface module supports 48 tri-speed (10/100/1000) ports. When the ports are
configured at lower speeds, for example, 100 Mbps or 10 Mbps, the maximum queue depth is less
than 256 when egress congestion occurs at the front-end Network Processor (NP) of the
SX-FI48GPP module. If egress congestion occurs at the back-end NP of the SX-FI48GPP module, the
maximum queue depth is 4095. The limit for buffer sharing is 4088.
Configuring the total transmit queue depth limit
for FastIron X Series devices
NOTE
To configure the total transmit queue depth limit on a third-generation module, refer to “Buffer
profile configuration” on page 587.
TABLE 104 Default maximum queue depth
Port type Maximum queue depth
per port, per priority 0 –
6
Maximum queue depth
per port, per priority 7
Total maximum queue
depth
1 Gbps port 96 224 896
10 Gbps port without jumbo enabled 400 704 3504
10 Gbps port with jumbo enabled 352 640 3104