Configuring Data Center Bridging Features 1009
ETS Theory of Operation
First Level of Scheduling
To understand the first level of scheduling, consider Table29-1. Assume that
we have eight ingress ports, each one receiving line rate traffic with one dot1p
priority each. The table shows the mapping of dot1p priorities to the cos-
queues, the min-bandwidth settings, and scheduler modes.
In this scenario, the input to the TCGs is as follows:
All of the pri0 traffic reaches the TCG0. This is because there is no other
priority contending for the same TCG. The minimum bandwidth setting
on the CoS queue does not have any effect.
TCG1 would receive 10% each of pri1 and pri3 and 80% of pri2. Even
though strict mode is enabled for pri2, the minimum bandwidth of pri1
and pri3 is first honored before applying strict mode on pri2.
TCG2 receives 25% each of pri4 and pri5 traffic and the other 50% can be
of pri6 or pri7. This is based on the minimum bandwidth ratio, which is
15:15:30, converted to 1:1:2.
Table 29-3. First Level of Scheduling
Dot1p
Priority
CoS Queue CoS Min-
bandwidth
Scheduler TCG Input to TCG
0 0 10 WRR 0100% of pri0
1 1 10 WRR 110% each of
2 2 10 Strict 1pri1, pri3 and
3 3 10 WRR 180% of pri2
4 4 15 WRR 225% each
5 5 15 WRR 2of pri4, pri5
6 6 30 WRR 2and 50% of
7 6 pri6/pri7