Users Guide

ETS Prerequisites and Restrictions
On an S6000 switch, ETS is enabled by default on Ethernet ports with equal bandwidth assigned to each 802.1p priority. You can
change the default ETS conguration only by using a DCB map.
The following prerequisites and restrictions apply when you congure ETS bandwidth allocation or strict-priority queuing in a DCB
map:
• When allocating bandwidth or conguring strict-priority queuing for dot1p priorities in a priority group on a DCBx CIN interface,
take into account the CIN bandwidth allocation (see Conguring Bandwidth Allocation for DCBx CIN) and dot1p-queue mapping.
• Although ETS bandwidth allocation or strict-priority queuing does not support weighted random early detection (WRED), explicit
congestion notication (ECN), rate shaping, and rate limiting because these parameters are not negotiated by DCBx with peer
devices, you can apply a QoS output policy with WRED and/or rate shaping on a DCBx CIN-enabled interface. In this case, the
WRED or rate shaping conguration in the QoS output policy must take into account the bandwidth allocation or queue
scheduler congured in the DCB map.
Priority-Group Conguration Notes
When you congure priority groups in a DCB map:
• A priority group consists of 802.1p priority values that are grouped together for similar bandwidth allocation and scheduling, and
that share the same latency and loss requirements. All 802.1p priorities mapped to the same queue must be in the same priority
group.
• In a DCB map, each 802.1p priority must map to a priority group.
• The maximum number of priority groups supported in a DCB map on an interface is equal to the number of data queues (4) on
the port. Each priority group can support more than one data queue.
• You can enable PFC on a maximum of two priority queues on an interface.
• If you congure more than one priority group as strict priority, the higher numbered priority queue is given preference when
scheduling data trac.
Hierarchical Scheduling in ETS Output Policies
ETS supports up to three levels of hierarchical scheduling.
For example, you can apply ETS output policies with the following congurations:
Priority group 1
Assigns trac to one priority queue with 20% of the link bandwidth and strict-priority scheduling.
Priority group 2 Assigns trac to one priority queue with 30% of the link bandwidth.
Priority group 3 Assigns trac to two priority queues with 50% of the link bandwidth and strict-priority scheduling.
In this example, the congured ETS bandwidth allocation and scheduler behavior is as follows:
Unused bandwidth
usage:
Normally, if there is no trac or unused bandwidth for a priority group, the bandwidth allocated to the group
is distributed to the other priority groups according to the bandwidth percentage allocated to each group.
However, when three priority groups with dierent bandwidth allocations are used on an interface:
• If priority group 3 has free bandwidth, it is distributed as follows: 20% of the free bandwidth to priority
group 1 and 30% of the free bandwidth to priority group 2.
• If priority group 1 or 2 has free bandwidth, (20 + 30)% of the free bandwidth is distributed to priority
group 3. Priority groups 1 and 2 retain whatever free bandwidth remains up to the (20+ 30)%.
Strict-priority
groups:
If two priority groups have strict-priority scheduling, trac assigned from the priority group with the higher
priority-queue number is scheduled rst. However, when three priority groups are used and two groups have
strict-priority scheduling (such as groups 1 and 3 in the example), the strict priority group whose trac is
Data Center Bridging (DCB)
255