12. Slope QoS policies
This chapter provides information to configure slope QoS policies using the command line interface.
12.1. Overview of buffer pools
Default buffer pool exists (logically) at each port. Buffer pools cannot be created or deleted in the 7210 SAS. The egress buffer pools are created as access uplink port egress buffer pool for access-uplink ports and access port egress buffer pool for access ports. Based on the maximum number of ports to be supported for access and access-uplink, the total buffer is distributed into the access port egress buffer pool and the access uplink port egress buffer pool. The distribution of the buffers into access and access-uplink port egress pools take care of the buffer requirements at the port level for various queue shaping/ scheduling mechanisms and for various packet sizes varying from 64 bytes to jumbo frames. Each port on the system gets an equal portion of the available buffers. From the buffers allocated to a port, each queue gets its CBS amount of buffers. The remaining buffers are allocated towards the shared MBS pool per port. All the queues of the port can use the buffers from the shared MBS pool. By default, each queue on the access port and access-uplink port is associated with slope-policy default which disables the slope parameters within the pool.
On 7210 SAS-D, WRED is supported to evaluate the packet’s eligibility to be allocated a buffer based on the slope parameters configured for the queue.
On 7210 SAS-Dxp, WRED is supported to evaluate the TCP packet’s eligibility to be allocated a buffer based on the slope parameters configured for the queue. WRED is not supported for non-TCP packets.
12.1.1. Configuration guidelines for 7210 SAS-D
7210 SAS-D provides an option to use either 2 slope per queue or 3 slopes per queue. This can be configured using the CLI command config>system>qos>no use-wred-slopes. The option to use only 2 WRED slopes per queue (port egress queues), allows differentiating in-profile and out-of-profile traffic flows. The option to use 3 WRED slopes per queues allows differentiating in-profile TCP traffic, out-of-profile TCP traffic and non-TCP traffic (both in and out-profile use a single slope). The slope does not get enabled by default. In order to maintain backward compatibility, by default the system uses 3 slopes option and user has to change it explicitly to use 2 slopes (if they desire).
The following table compares the WRED slope used for different traffic flows with 2 slopes or 3 slopes per queue.
Table 46: Slope behavior (applicable to 7210 SAS-D)
Slopes | TCP-non-TCP slope option (uses 3 WRED slopes per queue) | High-low slope option (uses 2 WRED slopes per queue) |
SAP Ingress TCP/IP traffic (Number of VLAN tags <=2) | High-priority TCP slope or low-priority TCP slope, based on packet profile | High-priority or low-priority slope, based on packet profile |
SAP Ingress non-TCP traffic (Number of VLAN tags does not matter) | Non-TCP slope - No in/out profile differentiation | High-priority or low-priority slope, based on packet profile |
SAP Ingress TCP/IP traffic (Number of VLAN tags>2) | Non-TCP slope - No in/out profile differentiation | High-priority or low-priority slope, based on packet profile |
12.1.2. Configuration guidelines for 7210 SAS-Dxp
7210 SAS-Dxp can use two slopes per queue. Using two WRED slopes per queue (port egress queues) allows differentiation of in-profile and out-of-profile TCP traffic flows. The slope is not enabled by default.
 | Note: All non-TCP traffic is tail-dropped if the egress queues are full when the traffic is being enqueued. |
The following table describes the WRED slope used for different traffic flows with two slopes per queue.
Table 47: Slope behavior (applicable to 7210 SAS-Dxp)
Slopes | High-low slope option (Uses 2 WRED slopes per queue) |
SAP Ingress TCP/IP traffic (Number of VLAN tags <=2) | High-priority or low-priority slope, based on packet profile |
SAP Ingress non-TCP traffic (Number of VLAN tags does not matter) | Tail-drop |
SAP Ingress TCP/IP traffic (Number of VLAN tags>2) | Tail-drop |
12.2. Basic configurations
A basic slope QoS policy must conform to the following:
- Each slope policy must have a unique policy ID.
- High slope, low slope and non-TCP slope are shut down (default).
- Default values can be modified but parameters cannot be deleted.
12.2.1. Create a slope QoS policy for 7210 SAS-D
Configuring and applying slope policies is optional. If no slope policy is explicitly applied to a port, a default slope policy is applied.
To create a new slope policy for 7210 SAS-D devices, define the following:
- a slope policy ID valueThe system will not dynamically assign a value.
- include a descriptionThe description provides a brief overview of policy features.
- Option to use either 2 slopes per queue or 3 slopes per queue.
- The high slope for the high priority Random Early Detection (RED) slope graph.
- The low slope for the low priority Random Early Detection (RED) slope graph.
- Optional non-TCP slope for the non-TCP Random Early Detection (RED) slope graph.
- The time average factor (TAF), a weighting exponent used to determine the portion of the shared buffer instantaneous utilization and shared buffer average utilization used to calculate the new shared buffer average utilization.
Use the following CLI syntax to configure a slope policy for both 2-slope per queue and 3-slope per queue option. When using 2-slope per queue, the non-tcp slope parameters are not used.
The following displays the slope policy configuration for 7210 SAS-D:
12.2.2. Create a slope QoS policy for 7210 SAS-Dxp
Configuring and applying slope policies is optional. If no slope policy is explicitly applied to a port, a default slope policy is applied.
To create a new slope policy for 7210 SAS-Dxp devices, define the following:
- a slope policy ID value; the system will not dynamically assign a value
- a description; the description provides a brief overview of policy features
- two slopes per queue:
- the high slope for the high priority Random Early Detection (RED) slope graph
- the low slope for the low priority Random Early Detection (RED) slope graph
- the time average factor (TAF), a weighting exponent used to determine the portion of the shared buffer instantaneous utilization and shared buffer average utilization used to calculate the new shared buffer average utilization
Use the following CLI syntax to configure a slope policy with two slopes per queue.
The following displays the slope policy configuration for 7210 SAS-Dxp:
12.3. Applying slope policies
Based on the 7210 SAS platform capabilities, the slope policies are associated with different entities:
The following CLI syntax examples may be used to apply slope policies to access ports or access-uplink ports:
12.3.1. Default slope policy values
This section provides the default values for slope policies.
12.3.1.1. Default slope values for 7210 SAS-D
The default access egress and access uplink egress policies are identified as policy-id “default”. The default policies cannot be edited or deleted. The following table lists default policy parameters.
Table 48: Slope policy defaults for 7210 SAS-D
Field | Default |
description | Default slope policy |
high (RED) slope | |
Administrative state | shutdown |
start-avg | 70% utilization |
max-avg | 90% utilization |
max-prob | 75% |
low (RED) slope | |
Administrative state | shutdown |
start-avg | 50% utilization |
max-avg | 75% utilization |
max-prob | 75% |
non-TCP (RED) slope | |
Administrative state | shutdown |
start-avg | 50% utilization |
max-avg | 75% utilization |
max-prob | 75% |
12.3.1.2. Default slope values for 7210 SAS-Dxp
The default access egress and access uplink egress policies are identified as policy-id “default”. The default policies cannot be edited or deleted. The following table lists default policy parameters.
Table 49: Slope policy defaults for 7210 SAS-Dxp
Field | Default |
description | Default slope policy |
high (RED) slope | |
Administrative state | shutdown |
start-avg | 70% utilization |
max-avg | 90% utilization |
max-prob | 75% |
low (RED) slope | |
Administrative state | shutdown |
start-avg | 50% utilization |
max-avg | 75% utilization |
max-prob | 75% |
12.4. Deleting QoS policies
A slope policy is associated by default with access and access uplink egress pools. A default policy may be replaced with a non-default policy, but a policy cannot be entirely removed from the configuration. When a non-default policy is removed, the policy association reverts to the default slope policy policy-id default. A QoS policy cannot be deleted until it is removed from all ports where it is applied or if the policies are using the slope-policy.
The following CLI syntax examples can be used to remove slope policies from ports:
12.4.1. Remove a policy from the QoS configuration
To delete a slope policy, enter the following command:
12.5. Copying and overwriting QoS policies
You can copy an existing slope policy, rename it with a new policy ID value, or overwrite an existing policy ID. The overwrite option must be specified or an error occurs if the destination policy ID exists.
The following output displays the copied policies for 7210 SAS-D devices:
The following output displays the copied policies for 7210 SAS-Dxp devices:
12.6. Editing QoS policies
You can change existing policies and entries in the CLI or NMS. The changes are applied immediately to all services where this policy is applied. To prevent configuration errors copy the policy to a work area, make the edits, and then write over the original policy.