Slope QoS policies

Overview

Default buffer pools exist (logically) at the port, FP or MDA, and node levels. Each physical port has three associated pool objects:

access ingress pool
access egress pool
network egress pool

Each FP has five associated pool objects:

access egress pool
access ingress pool
Acc-Ing MC path management
network ingress pool
network egress pool

By default, each pool is associated with slope-policy default, which has the slopes within the pool shutdown.

Slope policies are also applied when WRED is applied per queue and to the associated megapools; see WRED per queue.

Basic configurations

A basic slope QoS policy must conform to the following:

Each slope policy must have a unique policy ID.
Highplus slope, high slope, low slope, and exceed slope must be shut down (default).
Default values can be modified but parameters cannot be deleted.

Creating a slope QoS policy

Configuring and applying slope policies is optional. If no slope policy is explicitly applied to a pool, a default slope policy is applied.

To create a new slope policy, define the following:

a slope policy ID value. The system does not dynamically assign a value.
a description. The description provides a brief overview of policy features.
the highplus slope for the highplus Random Early Detection (RED) slope graph
the high slope for the high Random Early Detection (RED) slope graph
the low slope for the low Random Early Detection (RED) slope graph
the exceed slope for the exceed 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:

config>qos#
slope-policy name [create]
   description description-string   
   exceed-slope
     max-avg percent
     max-prob percent     
     start-avg percent     
     no shutdown
high-slope
     max-avg percent
     max-prob percent
     start-avg percent
     no shutdown
highplus-slope
     max-avg percent
     max-prob percent
     start-avg percent
     no shutdown
low-slope
     max-avg percent
     max-prob percent
     start-avg percent
     no shutdown
time-average-factor value

The following displays the slope policy configuration:

A:ALA-7>config>qos# info
#------------------------------------------
echo "QoS Slope/Queue Policies Configuration"
#------------------------------------------
...
        slope-policy "slopePolicy1" create
            description "Test"
            highplus-slope
                no shutdown
            exit
            high-slope
                no shutdown
            exit
            low-slope
                no shutdown
            exit
            exceed-slope
                no shutdown
            exit
        exit
...
#------------------------------------------
A:ALA-7>config>qos#

Applying slope policies

Apply slope policies to the following entities:

Ports

The following CLI syntax examples may be used to apply slope policies to port pools:

config>port>access>ingress>pool>slope-policy name

config>port>access>egress>pool>slope-policy name

config>port>network>egress>pool>slope-policy name

FP

The following CLI syntax examples may be used to apply slope policies to FP ingress network pools:

config>card>fp>ingress>network>pool>slope-policy name

For multicast path management, the slope policy is configured within a multicast management bandwidth policy that is then applied under the FP ingress multicast-path-management.

config>mcast-mgmt>bw-plcy>mcast-pool percent-of-total <percent-of-buffers> resv-cbs <percent-of-pool> slope-policy <policy-name>

Default slope policy values

The default access ingress and egress policies are identified as policy-id 1. The default policies cannot be edited or deleted. Slope policy defaults lists default policy parameters.

Table 1. Slope policy defaults
Parameter		Default
Description		‟Default slope policy”
Highplus-slope
	shutdown	shutdown
	start-avg	85
	max-avg	100
	max-prob	80
High-slope
	shutdown	shutdown
	start-avg	70
	max-avg	90
	max-prob	80
Low-slope
	shutdown	shutdown
	start-avg	50
	max-avg	75
	max-prob	80
Exceed-slope
	shutdown	shutdown
	start-avg	30
	max-avg	55
	max-prob	80
Time-average-factor		7

The following output displays the default configuration:

A:ALA-7>config>qos>slope-policy# info detail
----------------------------------------------
            description "Default slope policy."
            highplus-slope
                shutdown
                start-avg 85
                max-avg 100
                max-prob 80
            exit
            high-slope
                shutdown
                start-avg 70
                max-avg 90
                max-prob 80
            exit
            low-slope
                shutdown
                start-avg 50
                max-avg 75
                max-prob 80
            exit
            exceed-slope
                shutdown
                start-avg 30
                max-avg 55
                max-prob 80
            exit
 time-average-factor 7
----------------------------------------------
A:ALA-7>config>qos>slope-policy#

Monitoring buffer pool usage

SR OS supports monitoring of buffer pool usage using the following commands.

configure card fp ingress network pool monitor-pool-depth
configure port access egress pool monitor-pool-depth
configure port access ingress pool monitor-pool-depth
configure port network egress pool monitor-pool-depth

These commands configure alarm thresholds, as percentages, for the following categories of buffer pool usage:

total-pool-in-use
shared-pool-in-use
reserved-pool-in-use

The system raises an SNMP alarm or ON_CHANGE notification when any configured threshold is exceeded.

The following example displays buffer pool monitoring thresholds for an ingress network pool on a card.

MD-CLI

[ex:/configure card 1 fp 1 ingress network pool "default"]
A:admin@node-2# info
    monitor-pool-depth {
        admin-state enable
        alarm-thresholds {
            total-pool-in-use 90.3
            shared-pool-in-use 88.6
            reserved-pool-in-use 85.9
        }
    }

classic CLI

A:node-2>config>card>fp>ingress>network>pool# info
----------------------------------------------
              monitor-pool-depth
                  enable
                  alarm-thresholds
                      total-pool-in-use 90.30
                      shared-pool-in-use 88.60
                      reserved-pool-in-use 85.90
                  exit
              exit
----------------------------------------------

Service management tasks

This section discusses slope QoS policy service management tasks:

Deleting QoS policies

A slope policy is associated by default with XMAs and MDAs and port access and network 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 of the places where it is applied.

A:ALA-7>config>qos# no slope-policy slopePolicy1
MINOR: QoS #1902 Slope policy has references
A:ALA-7>config>qos#

Ports

The following CLI syntax examples can be used to remove slope policies from port pools:

config>port>access>ingress>pool# no slope-policy name

config>port>access>egress>pool# no slope-policy name

config>port>network>egress>pool# no slope-policy name

FP

The following CLI syntax examples may be used to apply slope policies to FP ingress network pools:

config>card>fp>ingress>network>pool>slope-policy name

For multicast path management, the slope policy is configured within a multicast management bandwidth policy that is then applied under the FP ingress multicast-path-management.

config>mcast-mgmt>bw-plcy>mcast-pool percent-of-total <percent-of-buffers> resv-cbs <percent-of-pool> slope-policy <policy-name>

Removing a policy from the QoS configuration

To delete a slope policy, enter the following command:

config>qos# no slope-policy policy-id

Example:

config>qos# no slope-policy slopePolicy1

Copying and overwriting QoS policies

An existing slope policy can be copied, renamed with a new policy ID value, or used to overwrite an existing policy ID. The overwrite option must be specified or an error occurs if the destination policy ID exists.

config>qos>copy {slope-policy} source-policy-id dest-policy-id [overwrite]

Editing QoS policies

Existing policies and entries in the CLI or NMS can be edited. 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, then write over the original policy.