Configuring Modular QoS Congestion Avoidance

Congestion avoidance techniques monitor traffic flow in an effort to anticipate and avoid congestion at common network bottlenecks. Avoidance techniques are implemented before congestion occurs as compared with congestion management techniques that control congestion after it has occurred.

Congestion avoidance is achieved through packet dropping. Cisco IOS XR software supports these quality of service (QoS) congestion avoidance techniques that drop packets:

  • Random early detection (RED)

  • Weighted random early detection (WRED)

  • Tail drop

The module describes the concepts and tasks related to these congestion avoidance techniques.

Feature History for Configuring Modular QoS Congestion Avoidance on Cisco IOS XR Software

Release

Modification

Release 5.0.0

This feature was introduced.

Prerequisites for Configuring Modular QoS Congestion Avoidance

This prerequisite is required for configuring QoS congestion avoidance on your network:

You must be in a user group associated with a task group that includes the proper task IDs. The command reference guides include the task IDs required for each command. If you suspect user group assignment is preventing you from using a command, contact your AAA administrator for assistance.

Information About Configuring Modular QoS Congestion Avoidance

Random Early Detection and TCP

The Random Early Detection (RED) congestion avoidance technique takes advantage of the congestion control mechanism of TCP. By randomly dropping packets prior to periods of high congestion, RED tells the packet source to decrease its transmission rate. Assuming the packet source is using TCP, it decreases its transmission rate until all packets reach their destination, indicating that the congestion is cleared. You can use RED as a way to cause TCP to slow transmission of packets. TCP not only pauses, but it also restarts quickly and adapts its transmission rate to the rate that the network can support.

RED distributes losses in time and maintains normally low queue depth while absorbing traffic bursts. When enabled on an interface, RED begins dropping packets when congestion occurs at a rate you select during configuration.

Average Queue Size for WRED

The router automatically determines the parameters to use in the WRED calculations. The average queue size is based on the previous average and current size of the queue. The formula is:

average = (old_average * (1-2 -x)) + (current_queue_size * 2 -x)

where x is the exponential weight factor.

For high values of x, the previous average becomes more important. A large factor smooths out the peaks and lows in queue length. The average queue size is unlikely to change very quickly, avoiding a drastic change in size. The WRED process is slow to start dropping packets, but it may continue dropping packets for a time after the actual queue size has fallen below the minimum threshold. The slow-moving average accommodates temporary bursts in traffic.


Note

The exponential weight factor, x , is fixed and is not user configurable.



Note

If the value of x gets too high, WRED does not react to congestion. Packets are sent or dropped as if WRED were not in effect.


For low values of x, the average queue size closely tracks the current queue size. The resulting average may fluctuate with changes in the traffic levels. In this case, the WRED process responds quickly to long queues. Once the queue falls below the minimum threshold, the process stops dropping packets.

If the value of x gets too low, WRED overreacts to temporary traffic bursts and drops traffic unnecessarily.

Tail Drop and the FIFO Queue

Tail drop is a congestion avoidance technique that drops packets when an output queue is full until congestion is eliminated. Tail drop treats all traffic flow equally and does not differentiate between classes of service. It manages the packets that are unclassified, placed into a first-in, first-out (FIFO) queue, and forwarded at a rate determined by the available underlying link bandwidth.

See the “Default Traffic Class” section of the “Configuring Modular Quality of Service Packet Classification on Cisco IOS XR Software” module.

Configuring Random Early Detection

This configuration task is similar to that used for WRED except that the random-detect precedence command is not configured and the random-detect command with the default keyword must be used to enable RED.

SUMMARY STEPS

  1. configure
  2. policy-map policy-map-name
  3. class class-name
  4. random-detect {cos value | default | discard-class value | dscp value | exp value | precedence value | min-threshold [units] max-threshold [units] }
  5. bandwidth {bandwidth [units] | percent value} or bandwidth remaining [percent value | ratio ratio-value
  6. shape average {percent percentage | value [units]}
  7. exit
  8. exit
  9. interface type interface-path-id
  10. service-policy {input | output} policy-map
  11. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose
Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters XR Config mode.

Step 2

policy-map policy-map-name

Example:


RP/0/RP0/CPU0:router(config)# policy-map policy1

Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy and enters the policy map configuration mode.

Step 3

class class-name

Example:


RP/0/RP0/CPU0:router(config-pmap)# class class1

Specifies the name of the class whose policy you want to create or change and enters the policy map class configuration mode.

Step 4

random-detect {cos value | default | discard-class value | dscp value | exp value | precedence value | min-threshold [units] max-threshold [units] }

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# random-detect default

Enables RED with default minimum and maximum thresholds.

Step 5

bandwidth {bandwidth [units] | percent value} or bandwidth remaining [percent value | ratio ratio-value

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth percent 30

or


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth remaining percent 20

(Optional) Specifies the bandwidth allocated for a class belonging to a policy map.

or

(Optional) Specifies how to allocate leftover bandwidth to various classes.

Step 6

shape average {percent percentage | value [units]}

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# shape average percent 50

(Optional) Shapes traffic to the specified bit rate or a percentage of the available bandwidth.

Step 7

exit

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# exit

Returns the router to policy map configuration mode.

Step 8

exit

Example:


RP/0/RP0/CPU0:router(config-pmap)# exit

Returns the router to XR Config mode.

Step 9

interface type interface-path-id

Example:


RP/0/RP0/CPU0:router(config)# interface HundGigE
 0/2/0/0

Enters the configuration mode and configures an interface.

Step 10

service-policy {input | output} policy-map

Example:


RP/0/RP0/CPU0:router(config-if)# service-policy output policy1

Attaches a policy map to an input or output interface to be used as the service policy for that interface. In this example, the traffic policy evaluates all traffic leaving that interface.

Step 11

Use the commit or end command.

commit —Saves the configuration changes, and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration mode, without committing the configuration changes.

Configuring Weighted Random Early Detection

WRED drops packets selectively based on IP precedence. Edge routers assign IP precedences to packets as they enter the network. WRED uses these precedences to determine how to treat different types of traffic.

When a packet arrives, the following actions occur:

  • The average queue size is calculated.

  • If the average queue size is less than the minimum queue threshold, the arriving packet is queued.

  • If the average queue size is between the minimum queue threshold for that type of traffic and the maximum threshold for the interface, the packet is either dropped or queued, depending on the packet drop probability for that type of traffic.

  • If the average queue size is greater than the maximum threshold, the packet is dropped.

Restrictions

  • You cannot configure WRED in a class that has been set for priority queueing (PQ).

  • You cannot use the random-detect command in a class configured with the priority command.

SUMMARY STEPS

  1. configure
  2. policy-map policy-name
  3. class class-name
  4. random-detect dscp dscp-value min-threshold [units] max-threshold [units]
  5. bandwidth {bandwidth [units] | percent value} or bandwidth remaining [percent value | ratio ratio-value]
  6. bandwidth {bandwidth [units] | percent value}
  7. bandwidth remaining percent value
  8. shape average {percent percentage | value [units]}
  9. queue-limit value [units]
  10. exit
  11. interface type interface-path-id
  12. service-policy {input | output} policy-map
  13. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose
Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters XR Config mode.

Step 2

policy-map policy-name

Example:


RP/0/RP0/CPU0:router(config)# policy-map policy1

Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy and enters the policy map configuration mode.

Step 3

class class-name

Example:


RP/0/RP0/CPU0:router(config-pmap)# class class1

Specifies the name of the class whose policy you want to create or change and enters the policy map class configuration mode.

Step 4

random-detect dscp dscp-value min-threshold [units] max-threshold [units]

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# random-detect dscp af11 1000000 bytes 2000000 bytes

Modifies the minimum and maximum packet thresholds for the DSCP value.

  • Enables RED.

  • dscp-value—Number from 0 to 63 that sets the DSCP value. Reserved keywords can be specified instead of numeric values.

  • min-threshold—Minimum threshold in the specified units. The value range of this argument is from 0 to 1073741823. When the average queue length reaches the minimum threshold, WRED randomly drops some packets with the specified DSCP value.

  • max-threshold—Maximum threshold in the specified units. The value range of this argument is from 0 to 1073741823. When the average queue length exceeds the maximum threshold, WRED drops all packets with the specified DSCP value.

  • units —Units of the threshold value. This can be bytes , gbytes , kbytes , mbytes , ms (milliseconds), packets , or us (microseconds). The default is packets .

  • This example shows that for packets with DSCP AF11, the WRED minimum threshold is 1,000,000 bytes and maximum threshold is 2,000,000 bytes.

Step 5

bandwidth {bandwidth [units] | percent value} or bandwidth remaining [percent value | ratio ratio-value]

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth percent 30

or


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth remaining percent 20

(Optional) Specifies the bandwidth allocated for a class belonging to a policy map.

or

(Optional) Specifies how to allocate leftover bandwidth to various classes.

Step 6

bandwidth {bandwidth [units] | percent value}

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth percent 30

(Optional) Specifies the bandwidth allocated for a class belonging to a policy map. This example guarantees 30 percent of the interface bandwidth to class class1.

Step 7

bandwidth remaining percent value

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth remaining percent 20

(Optional) Specifies how to allocate leftover bandwidth to various classes.

  • The remaining bandwidth of 70 percent is shared by all configured classes.

  • In this example, class class1 receives 20 percent of the 70 percent.

Step 8

shape average {percent percentage | value [units]}

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# shape average percent 50

(Optional) Shapes traffic to the specified bit rate or a percentage of the available bandwidth.

Step 9

queue-limit value [units]

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# queue-limit 50 ms

(Optional) Changes queue-limit to fine-tune the amount of buffers available for each queue. The default queue-limit is 100 ms of the service rate for a non-priority class and 10ms of the service rate for a priority class.

Note 

Even though this command is optional, it is recommended that you use it to fine-tune the queue limit, instead of relying on your system default settings. If the queue limit is too large, the buffer consumption goes up, resulting in delays. On the other hand, too small a queue limit may result in extra drops while allowing for faster rate adaption.

Step 10

exit

Example:


RP/0/RP0/CPU0:router(config-pmap)# exit

Returns the router to XR Config mode.

Step 11

interface type interface-path-id

Example:


RP/0/RP0/CPU0:router(config)# interface HundGigE
0/2/0/0

Enters the configuration mode and configures an interface.

Step 12

service-policy {input | output} policy-map

Example:


RP/0/RP0/CPU0:router(config-if)# service-policy output policy1

Attaches a policy map to an input or output interface to be used as the service policy for that interface.

  • In this example, the traffic policy evaluates all traffic leaving that interface.

  • Ingress policies are not valid; the bandwidth and bandwidth remaining commands cannot be applied to ingress policies.

Step 13

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Configuring Tail Drop

Packets satisfying the match criteria for a class accumulate in the queue reserved for the class until they are serviced. The queue-limit command is used to define the maximum threshold for a class. When the maximum threshold is reached, enqueued packets to the class queue result in tail drop (packet drop).

The queue-limit value uses the guaranteed service rate (GSR) of the queue as the reference value for the queue_bandwidth . If the class has bandwidth percent associated with it, the queue-limit is set to a proportion of the bandwidth reserved for that class.

If the GSR for a queue is zero, use the following to compute the default queue-limit :

  • 1 percent of the interface bandwidth for queues in a nonhierarchical policy.

  • 1 percent of parent maximum reference rate for hierarchical policy.

    The parent maximum reference rate is the minimum of parent shape, policer maximum rate, and the interface bandwidth.

The default queue-limit is set as follows:
  • For priority class

    default queue limit (in bytes) = (<10> ms * queue_bandwidth kbps) / 8

  • For non-priority class

    default queue limit (in bytes) = (<100> ms * queue_bandwidth kbps) / 8


Note

You can configure a maximum queue threshold up to 1GB, which translates to 80ms of buffering for 100Gbps queue.


Restrictions

  • When configuring the queue-limit command in a class, you must configure one of the following commands: priority , shape average , bandwidth , or bandwidth remaining , except for the default class.

SUMMARY STEPS

  1. configure
  2. policy-map policy-name
  3. class class-name
  4. queue-limit value [units]
  5. priority[ level priority-level ]
  6. police rate percent percentage
  7. class class-name
  8. bandwidth {bandwidth [units] | percent value}
  9. bandwidth remaining percent value
  10. exit
  11. exit
  12. interface type interface-path-id
  13. service-policy {input | output} policy-map
  14. Use the commit or end command.

DETAILED STEPS

  Command or Action Purpose
Step 1

configure

Example:


RP/0/RP0/CPU0:router# configure

Enters XR Config mode.

Step 2

policy-map policy-name

Example:


RP/0/RP0/CPU0:router(config)# policy-map policy1

Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy and also enters the policy map configuration mode.

Step 3

class class-name

Example:


RP/0/RP0/CPU0:router(config-pmap)# class class1

Specifies the name of the class whose policy you want to create or change and enters the policy map class configuration mode.

Step 4

queue-limit value [units]

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# queue-limit 1000000 bytes

Specifies or modifies the maximum the queue can hold for a class policy configured in a policy map. The default value of the units argument is packets . In this example, when the queue limit reaches 1,000,000 bytes, enqueued packets to the class queue are dropped.

Step 5

priority[ level priority-level ]

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# priority level 1

Specifies priority to a class of traffic belonging to a policy map.

Step 6

police rate percent percentage

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# police rate percent 30

Configures traffic policing.

Step 7

class class-name

Example:


RP/0/RP0/CPU0:router(config-pmap)# class class2

Specifies the name of the class whose policy you want to create or change. In this example, class2 is configured.

Step 8

bandwidth {bandwidth [units] | percent value}

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth percent 30

(Optional) Specifies the bandwidth allocated for a class belonging to a policy map. This example guarantees 30 percent of the interface bandwidth to class class2.

Step 9

bandwidth remaining percent value

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# bandwidth remaining percent 20

(Optional) Specifies how to allocate leftover bandwidth to various classes. This example allocates 20 percent of the leftover interface bandwidth to class class2.

Step 10

exit

Example:


RP/0/RP0/CPU0:router(config-pmap-c)# exit

Returns the router to policy map configuration mode.

Step 11

exit

Example:


RP/0/RP0/CPU0:router(config-pmap)# exit

Returns the router to XR Config mode.

Step 12

interface type interface-path-id

Example:


RP/0/RP0/CPU0:router(config)# interface HundredGigE 0/7/0/0

Enters the configuration mode and configures an interface.

Step 13

service-policy {input | output} policy-map

Example:


RP/0/RP0/CPU0:router(config-if)# service-policy output policy1

Attaches a policy map to an input or output interface to be used as the service policy for that interface. In this example, the traffic policy evaluates all traffic leaving that interface.

Step 14

Use the commit or end command.

commit —Saves the configuration changes and remains within the configuration session.

end —Prompts user to take one of these actions:
  • Yes — Saves configuration changes and exits the configuration session.

  • No —Exits the configuration session without committing the configuration changes.

  • Cancel —Remains in the configuration session, without committing the configuration changes.

Additional References

These sections provide references related to implementing QoS congestion avoidance.

Standards

Standards

Title

No new or modified standards are supported by this feature, and support for existing standards has not been modified by this feature.

RFCs

RFCs

Title

No new or modified RFCs are supported by this feature, and support for existing RFCs has not been modified by this feature.

Technical Assistance

Description

Link

The Cisco Technical Support website contains thousands of pages of searchable technical content, including links to products, technologies, solutions, technical tips, and tools. Registered Cisco.com users can log in from this page to access even more content.

http://www.cisco.com/cisco/web/support/index.html