Table Of Contents

Quality of Service Commands

access-list rate-limit

bandwidth (policy-map class)

bgp-policy

bump

bundle

class (policy-map)

class-bundle

class-map

clear ip rsvp reservation

clear ip rsvp sender

custom-queue-list

exponential-weighting-constant

fair-queue (class-default)

fair-queue (DWFQ)

fair-queue (WFQ)

fair-queue aggregate-limit

fair-queue individual-limit

fair-queue limit

fair-queue qos-group

fair-queue tos

fair-queue weight

frame-relay ip rtp priority

Quality of Service Commands

---

Use the commands in this chapter to configure quality of service (QoS), a measure of performance for a transmission system that reflects its transmission quality and service availability. The commands are arranged alphabetically.

For QoS configuration information and examples, refer to the Cisco IOS Quality of Service Solutions Configuration Guide.

access-list rate-limit

To configure an access list for use with committed access rate (CAR) policies, use the access-list rate-limit global configuration command. To remove the access list from the configuration, use the no form of this command.

access-list rate-limit acl-index {precedence | mac-address | mask prec-mask}

no access-list rate-limit acl-index {precedence | mac-address | mask prec-mask}

Syntax Description

acl-index	Access list number. Use any number from 1 to 99 to classify packets by precedence or precedence mask, and use any number from 100 to 199 to classify by MAC address.
precedence	IP Precedence.
mac-address	Address of the MAC.
mask prec-mask	IP Precedence mask; a two-digit hexadecimal number. Use this option when you want to assign multiple precedences to the same rate-limit access list.

Defaults

No CAR access lists are configured.

Command Modes

Global configuration

Command History

Release	Modification
11.1 CC	This command was introduced.

Usage Guidelines

This command classifies packets by the specified IP Precedence or MAC address for a particular CAR access list. You can then apply CAR policies, using the rate-limit command, to individual rate-limit access lists. Thus, packets with different IP Precedences or MAC addresses are treated differently by the CAR process.

You can specify only one command for each rate-limit access list. If you enter this command multiple times with the same access list number, the new command will overwrite the previous command.

Use the mask keyword to assign multiple IP Precedences to the same rate-limit access list. To determine the mask value, perform the following steps:

---

Step 1 Decide which precedences you want to assign to this rate-limit access list.

Step 2 Convert the precedences into an 8-bit number with each bit corresponding to one precedence. For example, an IP Precedence of 0 corresponds to 00000001, 1 corresponds to 00000010, 6 corresponds to 01000000, and 7 corresponds to 10000000.

Step 3 Add the 8-bit numbers for the selected precedences. For example, the mask for precedences 1 and 6 is 01000010.

Step 4 The command expects hexadecimal format. Convert the binary mask into the corresponding hexadecimal number. For example, 01000010 becomes 42. This value is used in the access-list rate-limit command. Any packets that have an IP Precedence of 1 or 6 will match this access list.

---

A mask of FF matches any precedence, and 00 does not match any precedence.

Examples

The following example assigns any packets with a MAC address of 00e0.34b0.7777 to rate-limit access list 100:

access-list rate-limit 100 00e0.34b0.7777

The following example assigns packets with an IP Precedence of 0, 1, or 2 to the rate-limit access list 25:

access-list rate-limit 25 mask 07

Related Commands

Command	Description
show access-lists rate-limit	Displays information about rate-limit access lists.
show ip cef	Displays entries in the FIB that are unresolved or displays a FIB summary.

bandwidth (policy-map class)

To specify or modify the bandwidth allocated for a class belonging to a policy map, use the bandwidth policy-map class configuration command. To remove the bandwidth specified for a class, use the no form of this command.

bandwidth {bandwidth-kbps | percent percent}

no bandwidth {bandwidth-kbps | percent percent}

Syntax Description

bandwidth-kbps	Amount of bandwidth (in kbps) to be assigned to the class.
percent percent	Percentage of available bandwidth to be assigned to the class.

Defaults

No default behavior.

Command Modes

Policy-map class configuration

Command History

Release	Modification
12.0(5)T	This command was introduced.
12.1(1)	The percent keyword was added.

Usage Guidelines

You use the bandwidth command when you configure a policy map for a class defined by the class-map command. The bandwidth command specifies the bandwidth for traffic in that class. Class-based weighted fair queueing (CBWFQ) derives the weight for packets belonging to the class from the bandwidth allocated to the class. CBWFQ then uses the weight to ensure that the queue for the class is serviced fairly.

Besides specifying the amount of bandwidth in kbps, you can assign bandwidth as a percentage of the available bandwidth. During periods of congestion, the classes are serviced in proportion to their configured bandwidth percentages. Available bandwidth is equal to the interface bandwidth minus the sum of all bandwidths reserved by Resource Reservation Protocol (RSVP), IP RTP Priority, and low latency queueing (LLQ).

---

Note It is important to remember that hard bandwidth guarantees may not be provided and only relative bandwidths are assured. That is, class bandwidths are always proportional to the specified percentages of the interface bandwidth. When the link bandwidth is fixed, class bandwidth guarantees are in proportion to the configured percentages. If the link bandwidth is unknown or variable, class bandwidth guarantees cannot be computed.

---

Configuring bandwidth in percentages is most useful when the underlying link bandwidth is unknown or the relative class bandwidth distributions are known. For interfaces that have adaptive shaping rates (such as available bit rate [ABR] virtual circuits), CBWFQ can be configured by configuring class bandwidths in percentages.

The following restrictions apply to the bandwidth command:

•If the percent keyword is used, the sum of the class bandwidth percentages cannot exceed 100 percent.

•The amount of bandwidth configured should be large enough to also accommodate Layer 2 overhead.

•A policy map can have all the class bandwidths specified in kbps or all the class bandwidths specified in percentages—but not a mix of both.

•The IP RTP Priority and RSVP features can only be configured in kbps.

•The priority class inside LLQ can have bandwidth specified only in kbps. The nonpriority classes inside LLQ can have bandwidths specified either in percentages or in kbps, but not a mix of both.

For more information on bandwidth allocation, refer to the chapter "Congestion Management Overview" in the Cisco IOS Quality of Service Solutions Configuration Guide.

Note that when the policy map containing class policy configurations is attached to the interface to stipulate the service policy for that interface, available bandwidth is assessed. If a policy map cannot be attached to a particular interface because of insufficient interface bandwidth, then the policy is removed from all interfaces to which it was successfully attached.

Examples

The following example modifies the bandwidth for a class called acl22. The default class belongs to a service policy map called polmap6.

policy-map polmap6

 class acl22

  bandwidth 2000

  queue-limit 30

CBWFQ Bandwidth Guarantee

The following example illustrates how bandwidth is guaranteed when only CBWFQ is configured:

! The following commands create a policy map with two classes:

policy-map policy1

 class class1

  bandwidth percent 50

  exit

 class class2

  bandwidth percent 25

  exit

 end

!The following commands attach the policy to interface s3/2:

interface s3/2

 service output policy1

 end

The following output from the show policy-map interface command shows that 50 percent of the interface bandwidth is guaranteed for class1 and 25 percent is guaranteed for class2:

Router# show policy-map interface s3/2

 Serial3/2  output :policy1

  Class class1

   Weighted Fair Queueing

       Output Queue:Conversation 265

         Bandwidth 50 (%) Packets Matched 0 Max Threshold 64 (packets)

         (discards/tail drops) 0/0

  Class class2

   Weighted Fair Queueing

       Output Queue:Conversation 266

         Bandwidth 25 (%) Packets Matched 0 Max Threshold 64 (packets)

         (discards/tail drops) 0/0

In this example, the entire interface bandwidth is available for CBWFQ because RSVP, IP RTP Priority, and LLQ are not enabled. If this policy map is attached to a physical interface, the available bandwidth is equal to the link bandwidth. During periods of congestion, 50 percent of the link bandwidth is guaranteed to class1 and 25 percent of the link bandwidth is guaranteed to class2. For example, if this policy map was attached to a 1 Mbps link, class1 would be guaranteed 500 kbps and class2 would be guaranteed 250 kbps during periods of congestion.

CBWFQ and LLQ Bandwidth Allocation

This example illustrates how bandwidth is guaranteed if LLQ is configured with CBWFQ. Remember, the available bandwidth for CBWFQ is the link bandwidth minus the sum of the bandwidths reserved by RSVP, LLQ, and IP RTP Priority.

In this example, LLQ is enabled in a third class called voice1:

! The following commands create a policy map with three classes:

policy map policy1

 class class1

  bandwidth percent 50

  exit

 class class2

  bandwidth percent 25

  exit

 end

 class voice1

  priority 500

  exit

 end

!The following commands attach the policy to interface s3/2:

interface s3/2

 service output policy1

 end

The following output from the show policy-map command shows that 50 percent of the interface bandwidth is guaranteed for class1, 25 percent is guaranteed for class2, and 500 kbps is guaranteed for voice1:

Router# show policy-map policy1

   Policy Map policy1

   Class class1

   Weighted Fair Queueing

       Bandwidth 50 (%)  Max Threshold 64 (packets)

   Class class2

   Weighted Fair Queueing

       Bandwidth 25 (%)  Max Threshold 64 (packets)

   Class voice1

   Weighted Fair Queueing

       Strict Priority

       Bandwidth 500 (kbps)  Max Threshold 64 (packets)

Because LLQ reserved 500 kbps of the interface bandwidth, if you attach this policy map to an interface with 2 Mbps, only 1.5 Mbps is available for CBWFQ classes. In this example, 50 percent of 1.5 Mbps (750 kbps) is guaranteed for class1 and 25 percent (375 kbps) is guaranteed for class2. The remaining 25 percent of the available bandwidth (375 kbps) is shared by class1, class2, and any best-effort traffic.

Related Commands

Command	Description
class (policy-map)	Specifies the name of the class whose policy you want to create or change, and the default class (commonly known as the class-default class) before you configure its policy.
max-reserved-bandwidth	Changes the percent of interface bandwidth allocated for CBWFQ, LLQ, and IP RTP Priority.
policy-map	Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy.
queue-limit	Specifies or modifies the maximum number of packets the queue can hold for a class policy configured in a policy map.
random-detect (interface)	Enables WRED or DWRED.

bgp-policy

To enable the Policy Propagation via Border Gateway Protocol (BGP) feature on the interface, use the bgp-policy interface configuration command. To disable the Policy Propagation via BGP feature, use the no form of this command.

bgp-policy ip-prec-map

no bgp-policy ip-prec-map

Syntax Description

ip-prec-map

QoS policy based on the IP Precedence.

Defaults

The Policy Propagation via BGP feature is disabled.

Command Modes

Interface configuration

Command History

Release	Modification
11.1 CC	This command was introduced.

Usage Guidelines

For the Policy Propagation via BGP feature to work, you must enable BGP and either Cisco Express Forwarding (CEF) or distributed CEF (dCEF). In addition, the proper route-map configuration must be in place to specify the IP Precedence (for example, the set ip precedence route-map configuration command).

---

Note If you specify both source and destination keywords on the interface, the software looks up the source address in the routing table and classifies the packet based on the source address first; then the software looks up the destination address in the routing table and reclassifies the packet based on the destination address.

---

To display QoS policy information for the interface, use the show ip interface command.

Examples

The following example enables the Policy Propagation via BGP feature on an interface based on the source address and the IP Precedence setting:

configure terminal

 interface ethernet 4/0/0

  bgp-policy ip-prec-map

  end

bump

To configure the bumping rules for a virtual circuit (VC) class that can be assigned to a VC bundle, use the bump vc-class configuration command. To remove the explicit bumping rules for the VCs assigned this class and default them to implicit bumping, use the no bump explicit command. To specify that the VC bundle members do not accept any bumped traffic, use the no bump traffic command.

To configure the bumping rules for a specific VC member of a bundle, use the bump bundle-vc configuration command. To remove the explicit bumping rules for the VC and default it to implicit bumping, use the no bump explicit command. To specify that the VC does not accept any bumped traffic, use the no form of this command.

bump {implicit | explicit precedence-level | traffic}

no bump {explicit precedence-level | traffic}

Syntax Description

implicit	Depending on the mode, applies implicit bumping rules, which is also the default, to a single VC bundle member (bundle-vc mode) or all VCs in the bundle (bundle mode). The (default) implicit bumping rule stipulates that bumped traffic is to be carried by a VC with a lower precedence.
explicit precedence-level	Specifies the precedence level to which traffic on a VC (bundle-vc mode) will be bumped when the VC goes down. Specifies a single number as the value of precedence-level.
traffic	In its positive form, specifies that the VC accepts bumped traffic. The no form stipulates that the VC does not accept any bumped traffic.

Defaults

Implicit bumping.

Bump traffic (VCs accept bumped traffic).

Command Modes

VC-class configuration (for a VC class).

Bundle-vc configuration (for a VC bundle member).

Command History

Release	Modification
12.0(3)T	This command was introduced.

Usage Guidelines

Use the bump command in bundle-vc configuration mode to configure bumping rules for a discrete VC bundle member or in vc-class configuration mode to configure a VC class that can be assigned to a bundle member.

The effects of different bumping configuration approaches are as follows:

•Implicit bumping: If you configure implicit bumping, bumped traffic is sent to the VC configured to handle the next lower precedence level. When the original VC that bumped the traffic comes back up, traffic it is configured to carry is restored to it. When no other positive forms of the bump command are configured, the bump implicit command takes effect.

•Explicit bumping: If you configure a VC with the bump explicit command, you can specify the precedence level to which traffic on a VC will be bumped when that VC goes down, and the traffic will be directed to a VC mapped with that precedence level. If the VC that picks up and carries the traffic goes down, the traffic is subject to the bumping rules for that VC. You can specify only one precedence level for bumping.

•Bumped traffic: The VC accepts bumped traffic. You can configure bumped traffic explicitly using either the bump traffic or the no bump traffic command, or let the default take effect by specifying neither.

•No bumped traffic: To configure a discrete VC to reject bumped traffic when the traffic is directed to the VC, use the no bump traffic command.

---

Note When no alternative VC can be found to handle bumped traffic, the bundle is declared down. To avoid this occurrence, configure explicitly the bundle member VC that has the lowest precedence level.

---

To use this command in vc-class configuration mode, you must enter the vc-class atm global configuration command before you enter this command.

To use this command to configure an individual bundle member in bundle-vc configuration mode, first enter the bundle command to enact bundle configuration mode for the bundle to which you want to add or modify the VC member to be configured. Then, use the pvc-bundle command to specify the VC to be created or modified and enter bundle-vc configuration mode.

This command has no effect if the VC class that contains the command is attached to a standalone VC, that is, if the VC is not a bundle member. In this case, the attributes are ignored by the VC.

VCs in a VC bundle are subject to the following configuration inheritance rules (listed in order of next highest precedence):

•VC configuration in bundle-vc mode

•Bundle configuration in bundle mode (with effect of assigned vc-class configuration)

•Subinterface configuration in subinterface mode

Examples

The following example configures the class premium-class to define parameters applicable to a VC in a bundle. Unless overridden with a bundle-vc bump configuration, the VC that uses this class will not allow other traffic to be bumped onto it.

vc-class atm premium-class

 no bump traffic 
 bump explicitly 7

Related Commands

Command	Description
class-vc	Assigns a VC class to an ATM PVC, SVC, or VC bundle member.
precedence	Configures precedence levels for a VC class that can be assigned to a VC bundle and thus applied to all VC members of that bundle.
protect	Configures a VC class with protected group or protected VC status for application to a VC bundle member.
ubr	Configures UBR QoS and specifies the output peak cell rate for an ATM PVC, SVC, VC class, or VC bundle member.
ubr+	Configures UBR QoS and specifies the output peak cell rate and output minimum guaranteed cell rate for an ATM PVC, SVC, VC class, or VC bundle member.
vbr-nrt	Configures the VBR-NRT QoS and specifies output peak cell rate, output sustainable cell rate, and output maximum burst cell size for an ATM PVC, SVC, VC class, or VC bundle member.

bundle

To create a bundle or modify an existing bundle to enter bundle configuration mode, use the bundle subinterface configuration command. To remove the specified bundle, use the no form of this command.

bundle bundle-name

no bundle bundle-name

Syntax Description

bundle-name

Specifies the name of the bundle to be created. Limit is 16 alphanumeric characters.

Defaults

None

Command Modes

Subinterface configuration

Command History

Release	Modification
12.0(3)T	This command was introduced.

Usage Guidelines

From within bundle configuration mode you can configure the characteristics and attributes of the bundle and its members, such as the encapsulation type for all virtual circuits (VCs) in the bundle, the bundle management parameters, the service type, and so on. Attributes and parameters you configure in bundle configuration mode are applied to all VC members of the bundle.

VCs in a VC bundle are subject to the following configuration inheritance rules (listed in order of next highest precedence):

•VC configuration in bundle-vc mode

•Bundle configuration in bundle mode

•Subinterface configuration in subinterface mode

To display status on bundles, use the show atm bundle and show atm bundle statistics commands.

Examples

The following example configures a bundle called new-york. The example specifies the IP address of the subinterface and the router protocol—the router uses IS-IS as an IP routing protocol—then configures the bundle.

interface a1/0.1 multipoint 
  ip address 10.0.0.1 255.255.255.0 
  ip router isis 
  bundle new-york

Related Commands

Command	Description
class-bundle	Configures a VC bundle with the bundle-level commands contained in the specified VC class.
oam-bundle	Enables end-to-end F5 OAM loopback cell generation and OAM management for all VC members of a bundle, or for a VC class that can be applied to a VC bundle.
pvc-bundle	Adds a PVC to a bundle as a member of the bundle and enters bundle-vc configuration mode in order to configure that PVC bundle member.

class (policy-map)

To specify the name of the class whose policy you want to create or change or to specify the default class (commonly known as the class-default class) before you configure its policy, use the class policy-map configuration command. To remove a class from the policy map, use the no form of this command.

class {class-name | class-default}

no class {class-name | class-default}

Syntax Description

class-name	The name of the class for which you want to configure or modify policy.
class-default	Specifies the default class so that you can configure or modify its policy.

Defaults

No default behavior.

Command Modes

Policy-map configuration

Command History

Release	Modification
12.0(5)T	This command was introduced.

Usage Guidelines

Enter the policy-map command to identify the policy map and enter policy map configuration mode before you use the class command. After you specify a policy map, you can configure policy for new classes or modify policy for any existing classes in that policy map.

The class name that you specify in the policy map ties the characteristics for that class—that is, its policy—to the class map and its match criteria, as configured using the class-map command.

When you configure policy for a class and specify its bandwidth and attach the policy map to an interface, class-based weighted fair queueing (CBWFQ) determines if the bandwidth requirement of the class can be satisfied. If so, CBWFQ creates the necessary internal data structures to maintain state for the class and allocates a queue for it.

When a class is removed, available bandwidth for the interface is incremented by the amount previously allocated to the class.

The maximum number of classes you can configure for a router—and, therefore, within a policy map—is 64.

The predefined default class called class-default is the class to which traffic is directed if that traffic does not satisfy the match criteria of other classes whose policy is defined in the policy map.

You can define a class policy to use either tail drop (by using the queue-limit command) or Weighted Random Early Detection (WRED) packet drop (by using the random-detect command). You cannot use the queue-limit and random-detect commands in the same class policy, but they can be used in two class policies in the same policy map.

You can configure the bandwidth command when either the queue-limit or the random-detect command is configured in a class policy. The bandwidth command specifies the amount of bandwidth allocated for the class.

For the default class, you can configure the fair-queue (class-default) command. The fair-queue command specifies the number of dynamic queues for the default class. The fair-queue command can be used in the same class policy as either the queue-limit or random-detect command. It cannot be used with the bandwidth command.

Examples

The following example configures three class policies included in the policy map called policy1. Class1 specifies policy for traffic that matches access control list 136. Class2 specifies policy for traffic on the interface ethernet101. The third class is the default class to which packets that do not satisfy configured match criteria are directed.

! The following commands create class-maps class1 and class2 

! and define their match criteria:

class-map class1

 match access-group 136

class-map class2

 match input-interface ethernet101

! The following commands create the policy map, which is defined to contain policy

! specification for class1, class2, and the default class:

policy-map policy1

class class1

 bandwidth 2000

 queue-limit 40

class class2

 bandwidth 3000

 random-detect

 random-detect exponential-weighting-constant 10

class class-default

 fair-queue 16

 queue-limit 20

Class1 has these characteristics: A minimum of 2000 kbps of bandwidth are expected to be delivered to this class in the event of congestion, and the queue reserved for this class can enqueue 40 packets before tail drop is enacted.

Class2 has these characteristics: A minimum of 3000 kbps of bandwidth are expected to be delivered to this class in the event of congestion, and a weight factor of 10 is used to calculate the average queue size. For congestion avoidance, WRED packet drop is used, not tail drop.

The default class has these characteristics: 16 dynamic queues are reserved for traffic that does not meet the match criteria of other classes whose policy is defined by the policy map policy1, and a maximum of 20 packets per queue are enqueued before tail drop is enacted to handle additional packets.

---

Note Note that when the policy map containing these classes is attached to the interface to stipulate the service policy for that interface, available bandwidth is assessed, taking into account all class policies and Resource Reservation Protocol (RSVP), if configured.

---

The following example configures policy for the default class included in the policy map called policy2. The default class has these characteristics: 20 dynamic queues are available for traffic that does not meet the match criteria of other classes whose policy is defined by the policy map policy2, and a weight factor of 14 is used to calculate the average queue size. For congestion avoidance, WRED packet drop is used, not tail drop.

policy-map policy2

class class-default

 fair-queue 20

 random-detect

 random-detect exponential-weighting-constant 14

Related Commands

Command	Description
bandwidth (policy-map class)	Specifies or modifies the bandwidth allocated for a class belonging to a policy map.
class-map	Creates a class map to be used for matching packets to a specified class.
policy-map	Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy.
queue-limit	Specifies or modifies the maximum number of packets the queue can hold for a class policy configured in a policy map.
random-detect (interface)	Enables WRED or DWRED.
random-detect exponential-weighting-constant	Configures the WRED and DWRED exponential weight factor for the average queue size calculation.
random-detect precedence	Configures WRED and DWRED parameters for a particular IP Precedence.

class-bundle

To configure a virtual circuit (VC) bundle with the bundle-level commands contained in the specified VC class, use the class-bundle bundle configuration command. The no form of this command removes the VC class parameters from a VC bundle.

class-bundle vc-class-name

no class-bundle vc-class-name

Syntax Description

vc-class-name

Name of the VC class you are assigning to your VC bundle.

Defaults

No VC class is assigned to the VC bundle.

Command Modes

Bundle configuration

Command History

Release	Modification
12.0 T	This command was introduced, replacing the class command for configuring ATM VC bundles.

Usage Guidelines

To use this command, you must first enter the bundle command to create the bundle and enter bundle configuration mode.

Use this command to assign a previously defined set of parameters (defined in a VC class) to an ATM VC bundle. Parameters set through bundle-level commands contained in a VC class are applied to the bundle and its VC members.

You can add the following commands to a VC class to be used to configure a VC bundle: oam-bundle, broadcast, encapsulation, protocol, oam retry, and inarp.

Bundle-level parameters applied through commands configured directly on a bundle supersede bundle-level parameters applied through a VC class by the class-bundle command. Some bundle-level parameters applied through a VC class or directly to the bundle can be superseded by commands that you directly apply to individual VCs in bundle-vc configuration mode.

Examples

In the following example, a class called class1 is first created and then applied to the bundle called bundle1:

! The following commands create the class class1:

vc-class atm class1

 encapsulation aal5snap

 broadcast

 protocol ip inarp

 oam-bundle manage 3

 oam 4 3 10

! The following commands apply class1 to the bundle called bundle1:

bundle bundle1

 class-bundle class1

Taking into account hierarchy precedence rules, VCs belonging to the bundle1 bundle will be characterized by these parameters: aal5snap, encapsulation, broadcast on, use of Inverse Address Resolution Protocol (ARP) to resolve IP addresses, and Operation, Administration, and Maintenance (OAM) enabled.

Related Commands

Command	Description
broadcast	Configures broadcast packet duplication and transmission for an ATM VC class, PVC, SVC, or VC bundle.
bundle	Creates a bundle or modifies an existing bundle to enter bundle configuration mode.
class-int	Assigns a VC class to an ATM main interface or subinterface.
class-vc	Assigns a VC class to an ATM PVC, SVC, or VC bundle member.
encapsulation	Sets the encapsulation method used by the interface.
inarp	Configures the Inverse ARP time period for an ATM PVC, VC class, or VC bundle.
oam-bundle	Enables end-to-end F5 OAM loopback cell generation and OAM management for all VC members of a bundle, or for a VC class that can be applied to a VC bundle.
oam retry	Configures parameters related to OAM management for an ATM PVC, SVC, VC class, or VC bundle.
protocol (ATM)	Configures a static map for an ATM PVC, SVC, VC class, or VC bundle. Enables Inverse ARP or Inverse ARP broadcasts on an ATM PVC by either configuring Inverse ARP directly on the PVC, on the VC bundle, or in a VC class (applies to IP and IPX protocols only).
pvc-bundle	Adds a PVC to a bundle as a member of the bundle and enters bundle-vc configuration mode in order to configure that PVC bundle member.

Command	Description
show atm bundle	Displays the bundle attributes assigned to each bundle VC member and the current working status of the VC members.
show atm bundle statistics	Displays statistics on the specified bundle.
show atm map	Displays the list of all configured ATM static maps to remote hosts on an ATM network.
vc-class atm	Configures a VC class for an ATM VC or interface.

class-map

To create a class map to be used for matching packets to the class whose name you specify, use the class-map global configuration command. To remove an existing class map from the router, use the no form of this command.

class-map [match-all | match-any] class-map-name

no class-map [match-all | match-any] class-map-name

Syntax Description

match-all | match-any	Determines how packets are evaluated when multiple match criteria exist. Packets must either meet all of the match criteria (match-all) or one of the match criteria (match-any) in order to be considered a member of the class.
class-map-name	Name of the class for the class map. The name can be a maximum of 40 alphanumeric characters. The class name is used for both the class map and to configure policy for the class in the policy map.

Defaults

No default behavior.

Command Modes

Global configuration

Command History

Release	Modification
12.0(5)T	This command was introduced.

Usage Guidelines

Use this command to specify the name of the class for which you want to create or modify class map match criteria. Use of the class-map command enables class-map configuration mode in which you can enter one of the match commands to configure the match criteria for this class. Packets arriving at the output interface are checked against the match criteria configured for a class map to determine if the packet belongs to that class.

You can use one of the following commands in a class map:

•match access-group

•match input-interface

•match mpls experimental

•match protocol

If you specify more than one command in the class map, only the last command entered applies. The last command overrides the previously entered commands.

Examples

The following example specifies class101 as the name of a class, and it defines a class map for this class. The class101 class specifies policy for traffic that matches access control list 101.

class-map class101

  match access-group 101 

Related Commands

Command	Description
class (policy-map)	Specifies the name of the class whose policy you want to create or change, and the default class (commonly known as the class-default class) before you configure its policy.
match access-group	Configures the match criteria for a class map based on the specified ACL number.
match input-interface	Configures a class map to use the specified input interface as a match criterion.
match mpls experimental	Configures a class map to use the specified EXP field value as a match criterion.
match protocol	Configures the match criteria for a class map based on the specified protocol.
policy-map	Creates or modifies a policy map that can be attached to one or more interfaces to specify a service policy.

clear ip rsvp reservation

To remove Resource Reservation Protocol (RSVP) RESV-related receiver information currently in the database, use the clear ip rsvp reservation command in EXEC mode.

clear ip rsvp reservation {session-ip-address sender-ip-address {tcp | udp | ip-protocol} session-dport sender-sport | *}

Syntax Description

session-ip-address	For unicast sessions, this is the address of the intended receiver; for multicast sessions, it is the IP multicast address of the session.
sender-ip-address	The IP address of the sender.
tcp | udp | ip-protocol	TCP, User Datagram Protocol (UDP), or IP protocol in the range from 0 to 65535.
session-dport	The destination port. Note Port numbers are specified in all cases, because the use of 16-bit ports following the IP header is not limited to UDP or TCP. If destination is zero, source must be zero, and the implication is that ports are not checked. If destination is nonzero, source must be nonzero (except for wildcard filter (wf) reservations, for which the source port is always ignored and can therefore be zero).
sender-sport	The source port. Note Port numbers are specified in all cases, because the use of 16-bit ports following the IP header is not limited to UDP or TCP. If destination is zero, source must be zero, and the implication is that ports are not checked. If destination is nonzero, source must be nonzero (except for wildcard filter (wf) reservations, for which the source port is always ignored and can therefore be zero).
*	Wildcard used to clear all senders.

Command Modes

EXEC

Command History

Release	Modification
11.2	This command was introduced.

Usage Guidelines

Use the clear ip rsvp reservation command to remove the RESV-related sender information currently in the database so that when reservation requests arrive, based on the RSVP admission policy, the relevant ones can be reestablished.

Whenever you change the clockrate or bandwidth of an interface, RSVP does not update its database to reflect the change. This is because such a change requires that RSVP reestablish reservations based on the new clockrate or bandwidth value and arbitrarily dropping some reservations while retaining others is not desired. The solution is to clear the RESV state by issuing the clear ip rsvp reservation command.

The clear ip rsvp reservation command clears the RESV state from the router on which you issued the command and causes the router to send a PATH TEAR message to the upstream routers thereby clearing the RESV state for that reservation on all the upstream routers.

Examples

The following example clears all the RESV-related receiver information currently in the database:

Router# clear ip rsvp reservation *

The following example clears all the RESV-related receiver information for a specified reservation currently in the database:

Router# clear ip rsvp reservation 10.2.1.1 10.1.1.2 udp 10 20

Related Commands

Command	Description
clear ip rsvp sender	Removes RSVP PATH-related sender information currently in the database.

clear ip rsvp sender

To remove Resource Reservation Protocol (RSVP) PATH-related sender information currently in the database, use the clear ip rsvp sender command in EXEC mode.

clear ip rsvp sender {session-ip-address sender-ip-address {tcp | udp | ip-protocol} session-dport sender-sport | *}

Syntax Description

session-ip-address	For unicast sessions, this is the address of the intended receiver; for multicast sessions, it is the IP multicast address of the session.
sender-ip-address	The IP address of the sender.
tcp | udp | ip-protocol	TCP, User Datagram Protocol (UDP), or IP protocol in the range from 0 to 65535.
session-dport	The destination port. Note Port numbers are specified in all cases, because the use of 16-bit ports following the IP header is not limited to UDP or TCP. If destination is zero, source must be zero, and the implication is that ports are not checked. If destination is nonzero, source must be nonzero (except for wildcard filter (wf) reservations, for which the source port is always ignored and can therefore be zero).
sender-sport	The source port. Note Port numbers are specified in all cases, because the use of 16-bit ports following the IP header is not limited to UDP or TCP. If destination is zero, source must be zero, and the implication is that ports are not checked. If destination is nonzero, source must be nonzero (except for wildcard filter (wf) reservations, for which the source port is always ignored and can therefore be zero).
*	Wildcard used to clear all senders.

Command Modes

EXEC

Command History

Release	Modification
11.2	This command was introduced.

Usage Guidelines

Use the clear ip rsvp sender command to remove the PATH-related sender information currently in the database so that when reservation requests arrive, based on the RSVP admission policy, the relevant ones can be reestablished.

Whenever you change the clockrate or bandwidth of an interface, RSVP does not update its database to reflect the change. This is because such a change requires that RSVP reestablish reservations based on the new clockrate or bandwidth value and arbitrarily dropping some reservations while retaining others is not desired. The solution is to clear the PATH state by issuing the clear ip rsvp sender command.

The clear ip rsvp sender command clears the PATH state from the router on which you issued the command and causes the router to send a PATH TEAR message to the downstream routers thereby clearing the PATH state for that reservation on all the downstream routers.

Examples

The following example clears all the PATH-related sender information currently in the database:

Router# clear ip rsvp sender *

The following example clears all the PATH-related sender information for a specified reservation currently in the database:

Router# clear ip rsvp sender 10.2.1.1 10.1.1.2 udp 10 20

Related Commands

Command	Description
clear ip rsvp reservation	Removes RSVP RESV-related receiver information currently in the database.

custom-queue-list

To assign a custom queue list to an interface, use the custom-queue-list interface configuration command. To remove a specific list or all list assignments, use the no form of this command.

custom-queue-list list

no custom-queue-list [list]

Syntax Description

list	Any number from 1 to 16 for the custom queue list.

Defaults

No custom queue list is assigned.

Command Modes

Interface configuration

Command History

Release	Modification
10.0	This command was introduced.

Usage Guidelines

Only one queue list can be assigned per interface. Use this command in place of the priority-list command (not in addition to it). Custom queueing allows a fairness not provided with priority queueing. With custom queueing, you can control the bandwidth available on the interface when the interface is unable to accommodate the aggregate traffic enqueued. Associated with each output queue is a configurable byte count, which specifies how many bytes of data should be delivered from the current queue by the system before the system moves on to the next queue. When a particular queue is being processed, packets are sent until the number of bytes sent exceeds the queue byte count or until the queue is empty.

Use the show queueing custom and show interfaces commands to display the current status of the custom output queues.

Examples

In the following example, custom queue list number 3 is assigned to serial interface 0:

interface serial 0

 custom-queue-list 3

Related Commands

Command	Description
queue-list default	Assigns a priority queue for those packets that do not match any other rule in the queue list.
queue-list interface	Establishes queueing priorities on packets entering on an interface.
queue-list queue byte-count	Specifies how many bytes the system allows to be delivered from a given queue during a particular cycle.
queue-list queue limit	Designates the queue length limit for a queue.
show interfaces	Displays statistics for all interfaces configured on the router or access server.
show queue	Displays the contents of packets inside a queue for a particular interface or VC.
show queueing	Lists all or selected configured queueing strategies.

exponential-weighting-constant

To configure the exponential weight factor for the average queue size calculation for a Weighted Random Early Detection (WRED) parameter group, use the exponential-weighting-constant random-detect-group configuration command. To return the exponential weight factor for the group to the default, use the no form of this command.

exponential-weighting-constant exponent

no exponential-weighting-constant

Syntax Description

exponent

Exponent from 1 to 16 used in the average queue size calculation. The default is 9.

Defaults

The weight factor is 9.

Command Modes

Random-detect-group configuration

Command History

Release	Modification
11.1(22)CC	This command was introduced.

Usage Guidelines

When used, this command is issued after the random-detect-group command is entered.

Use this command to change the exponent used in the average queue size calculation for a WRED parameter group. The average queue size is based on the previous average and the current size of the queue. The formula is:

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

where x is the exponential weight factor specified in this command. Thus, the higher the factor, the more dependent the average is on the previous average.

---

Note The default WRED parameter values are based on the best available data. We recommend that you do not change the parameters from their default values unless you have determined that your applications would benefit from the changed values.

---

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. The WRED process will be 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 resulting slow-moving average will accommodate temporary bursts in traffic.

If the value of x gets too high, WRED will not react to congestion. Packets will be 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 will respond quickly to long queues. Once the queue falls below the minimum threshold, the process will stop dropping packets.

If the value of x gets too low, WRED will overreact to temporary traffic bursts and drop traffic unnecessarily.

Examples

The following example configures the WRED group sanjose with a weight factor of 10:

random-detect-group sanjose

  exponential-weighting-constant 10

Related Commands

Command	Description
protect	Configures a VC to belong to the protected group of the bundle or to be an individually protected VC bundle member.
random-detect exponential-weighting-constant	Configures the WRED and DWRED exponential weight factor for the average queue size calculation.
random-detect-group	Defines the WRED or DWRED parameter group.
show queueing	Lists all or selected configured queueing strategies.
show queueing interface	Displays the queueing statistics of an interface or VC.

fair-queue (class-default)

To specify the number of dynamic queues to be reserved for use by the class-default class as part of the default class policy, use the fair-queue policy-map class configuration command. To delete the configured number of dynamic queues from the class-default policy, use the no form of this command.

fair-queue [number-of-dynamic-queues]

no fair-queue [number-of-dynamic-queues]

Syntax Description

number-of-dynamic-queues

(Optional) A power of 2 number in the range of 16 to 4096 specifying the number of dynamic queues.

Defaults

The number of dynamic queues is derived from the interface or ATM permanent virtual circuit (PVC) bandwidth. See Table 3 for the default number of dynamic queues that weighted fair queueing (WFQ) and class-based WFQ (CBWFQ) use when they are enabled on an interface. See Table 4 for the default number of dynamic queues used when WFQ or CBWFQ is enabled on an ATM PVC.

Table 3 Default Number of Dynamic Queues as a Function of Interface Bandwidth

Bandwidth Range	Number of Dynamic Queues
Less than or equal to 64 kbps	16
More than 64 kbps and less than or equal to 128 kbps	32
More than 128 kbps and less than or equal to 256 kbps	64
More than 256 kbps and less than or equal to 512 kbps	128
More than 512 kbps	256

Table 4 Default Number of Dynamic Queues as a Function of ATM PVC Bandwidth

Bandwidth Range	Number of Dynamic Queues
Less than or equal to 128 kbps	16
More than 128 kbps and less than or equal to 512 kbps	32
More than 512 kbps and less than or equal to 2000 kbps	64
More than 2000 kbps and less than or equal to 8000 kbps	128
More than 8000 kbps	256

Command Modes

Policy-map class configuration

Command History

Release	Modification
12.0(5)T	This command was introduce