cc/td/doc/product/software/ios120/120newft/120limit/120st
hometocprevnextglossaryfeedbacksearchhelp
PDF

Table of Contents

MPLS Traffic Engineering (TE)—Scalability Enhancements
Feature Overview
Supported Platforms
Supported Standards, MIBs, and RFCs
Prerequisites
Configuration Tasks
Monitoring and Maintaining Scalability Enhancements
Configuration Examples
Command Reference
clear ip rsvp counters
clear ip rsvp msg-pacing
clear mpls traffic-eng tunnel counters
ip rsvp msg-pacing
mpls traffic-eng scanner
mpls traffic-eng topology holddown sigerr
show ip rsvp counters
show mpls traffic-eng tunnels statistics
Glossary

MPLS Traffic Engineering (TE)—Scalability Enhancements


This feature module describes the MPLS traffic engineering scalability enhancements for Release 12.0(14)ST. Scalability indicates how quickly some measure of resource usage increases as a network gets larger.

This document contains the following sections:

Feature Overview

Implementation of multiprotocol label switching (MPLS) traffic engineering scalability has been improved so that scalability performs better for large numbers of traffic engineering tunnels. These improvements

User-observable scalability enhancements include the following:

Pacing for RSVP Messages

A burst of Resource Reservation Protocol (RSVP) traffic engineering signaling messages can overflow the input queue of a receiving router, causing some messages to be dropped. Dropped messages cause a substantial delay in completing label-switched path (LSP) signaling.

A new mechanism controls the transmission rate for RSVP messages and reduces the likelihood of input drops on the receiving router. The default transmission rate is 200 RSVP messages per second to a given neighbor. The rate is configurable.

Signaling and Management for MPLS Traffic Engineering Tunnels

The following changes improve the responsiveness of LSP recovery when a link used by an LSP fails:

Controlling IS-IS and MPLS Traffic Engineering Topology Database Interactions

The delay between when the Intermediate System-to-Intermediate System (IS-IS) protocol receives an IGP update and when it delivers the update to the MPLS traffic engineering topology database has been reduced in most situations.

Previously, when IS-IS received a new LSP that contained traffic engineering type, length, and value (TLV) objects, a delay of several seconds could occur before IS-IS passed the traffic engineering TLVs to the traffic engineering database. The purpose of the delay was to provide better scalability during periods of network instability and to give the router an opportunity to receive more fragments of the LSP before passing the information to the traffic engineering database. However, this delay increased the convergence time for the traffic engineering database.

Now IS-IS extracts traffic engineering TLVs from received LSPs and passes them to the traffic engineering database immediately. The exception to this occurs when there are large numbers of LSPs to process and it is important to limit CPU consumption, such as during periods of network instability. The arguments that control IS-IS delivery of traffic engineering TLVs to the traffic engineering topology database are configurable.

Improved Diagnostic Capabilities for MPLS Traffic Engineering and RSVP Signaling

The following enhancements improve diagnostic and trouble shooting capabilities for MPLS traffic engineering and RSVP:

Benefits

The scalability enhancements provide the following benefits:

Restrictions

The number of tunnels that a particular platform can support may vary depending on

Related Features and Technologies

The MPLS traffic engineering feature is related to the IS-IS, OSPF, RSVP, and MPLS features. These features are presented in Cisco product documentation (see the "Related Documents" section).

You can use the scalability enhancements with

Related Documents

Supported Platforms

The scalability enhancements are supported on the following platforms:

Supported Standards, MIBs, and RFCs

Standards

This feature does not support any new or modified standards.

MIBs

This feature supports the Traffic Engineering MIB.

To obtain lists of MIBs supported by platform and Cisco IOS release and to download MIB modules, go to the Cisco MIB web site on Cisco Connection Online (CCO) at http://www.cisco.com/public/sw-center/netmgmt/cmtk/mibs.shtml.

RFCs

This feature does not support any new or modified RFCs.

Prerequisites

Your network must support the following Cisco IOS features before you enable MPLS traffic engineering:

Configuration Tasks

To configure traffic engineering, refer to the MPLS Traffic Engineering and Enhancements manual.

To enable the scalability enhancements, perform the following steps:

1. Turn on RSVP message pacing.

2. Verify that RSVP message pacing is turned on.

The configuration tasks are described below.

Turning on RSVP Message Pacing

RSVP message pacing maintains, on an outgoing interface basis, a count of the messages it has been forced to drop because the output queue for the interface used for message pacing was full.

To turn on RSVP message pacing, enter the following global configuration command:

ip rsvp msg-pacing [period msec [burst msgs [max_size qsize]]]

Verifying That RSVP Message Pacing Is Turned On

To verify that RSVP message pacing is turned on, enter the following command:

show ip rsvp neighbor

Monitoring and Maintaining Scalability Enhancements

To monitor and maintain the scalability enhancements, use the commands shown below. For detailed command descriptions, go to the "Command Reference" section.

Command Purpose
router# show ip rsvp neighbor

Verifies that RSVP message pacing is turned on.

router# show ip rsvp counters

Displays the counts of RSVP messages that were sent and received.

router# clear ip rsvp counters

Clears (sets to zero) all Internet Protocol (IP) RSVP counters that are being maintained.

router# clear ip rsvp msg-pacing

Clears (sets to zero) counts of the messages that message pacing was forced to drop because the output queue for the interface used for message pacing was full.

router# show mpls traffic-eng tunnels statistics

Displays event counters for one or more MPLS traffic engineering tunnels.

router# clear mpls traffic-eng tunnel counters

Clears the counters for all MPLS traffic engineering tunnels.

Configuration Examples

The following command turns on RSVP message pacing:

ip rsvp msg-pacing

Enter the following command to verify that RSVP message pacing is turned on:

show ip rsvp neighbor

Following is sample output that traffic engineering displays if RSVP message pacing is turned on:

RSVP Msg-Pacing:Period 20, Burst 4, Rate 200, MaxSize 500
Interface Neighbor Encapsulation OQueue OQueue-drops
AT0.1 10.12.9.12 RSVP 0 0
Hs0 10.13.9.13 RSVP 0 0
tagsw4500-9#

If RSVP message pacing is not turned on, there is not an OQueue or OQueue-drops column.

Command Reference

This section documents the following new or modified commands:

All other commands used with this feature are documented in the Cisco IOS Release 12.0 and Release 12.1 command reference publications.

clear ip rsvp counters

To clear (set to zero) all IP RSVP counters that are being maintained, use the clear ip rsvp counters EXEC command.

clear ip rsvp counters [confirm]

Syntax Description

confirm

(Optional) Requests a confirmation that all IP RSVP counters were cleared.

Defaults

No default behavior or values.

Command Modes

EXEC

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

This command allows you to set all IP RSVP counters to zero so that you can see changes easily.

Examples

In the following example, all IP RSVP counters that are being maintained are cleared:

router# clear ip rsvp counters
Clear rsvp counters [confirm]
router#

Related Commands

Command Description

show ip rsvp counters

Displays counts of RSVP messages that were sent and received.

clear ip rsvp msg-pacing

To clear the RSVP message pacing output from the show ip rsvp neighbor command, use the clear ip rsvp msg-pacing EXEC command.

clear ip rsvp msg-pacing

Syntax Description

This command has no arguments or keywords.

Defaults

No default behavior or values.

Command Modes

EXEC

Command History

Release Modification

12.0(14)ST

This command was introduced.

Examples

The following example clears the RSVP message pacing output:

router# clear ip rsvp msg-pacing

Related Commands

Command Description

show ip rsvp counters

Displays counts of RSVP messages that were sent and received.

show ip rsvp neighbor

Shows the current RSVP neighbors and indicates whether the neighbor is using IP or User Data Protocol (UDP) encapsulation for a specified interface or for all interfaces.

clear mpls traffic-eng tunnel counters

To clear the counters for all MPLS traffic engineering tunnels, use the clear mpls traffic-eng tunnel counters EXEC command.

clear mpls traffic-eng tunnel counters

Syntax Description

This command has no optional parameters or keywords.

Defaults

No default behavior or values.

Command Modes

EXEC

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

This command allows you to set the MPLS traffic engineering tunnel counters to zero so that you can see changes to the counters easily.

Examples

In the following example, the counters for all MPLS traffic engineering tunnels are cleared and a request is made for confirmation that the specified action occurred:

router# clear mpls traffic-eng tunnel counters
Clear traffic engineering tunnel counters [confirm]
router#

Related Commands

Command Description

show mpls traffic-eng tunnels statistics

Displays event counters for one or more MPLS traffic engineering tunnels.

ip rsvp msg-pacing

To set up message pacing (that is, to control the transmission rate for RSVP messages), use the ip rsvp msg-pacing configuration command. To disable this feature, use the no form of this command.

ip rsvp msg-pacing [period msec [burst msgs [max_size qsize]]]
no rsvp msg-pacing

Syntax Description

period msec

(Optional) Length of the interval, in seconds, during which a router can send the number of RSVP messages specified in the burst argument. The value can be from 1 to 1000 milliseconds.

burst msgs

(Optional) Maximum number of RSVP messages that a router can send to an output interface during each interval specified in the period argument. The value can be from 1 to 2000.

max_size qsize

(Optional) Size of per-interface output queues in the sending router. Valid values are from 1 to 2000.

Defaults

RSVP messages are not paced.

If you enter the command without the optional arguments, the transmission rate for RSVP messages is limited to 200 messages per second per outgoing interface.

The default output queue size, specified in the max_size keyword, is 500.

Command Modes

Configuration

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

You can use this command to prevent a burst of RSVP traffic engineering signaling messages from overflowing the input queue of a receiving router, which would cause the router to drop some messages. Dropped messages substantially delay the completion of signaling for LSPs for which messages have been dropped.

Examples

In the following example, a router can send a maximum of 150 RSVP traffic engineering signaling messages in 1 second to a neighbor, and the size of the output queue is 750:

router(config)# ip rsvp msg-pacing period 1 burst 150 maxsize 750

Related Commands

Command Description

clear ip rsvp msg-pacing

Clears the RSVP message pacing output from the show ip rsvp neighbor command.

mpls traffic-eng scanner

To specify how often IS-IS should extract traffic engineering TLVs from flagged LSPs and pass them to the traffic engineering topology database, and the maximum number of LSPs that the router can process immediately, use the mpls traffic-eng scanner router IS-IS configuration subcommand. To disable this feature, use the no form of this command.

mpls traffic-eng scanner [interval sec] [max-flash #_LSPs]
no mpls traffic-eng scanner

Syntax Description

interval sec

(Optional) Frequency, in seconds, at which IS-IS transmits traffic engineering TLVs into the traffic engineering database. The value can be from 1 to 60.

max-flash #_LSPs

(Optional) Maximum number of LSPs that the router can process immediately without incurring a delay. The value can be from 0 to 200.

Defaults

The default interval is 5 seconds.

The default max-flash value is 15 LSPs.

The first 15 LSPs are sent without a delay into the traffic engineering database. If more LSPs are received, the default delay of 5 seconds applies.

If you specify the no form of this command, there is a delay of 5 seconds before IS-IS scans its database and passes traffic engineering TLVs associated with flagged LSPs to the traffic engineering database.

Command Modes

Router IS-IS configuration subcommand

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

When IS-IS receives a new LSP, it inserts it into the IS-IS database. If the LSP contains traffic engineering TLVs, IS-IS flags the LSPs for transmission to the traffic engineering database. At the default or user-specified interval, traffic engineering TLVs are extracted and sent to the traffic engineering database. Users can also specify the maximum number of LSPs that the router can process immediately.

Examples

In the following example, the router is allowed to process up to 50 IS-IS LSPs without any delay. Processing entails checking for traffic engineering TLVs, extracting them, and passing them to the traffic engineering database. If more than 50 LSPs need to be processed, there is a delay of 5 seconds for subsequent LSPs.

router(isis)# mpls traffic-eng scanner interval 5 max-flash 50

Related Commands

Command Description

mpls traffic-eng

Configures a router running IS-IS so that it floods MPLS traffic engineering link information into the indicated IS-IS level.

mpls traffic-eng router-id

Specifies that the traffic engineering router identifier for the node is the IP address associated with a given interface.

mpls traffic-eng topology holddown sigerr

To specify the amount of time that a router should ignore a link in its traffic engineering topology database in tunnel path Constrained Shortest Path First (CSPF) computations following a traffic engineering tunnel error on the link, use the mpls traffic-eng topology holddown sigerr global configuration command. To disable this feature, use the no form of this command.

mpls traffic-eng topology holddown sigerr seconds
no mpls traffic-eng topology holddown sigerr

Syntax Description

seconds

Specifies how long (in seconds) a router should ignore a link during tunnel path calculations following a traffic engineering tunnel error on the link. The value can be from 0 to 300.

Defaults

If you do not specify this command, tunnel path calculations ignore a link on which there is a traffic engineering error until either 10 seconds have elapsed or a topology update is received from the IGP routing protocol.

Command Modes

Global configuration

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

A router that is at the headend for traffic engineering tunnels might receive an RSVP No Route error message for an existing tunnel or for one being signaled due to the failure of a link the tunnel traverses before the router receives a topology update from the IGP routing protocol announcing that the link is down. In such a case, the headend router ignores the link in subsequent tunnel path calculations to avoid generating paths that include the link and are likely to fail when signaled. The link is ignored until the router receives a topology update from its IGP or a link holddown timeout occurs. You can use the mpls traffic-eng topology holddown sigerr command to change the link holddown time from its 10 second default value.

Examples

In the following example, the link holddown time for signaling errors is set at 15 seconds:

router(config)# mpls traffic-eng topology holddown sigerr 15

Related Commands

Command Description

show mpls traffic-eng topology

Show the MPLS traffic engineering global topology as currently known at this node.

show ip rsvp counters

To display the counts of RSVP messages that were sent and received, use the show ip rsvp counters EXEC command.

show ip rsvp counters [interface interface_unit | summary]

Syntax Description

interface interface_unit

 

(Optional) Shows the number of RSVP messages sent and received for the specified interface name.

summary

(Optional) Shows the cumulative number of RSVP messages sent and received by the platform.

Defaults

If you enter the command without a keyword, the command displays the number of RSVP messages that were sent and received for each interface for which RSVP is configured.

Command Modes

EXEC

Command History

Release Modification

12.0(14)ST

This command was introduced.

Examples

In the following example, values are shown for the number of RSVP messages of each type (such as Path and Resv) that were sent and received on POS2/2:

router# sh ip rsvp counters interface pos2/2
POS2/2 Recv Xmit Recv Xmit
Path 0 112876 Resv 74414 0
PathError 0 0 ResvError 0 0
PathTear 0 7 ResvTear 4 0
ResvConfirm 0 0 ResvTearConfirm 0 4
UnknownMsg 0 0 Errors 0 0
router#

In the following example, values are shown for the number of RSVP messages of each type that were sent and received by the router over all interfaces:

router# show ip rsvp counters summary
All Interfaces Recv Xmit Recv Xmit
Path 512 224962 Resv 148446 385
PathError 0 2 ResvError 0 0
PathTear 8 12 ResvTear 4 4
ResvConfirm 0 0 ResvTearConfirm 4 4
UnknownMsg 0 0 Errors 0 0
router#

Table 1 describes the fields displayed in the above examples.

Table 1   show ip rsvp counters Field Descriptions

Field Description

Recv

Number of messages received at the specified interface or at all interfaces

Xmit

Number of messages transmitted from the specified interface or from all interfaces

All Interfaces

Specifies that the number of messages displayed is for all interfaces

Related Commands

Command Description

clear ip rsvp counters

Clears (sets to zero) all IP RSVP counters that are being maintained.

show mpls traffic-eng tunnels statistics

To display event counters for one or more MPLS traffic engineering tunnels, use the show mpls traffic-eng tunnels statistics EXEC command.

show mpls traffic-eng tunnels [tunnel tunnel_name] statistics [summary]

Syntax Description

tunnel tunnel_name

(Optional) Displays event counters for the specified tunnel.

summary

(Optional) Displays event counters accumulated for all tunnels.

Defaults

If you enter the command without any keywords, the command displays the event counters for every MPLS traffic engineering tunnel interface configured on the router.

Command Modes

EXEC

Command History

Release Modification

12.0(14)ST

This command was introduced.

Usage Guidelines

A label switching router (LSR) maintains counters for each MPLS traffic engineering tunnel headend that counts significant events for the tunnel, such as state transitions for the tunnel, changes to the tunnel path, and various signaling failures. You can use the show mpls traffic-eng tunnels statistics command to display these counters for a single tunnel, for every tunnel, or for all tunnels (accumulated values). Displaying the counters is often useful for troubleshooting tunnel problems.

Examples

The following are examples of output from the show mpls traffic-eng tunnels statistics command:

router# show mpls traffic-eng tunnels tunnel1001 statistics
Tunnel1001 (Destination 8.8.8.8; Name Router_t1001)
Management statistics:
Path: 25 no path, 1 path no longer valid, 0 missing ip exp path
5 path changes
State: 3 transitions, 0 admin down, 1 oper down
Signalling statistics:
Opens: 2 succeeded, 0 timed out, 0 bad path spec
0 other aborts
Errors: 0 no b/w, 0 no route, 0 admin
0 bad exp route, 0 rec route loop, 0 other

router# show mpls traffic-eng tunnels statistics
Tunnel1001 (Destination 8.8.8.8; Name Router_t1001)
Management statistics:
Path: 25 no path, 1 path no longer valid, 0 missing ip exp path
5 path changes
State: 3 transitions, 0 admin down, 1 oper down
Signalling statistics:
Opens: 2 succeeded, 0 timed out, 0 bad path spec
0 other aborts
Errors: 0 no b/w, 0 no route, 0 admin
0 bad exp route, 0 rec route loop, 0 other

...

Tunnel7050 (Destination 8.8.8.8; Name Router_t7050)
Management statistics:
Path: 19 no path, 1 path no longer valid, 0 missing ip exp path
3 path changes
State: 3 transitions, 0 admin down, 1 oper down
Signalling statistics:
Opens: 2 succeeded, 0 timed out, 0 bad path spec
0 other aborts
Errors: 0 no b/w, 0 no route, 0 admin
0 bad exp route, 0 rec route loop, 0 other
router#
router# show mpls traffic-eng tunnels statistics summary
Management statistics:
Path: 2304 no path, 73 path no longer valid, 0 missing ip exp path
432 path changes
State: 300 transitions, 0 admin down, 100 oper down
Signalling statistics:
Opens: 200 succeeded, 0 timed out, 0 bad path spec
0 other aborts
Errors: 0 no b/w, 18 no route, 0 admin
0 bad exp route, 0 rec route loop, 0 other
router#

Table 2 describes the fields displayed in the above examples.

Table 2   show mpls traffic-eng tunnels statistics Field Descriptions

Field Description

Tunnelnnnn

Name of the tunnel interface.

Destination

IP address of the tunnel tailend.

Name

Internal name for the tunnel, composed of the router name and the tunnel interface number.

Path

Heading for counters for tunnel path events, including:

  • No path—Number of unsuccessful attempts to calculate a path for the tunnel.
  • Path no longer valid—Number of times a previously valid path for the tunnel became invalid.
  • Missing ip exp path—Number of times that attempts to use "obtain a path for the tunnel" failed because no path was configured (and there was no dynamic path option for the tunnel).
  • Path changes—Number of times the tunnel path changed.

State

Heading for counters for tunnel state transitions.

Opens

Heading for counters for tunnel open attempt events.

Errors

Heading for various tunnel signaling errors, such as no bandwidth, no route, admin (preemption), a bad explicit route, a loop in the explicit route, and so forth.

Related Commands

Command Description

clear mpls traffic-eng tunnel counters

Clears the counters for all MPLS traffic engineering tunnels.

Glossary

CEF—Cisco express forwarding. A means for accelerating the forwarding of packets within a router, by storing route lookup information in several data structures instead of in a route cache.

enterprise network—A large and diverse network connecting most major points in a company or other organization.

headend—The end point of a broadband network. All stations transmit toward the headend; the headend then transmits toward the destination stations.

IGP—Interior Gateway Protocol. An Internet protocol used to exchange routing information within an autonomous system. Examples of common Internet IGPs include IGRP, OSPF, and RIP.

interface—A network connection.

IS-IS—Intermediate System-to-Intermediate System. OSI link-state hierarchical routing protocol based on DECnet Phase V routing, where ISs (routers) exchange routing information based on a single metric, to determine the network topology.

label-switched path (LSP)—A sequence of hops (R0...Rn) in which a packet travels from R0 to Rn through label switching mechanisms. A label-switched path can be chosen dynamically, based on normal routing mechanisms, or through configuration.

MPLS—Multiprotocol Label Switching (formerly known as tag switching). A method for directing packets primarily through Layer 2 switching rather than Layer 3 routing. In MPLS, packets are assigned short fixed-length labels at the ingress to an MPLS cloud by using the concept of forwarding equivalence classes. Within the MPLS domain, the labels are used to make forwarding decisions mostly without recourse to the original packet headers.

router—A network layer device that uses one or more metrics to determine the optimal path along which network traffic should be forwarded. Routers forward packets from one network to another based on network layer information.

RSVP—Resource Reservation Protocol. A protocol that supports the reservation of resources across an IP network.

scalability—An indicator showing how quickly some measure of resource usage increases as a network gets larger.

topology—The physical arrangement of network nodes and media within an enterprise networking structure.

traffic engineering—Techniques and processes that cause routed traffic to travel through the network on a path other than the one that would have been chosen if standard routing methods were used.

traffic engineering tunnel—A label-switched tunnel that is used for traffic engineering. Such a tunnel is set up through means other than normal Layer 3 routing; it is used to direct traffic over a path different from the one that Layer 3 routing would cause the tunnel to take.


hometocprevnextglossaryfeedbacksearchhelp
Posted: Thu Jan 16 21:48:03 PST 2003
All contents are Copyright © 1992--2002 Cisco Systems, Inc. All rights reserved.
Important Notices and Privacy Statement.