AR150 - Configuration Guide - QoS(V200R002C00_02)

8/10/2019 AR150 - Configuration Guide - QoS(V200R002C00_02)

1/111

Huawei AR150&200 Series Enterprise Routers

V200R002C00

Configuration Guide - QoS

Issue 02

Date 2012-03-30

HUAWEI TECHNOLOGIES CO., LTD.


2/111

Copyright Huawei Technologies Co., Ltd. 2012. All rights reserved.

No part of this document may be reproduced or transmitted in any form or by any means without prior writtenconsent of Huawei Technologies Co., Ltd. Trademarks and Permissions

and other Huawei trademarks are trademarks of Huawei Technologies Co., Ltd. All other trademarks and trade names mentioned in this document are the property of their respective holders. Notice

The purchased products, services and features are stipulated by the contract made between Huawei and thecustomer. All or part of the products, services and features described in this document may not be within the

purchase scope or the usage scope. Unless otherwise specified in the contract, all statements, information,and recommendations in this document are provided "AS IS" without warranties, guarantees or representationsof any kind, either express or implied.

The information in this document is subject to change without notice. Every effort has been made in thepreparation of this document to ensure accuracy of the contents, but all statements, information, andrecommendations in this document do not constitute the warranty of any kind, express or implied.

Huawei Technologies Co., Ltd. Address: Huawei Industrial Base

Bantian, LonggangShenzhen 518129People's Republic of China

Website: http://www.huawei.com

Email: [email protected]

Issue 02 (2012-03-30) Huawei Proprietary and ConfidentialCopyright Huawei Technologies Co., Ltd.

i
http://www.huawei.com/


3/111

About This Document

Intended AudienceThis document describes the concepts and configuration procedures of QoS features on theAR150/200, and provides the configuration examples.

This document provides guidance for configuring QoS features.

This document is intended for:

l Data configuration engineersl Commissioning engineersl Network monitoring engineersl System maintenance engineers

Symbol ConventionsThe symbols that may be found in this document are defined as follows.

Symbol Description

DANGER

Indicates a hazard with a high level of risk, which if notavoided, will result in death or serious injury.

WARNINGIndicates a hazard with a medium or low level of risk, whichif not avoided, could result in minor or moderate injury.

CAUTION

Indicates a potentially hazardous situation, which if notavoided, could result in equipment damage, data loss,

performance degradation, or unexpected results.

TIP Indicates a tip that may help you solve a problem or savetime.

NOTE Provides additional information to emphasize or supplementimportant points of the main text.

Huawei AR150&200 Series Enterprise RoutersConfiguration Guide - QoS About This Document


ii


4/111

Command ConventionsThe command conventions that may be found in this document are defined as follows.

Convention DescriptionBoldface The keywords of a command line are in boldface .

Italic Command arguments are in italics .

[ ] Items (keywords or arguments) in brackets [ ] are optional.

{ x | y | ... } Optional items are grouped in braces and separated byvertical bars. One item is selected.

[ x | y | ... ] Optional items are grouped in brackets and separated byvertical bars. One item is selected or no item is selected.

{ x | y | ... } * Optional items are grouped in braces and separated byvertical bars. A minimum of one item or a maximum of allitems can be selected.

[ x | y | ... ] * Optional items are grouped in brackets and separated byvertical bars. Several items or no item can be selected.

& The parameter before the & sign can be repeated 1 to n times.

# A line starting with the # sign is comments.

Interface Numbering ConventionsInterface numbers used in this manual are examples. In device configuration, use the existinginterface numbers on devices.

Change History

Changes between document issues are cumulative. Therefore, the latest document versioncontains all updates made to previous versions.

Changes in Issue 02 (2012-03-30)

Based on issue 01 (2011-12-30), the document is updated as follows:

The following information is modified:l 1.5.5 Checking the Configuration

Changes in Issue 01 (2011-12-30)

Initial commercial release.

Huawei AR150&200 Series Enterprise RoutersConfiguration Guide - QoS About This Document


iii


5/111


6/111

1.8.2 Configuring a Traffic Policy....................................................................................................................36

1.8.2.1 Configuring a Sub Traffic Policy...................................................................................................36

1.8.2.2 Configuring a Traffic Policy...........................................................................................................36

1.8.2.3 Applying the Traffic Policy to an Interface....................................................................................38

1.8.3 (Optional) Configuring Traffic Policing on an Interface.........................................................................38

1.8.4 (Opt ional) Configuring Traffic Shaping on an Interface.........................................................................38

1.8.5 Chec king the Configuration.....................................................................................................................39

1.9 Maintaining Q oS...............................................................................................................................................39

1.9.1 Displayi ng the Queue-based Traffic Statistics........................................................................................40

1.9.2 Clearing the Queue-based Traffic Statistics............................................................................................40

1.10 Configuratio n Examples.................................................................................................................................40

1.10.1 Examp le for Configuring Priority Mapping..........................................................................................40

1.10.2 Examp le for Configuring Traffic Policing............................................................................................44

1.10.3 Examp le for Configuring Traffic Shaping.............................................................................................491.10.4 Examp le for Configuring Congestion Avoidance and Congestion Management.................................53

1.10.5 Examp le for Configuring HQoS............................................................................................................59

2 Traffic Policy Co nfiguration.....................................................................................................702.1 Traffic Policy Overview...................................................................................................................................71

2.2 Traffic Po licy Features Supported by the AR150/200.....................................................................................71

2.3 Configuring a Traffic Classifier.......................................................................................................................74

2.3.1 Establis hing the Configuration Task.......................................................................................................74

2.3.2 (Optional) C onfiguring SAC...................................................................................................................75

2.3.2.1 Config uring a Signature File..........................................................................................................752.3.2.2 Config uring an SAC Group............................................................................................................76

2.3.2.3 Config uring the SAC Statistics Function.......................................................................................77

2.3.2.4 Ch ecking the Configuration............................................................................................................78

2.3.3 Configu ring a Traffic Classifier..............................................................................................................78

2.3.4 Chec king the Configuration.....................................................................................................................80

2.4 Configuri ng a Traffic Behavior........................................................................................................................80

2.4.1 Esta blishing the Configuration Task.................................................................. .....................................81

2.4 .2 Configuring a Permit or Deny Action.....................................................................................................82

2.4.3 Configu ring Redirection..........................................................................................................................822.4.4 Configu ring Re-marking..........................................................................................................................83

2.4.5 Configu ring Traffic Policing...................................................................................................................84

2.4.6 Configu ring Traffic Shaping...................................................................................................................85

2.4.7 Configu ring Congestion Management.....................................................................................................85

2.4.8 Conf iguring Congestion Avoidance........................................................................................................87

2.4.9 Binding a Sub Traffic Policy...................................................................................................................88

2.4.10 Config uring Traffic Statistics................................................................................................................90

2.4.11 Ch ecking the Configuration...................................................................................................................90

2.5 Configuring a Traffic Policy.............................................................................................................................91

2.6 Maintaining a Traffic Policy.............................................................................................................................92

Huawei AR150&200 Series Enterprise RoutersConfiguration Guide - QoS Contents


v


7/111

2.6.1 Displaying the Flow-based Traffic Statistics...........................................................................................92

2.6.2 Clearing the Flow-based Traffic Statistics..............................................................................................92

2.7 Configuration Examples...................................................................................................................................93

2.7.1 Example for Configuring Re-marking.....................................................................................................93

2.7.2 Example for Configuring Traffic Statistics.............................................................................................97

2.7.3 Example for Preventing BT Download.................................................................................................100

Huawei AR150&200 Series Enterprise RoutersConfiguration Guide - QoS Contents


vi


8/111

1 QoS ConfigurationAbout This Chap ter

This chapter describes common QoS functions on interfaces of the AR150/200: prioritymapping, traf fic policing, traffic shaping, congestion management, congestion avoidance, andHierarchical Quality of Service (HQoS). It also provides configuration methods and examples.

1.1 QoS Ove rviewQoS is design ed to provide differentiated serv ices based on networking requirements.

1.2 QoS Feat ures Supported by the AR150/200This section describes QoS features supported by the AR150/200.

1.3 Configuri ng Priority MappingAfter priority mapping is configured, the AR150 /200 determines the queues and output prioritiesof the receive d packets based on packet priorities or the default 802.1p priority of the interface.By doing this , the AR150/200 provides differentiated service s.

1.4 Configuri ng Traffic PolicingThe AR150/2 00 supports interfa ce-based traffic policing and flow-based traffic policing.

1.5 Configuring Traffic ShapingTraffic shaping enables outgoing traffic to be sent out at an even rate and reduces the number of discarded packets whose rate exceeds the rate limit.

1.6 Configuring Congestion ManagementIf congestion occurs on a network after congestion management is configured, the AR150/200determines the sequence at which packets are forwarded according to the defined scheduling

policy.

1.7 Configuring Congestion AvoidanceAfter congestion avoidance is complete, the AR150/200 discards the packets whose rate exceedsthe rate limit based on the WRED configuration.

1.8 Configuring HQoSTo provide differentiated QoS services, configure HQoS on the AR150/200.

1.9 Maintaining QoSThis section describes how to maintain QoS.

Huawei AR150&200 Series Enterprise RoutersConfiguration Guide - QoS 1 QoS Configuration


1


9/111

1.10 Configuration ExamplesThis section provides several configuration examples of traffic policing, traffic shaping,congestion avoidance, and congestion management.



2


10/111

1.1 QoS OverviewQoS is designed to provide differentiated services based on networking requirements.

QoS evaluates the capabilities of the service supplied to meet customer requirements. On theInternet, QoS is used to evaluate capabilities of a network to transmit packets. The network

provides various services, and QoS evaluates services from different aspects. Generally, QoSevaluates core items, including the delay, jitter, and packet loss ratio during packet transmission.

QoS uses the following models:l Best-Effort (BE) model: The BE model is a service model that handles all packets equally.

In the BE model, a network attempts to send packets, but cannot ensure performance suchas delay and reliability. The BE model can be applied to various network applications, suchas FTP and email. The BE model can be implemented using the First in First out (FIFO)queue.

l IntServ model: The IntServ model is an integrated service model. A device must submit arequest to a network before sending packets. Combined with multicast, the IntServ modelcan be used in real-ti me multimedia applications that require high bandwidth and low delay,such as video conference and video on demand (VoD).

l DiffServ model: The DiffServ model is a multi-service model and can meet different QoSrequirements , which ensures that appl ications are provided with differentiated QoS levels.QoS can be s pecified based on various information, such as the IP precedence, sourceaddress, and destination addres s of packets. Network devices perform traffic classification,traffic shaping, traffic policing, and queue scheduling based on the information.

1.2 QoS Features Supported by the AR150/200This section describes QoS features supported by the AR150/200.

The AR150/200 supports the following functions:l Priority Mappingl Traffic Policingl Traffic Shapingl Congestion Avoidancel Congestion Managementl

HQoS

Table 1-1 Implementation of QoS features

Feature Implementation

Priority Mapping Priority mapping can be applied to inboundand outbound interfaces.



3


11/111

Feature Implementation

Traffic policing Interface-based traffic policing can beapplied to inbound and outbound WAN-sideinterfaces, but can only be applied to inboundLAN-side interfaces.

Flow-based traffic policing can be applied toinbound and outbound interfaces.

Traffic shaping Interface-based traffic shaping can only beapplied to outbound interfaces.

Queue-based traffic shaping can only beapplied to outbound interfaces.

Flow-based traffic shaping can only beapplied to outbound WAN-side interfaces.

Congestion avoidance Queue-based congestion avoidance can only be applied to outbound WAN-side interfaces.

Flow-based congestion avoidance can only be applied to outbound WAN-side interfaces.

Congestion management Queue-based congestion management canonly be applied to outbound interfaces.

Class-based congestion management canonly be applied to outbound WAN-sideinterfaces.

HQoS Flow-based hierarchical QoS (HQoS) canonly be applied to outbound WAN-sideinterfaces.

Priority Mapping Different packets carry different precedence fields. For example, VLAN packets carry the 802.1pfield, IP packets carry the DSCP field. The mappings between priority fields must be configuredon gateways to retain priorities of packets when the packets traverse different networks.

To ensure QoS for different packets, the AR150/200 determines the queues that received packetsenter based on 802.1p priorities or DSCP priorities in packets or the default 802.1p priority of an interface. The AR150/200 can re-mark packet priorities so that the connected device can

provide differentiated QoS based on precedence fields of the packets.

The AR150/200 sends packets to different interface queues based on the 802.1p priority, and performs traffic shaping, congestion avoidance, and queue scheduling for the queues. The layer 2 FE interfaces on the AR150/200 support only four queues, whereas other interfaces supporteight queues. The mappings between local priorities and queues on the Layer 2 FE interfaceson the AR150/200 are different. Table 1-2 lists the mappings between 802.1p priorities andqueues on the layer 2 FE interfaces on the AR150/200. Table 1-3 lists the mappings between802.1p priorities and queues on other interfaces.



4


12/111

Table 1-2 Mappings between 802.1p priorities and queues

802.1p Priority Queue Index

0 0

1 0

2 1

3 1

4 2

5 2

6 3

7 3

Table 1-3 Mappings between 802.1p priorities and queues

802.1p Priority Queue Index

0 0

1 1

2 2

3 34 4

5 5

6 6

7 7

Traffic Policing Traffic policing limits the traffic and resource usage by monitoring the rate limit.

Traffic policing discards the excess traffic to limit traffic within a proper range and to protectnetwork resources.

Traffic policing limits the traffic entering the Internet Service Provider (ISP).

l Token bucket and traffic measurement

The AR150/200 needs to determine whether the traffic rate exceeds the rate limit before performing traffic policing. Token buckets are usually used to measure traffic.

With a certain capacity, a token bucket stores tokens. The system places tokens into a token

bucket at the configured rate (one token occupies one bit). If the token bucket is full, excesstokens overflow and no token is added.



5


13/111

When measuring traffic, a token bucket forwards packets based on the number of tokensin the token bucket. If there are enough tokens in the token bucket for forwarding packets,the traffic rate is within the rate limit. Otherwise, the traffic rate is not within the rate limit.

Figure 1-1 Measuring traffic using token buckets

Packets sent by theinterface Continue to send

packets

Tokenbucket

Put tokens into the bucket at thespecified rate

Classification

Drop

Packets thatneed to be

policed

Packets that donot need to be

policed

The AR150/200 supports dual token bucket technology:

The dual token bucket technology uses the following parameters: Committed Burst Size (CBS): capacity of bucket C, that is, the maximum volume of

burst traffic (in bytes) allowed by bucket C each time. Committed Information Rate (CIR): rate at which tokens are put into bucket C, that is,

the average rate of traffic (in kbit/s) allowed by bucket C. Peak Burst Size (PBS): capacity of bucket P, that is, the maximum volume of burst

traffic (in bytes) allowed by bucket P each time. Peak Information Rate (PIR): rate at which tokens are put into bucket P, that is, the

average rate of traffic (in kbit/s) allowed by bucket P.

When dual token buckets are used, the packets whose rate is less than the CIR value arecolored green, the packets whose rate is greater than the PIR value are colored red, and the

packets whose rate is greater than the CIR value, but less than the PIR value are coloredyellow.

Traffic policing actions include permit, deny, and forward after a priority change. Bydefault, green packets and yellow packets are allowed to pass through, and red packets arediscarded.

l Traffic policing features supported by the AR150/200

The AR150/200 supports the following traffic policing features: Interface-based traffic policing limits the incoming or outgoing service traffic on an

interface.



6


14/111

NOTE

You can use interface-based traffic policing to:

l Limit all the service traffic on an interface.l Limit the service traffic matching a specified ACL rule.l Limit the service traffic whose source and destination IP addresses within a specified range.

Flow-based traffic policing limits incoming or outgoing service traffic matching trafficclassification rules.

Traffic Shaping

Traffic shaping also limits traffic and resource usage by monitoring the rate limit. It also usesthe token bucket technology to measure traffic.

Traffic shaping adjusts the speed of outgoing traffic so that the downstream device hascapabilities to process traffic.

l Differences between traffic shaping and traffic policing

Traffic policing directly discards the excess packets. Traffic shaping buffers the packetswhose rate is greater than the traffic shaping rate in queues; therefore, traffic shapingreduces the number of discarded packets. As shown in Figure 1-2 , when there are sufficienttokens in the token bucket, the buffered packets are forwarded at an even rate. If the tokensare insufficient, packets are buffered continuously. If the number of packets to be bufferedis greater than the queue length, excess packets are discarded.

Figure 1-2 Traffic shaping

Packets sent by the

interfaceFlows

Tokenbucket

Put tokens intotoken bucket atspecified rate

Simple trafficclassification

Packetswithin therate limit

Packets notwithin the rate

limit

Token

Buffer packets in

queues

Discardedpackets when

the token bucketis full

Packets that donot need to be

shaped

Packets thatneed to be

shaped



7


15/111

Traffic shaping increases the delay because it buffers packets in queues, whereas traffic policing does not.

l Traffic shaping features supported by the AR150/200

The AR150/200 supports the following traffic shaping features: Traffic shaping on an interface or a sub-interface

Shapes all the packets that pass through an interface or a sub-interface. Queue-based traffic shaping

Shapes packets in a specified queue on an interface so that packets are shaped based on priorities.

Flow-based traffic shaping

Shapes packets of a specified type that pass through an interface so that packetsmatching traffic classification rules can be shaped.

Congestion AvoidanceCongestion avoidance is a flow control mechanism. A system configured with congestionavoidance monitors network resources such as queues and memory buffers. When congestionoccurs or aggravates, the system discards packets.

Congestion avoidance policies include tail drop, Random Early Detection (RED), and WeightedRandom Early Detection (WRED):l Tail drop

The traditional packet loss policy uses the tail drop method. This method processes all packets equally without classifying the packets into different types. When congestion

occurs, packets at the end of a queue are discarded until the congestion problem is solved.This policy leads to global TCP synchronization. When packets of multiple TCPconnections are discarded simultaneously, these TCP connections enter the congestionavoidance and slow start state. After a while, the peak of these TCP connections occurs.The volume of traffic varies greatly, affecting link usage.

l RED

The RED technique randomly discards packets to prevent the transmission speed of multiple TCP connections from being reduced simultaneously.

As specified by the RED algorithm, the upper drop threshold and lower drop threshold areset. RED processes packets as follows:

When the queue length is shorter than the lower drop threshold, no packet is discarded. When the queue length is longer than the upper drop threshold, all packets are discarded. When the queue length is between the lower drop threshold and the upper drop threshold,

incoming packets are discarded randomly. RED generates a random number for eachincoming packet and compares it with the drop probability of the current queue. If therandom number is greater than the drop probability, the packet is discarded. A longer queue indicates a higher drop probability.

l WRED

WRED also discards packets randomly to prevent global TCP synchronization. WRED,however, generates random numbers of packets based on packet priorities. WRED discards

packets based on packet priorities, so the drop probability of packets with higher prioritiesis low.



8


16/111

By default, the AR150/200 uses tail drop. The AR150/200 supports queue-based WRED andflow-based RED.

Congestion Management

If a network transmitting both delay-sensitive and delay-insensitive services is congestedintermittently, congestion management is required. However, if a network is always congested,

bandwidth needs to be increased.

Generally, the AR150/200 uses the following queue scheduling mechanisms:l PQ schedulingl WRR schedulingl DRR schedulingl WFQ schedulingl PQ+WRR/PQ+DRR/PQ+WFQ schedulingl CBQ scheduling

l PQ scheduling

Priority queuing (PQ) schedules packets in descending order of priorities. Queues withlower priories are processed only after all the queues with higher priorities have been

processed.

By using PQ scheduling, the AR150/200 puts packets of delay-sensitive services intoqueues with higher priorities and packets of other services into queues with lower priorities.In this manner, packets of key services can be transmitted first.

PQ scheduling has a disadvantage. If a lot of packets exist in queues with higher prioritieswhen congestion occurs, packets in queues with lower priorities cannot be transmitted for a long time.

l WRR scheduling

Weighted Round Robin (WRR) scheduling ensures that packets in all the queues arescheduled in turn.

For example, eight queues are configured on an interface. Each queue is configured witha weight: w7, w6, w5, w4, w3, w2, w1, and w0. The weight value represents the percentageof obtaining resources. The following scenario assumes that the weights of queues on the100M interface are 50, 50, 30, 30, 10, 10, 10, and 10, which match w7, w6, w5, w4, w3,w2, w1, and w0. Therefore, the queue with the lowest priority can obtain at least 5 Mbit/s

bandwidth. This ensures that packets in all the queues can be scheduled.

In addition, WRR can dynamically change the time of scheduling packets in queues. For example, if a queue is empty, WRR ignores this queue and starts to schedule the next queue.This ensures efficient use of ba ndwidth.

WRR schedu ling has two disadvantages: WRR sch edules packets bas ed on the number of packets. When the average packet

length in each queue is the same or known, you can obtain the required bandwidth bysetting W RR weight values. When the average packet length in each queue is variable,you canno t obtain the requir ed bandwidth by setting WRR weight values.

Delay-sen sitive services, such as voice services, cannot be scheduled in a timelymanner.

l DRR scheduling

Implementation of Deficit Round Robin (DRR) is similar to that of WRR.



9


17/111

The difference between DRR and WRR is as follows: WRR schedules packets based onthe number of packets, whereas DRR schedules packets based on the packet length. If the

packet length is too long, DRR allows the negative weight value so that long packets can be scheduled. In the next round, the queue with the negative weight value is not scheduleduntil its weight value becomes positive.

DRR offsets the disadvantages of PQ scheduling and WRR scheduling. That is, in PQscheduling, packets in queues with lower priorities cannot be scheduled for a long time; inWRR scheduling, bandwidth is allocated improperly when the packet length of each queueis different or variable.

DRR cannot schedule delay-sensitive services such as voice services in time.l WFQ scheduling

Fair Queue (FQ) ensures that network resources are allocated evenly to optimize the delayand jitter of all flows. Weighted FQ (WFQ) schedules packets based on priorities, andschedules more packets with higher priorities than packets with lower priorities.

WFQ can automatically classify flows based on the session information, including the protocol type, source and destination TCP or UDP port numbers, source and destination IPaddresses, and precedence field in the ToS field. In addition, WFQ provides a large number of queues and evenly puts flows into queues to smooth out the delay. When flows leavequeues, WFQ allocates the bandwidth on the outbound interface for each flow based onthe precedence of each flow. Flows with the lowest priorities obtain the least bandwidth.

l PQ+WRR/PQ+DRR/PQ+WFQ scheduling

PQ, WRR, DRR, and WFQ have their own advantages and disadvantages. If only PQscheduling is used, packets in queues with lower priorities cannot obtain bandwidth. If onlyWRR, DRR, or WFQ scheduling is used, delay-sensitive services cannot be scheduled intime. PQ+WRR, PQ+DRR, or PQ+WFQ scheduling integrates the advantages of PQscheduling and WRR or DWRR scheduling and offsets their disadvantages.

By using PQ+WRR, PQ+DRR, or PQ+WFQ scheduling, the AR150/200 puts important packets, such as protocol packets and packets of delay-sensitive services to the PQ queue,and allocates bandwidth to the PQ queue. Then the AR150/200 can put other packets intoWRR, DRR, or WFQ queues based on the packet priority. Packets in WRR, DRR, or WFQqueues can be scheduled in turn.

l CBQ scheduling

Class-based queueing (CBQ) is an extension of WFQ and matches packets with trafficclassifiers. CBQ classifies packets based on the IP precedence or DSCP priority, inboundinterface, or 5-tuple (protocol type, source IP address and mask, destination IP address andmask, source port range, and destination port range). Then CBQ puts packets into differentqueues. If packets do not match any configured traffic classifiers, CBQ matches packets

with the default traffic classifier.CBQ provides the following types of queues: Expedited Forwarding (EF) queues are applied to short-delay services.

An EF queue has the highest priority. You can put one or more types of packets into EFqueues and set different bandwidth for different types of packets. During scheduling, if EF queues have packets, packets in EF queues are sent first. Packets in other queues aresent only after EF queues are empty or the maximum reservable bandwidth of EF queuesis exceeded.

UDP packets of VoIP services often exist in EF queues; therefore, use the tail dropmethod but not WRED.

Assured Forwarding (AF) queues are applied to key data services that require assured bandwidth.



10


18/111

Each AF queue corresponds to one type of packets. You can set bandwidth for eachtype of packets. During scheduling, the system sends packets based on the configured

bandwidth. AF implements fair scheduling. If an interface has remaining bandwidth, packets in AF queues obtain the remaining bandwidth based on weights. Whencongestion occurs, each type of packets can obtain the minimum bandwidth.

If the length of an AF queue reaches the maximum value, the tail drop method is used by default. You can choose to use WRED.

Best-Effort (BE) queues are applied to best-effort services that require no strict QoSassurance.

If packets do not match any configured traffic classifiers, packets match the defaulttraffic classifier defined by the system. You are allowed to configure AF queues and

bandwidth for the default traffic classifier, whereas BE queues are configured in mostsituations. BE uses WFQ scheduling so that the system schedules packets matching thedefault traffic classifier based on flows.

If the length of a BE queue reaches the maximum value, the tail drop method is used by default. You can choose to use WRED.

HQoS

The traditional QoS technology schedules packets based on interfaces. An interface, however,can identify only priorities of different services but cannot identify services of different users.Packets of the same priority are placed into the same queue on an interface, and compete for thesame queue resource. Therefore, the traditional QoS technology is unable to providedifferentiated services based on types of traffic and users.

As the number of users increases continuously and services develop, users require differentiatedservices to have better QoS and gain more profits. Hierarchical QoS (HQoS) implementshierarchical scheduling based on queues and differentiates services and users. It provides QoSguarantee and saves network operation and maintenance costs.

l Queues supported by HQoS

The AR150/200 supports three levels of queues, that is, level-3 flow queue (FQ), level-2subscriber queue (SQ), and level-1 port queue (PQ). The HQoS hierarchy is a tree structure.The flow queue is taken as the leaf node and the port queue is taken as the root node. When

packets pass through an interface using HQoS, the packets are classified so that theytraverse the branches of the tree. Packets arrive at the top of the tree and are classified onone of the leaves. Packets then traverse down the tree until they are transmitted out theinterface at the root.

l HQoS implementationHQoS is implemented by traffic policy nesting supported by the AR150/200.

A traffic policy can be nested into another traffic policy. That is, the traffic behavior or action in a traffic policy is a sub traffic policy. When a sub traffic policy is bound to a traffic

policy, the traffic behavior in the traffic policy is taken for packets matching the trafficclassifier associated with the traffic behavior. Then the packets are classified by the subtraffic policy and the traffic behavior in the sub traffic policy is taken for the classified

packets.

A traffic classifier in the traffic policy differentiates users. That is, the packets that matchthe traffic classifier in the traffic policy enter the same subscriber queue.

The traffic classifier in a sub traffic policy differentiates services. That is, the packets thatmatch the traffic classifier in the sub traffic policy enter the same flow queue.



11


19/111

l HQoS scheduling

HQoS implements hierarchical scheduling and provides good service support. Scheduler levels are relevant to the topology.

The AR150/200 provides three levels of schedulers, that is, flow queue scheduler,subscriber queue scheduler, and port queue scheduler. The flow queue scheduler andsubscriber queue scheduler support PQ scheduling, WFQ scheduling, and PQ+WFQscheduling. The port queue scheduler uses RR scheduling.

HQoS deployment for enterprise users is used as an example. Enterprise users have VoIPservices, video conference (VC) services, and data services. Each subscriber queuecorresponds to one enterprise user and each flow queue corresponds to a type of services.By deploying HQoS, the AR150/200 implements the following functions: Controlling traffic scheduling among the three types of services of a single enterprise

user Controlling total bandwidth of the three types of services of a single enterprise user

Controlling bandwidth allocation between multiple enterprise users Controlling total bandwidth of multiple enterprise users

1.3 Configuring Priority Mapping After priority mapping is configured, the AR150/200 determines the queues and output prioritiesof the received packets based on packet priorities or the default 802.1p priority of the interface.By doing this, the AR150/200 provides differentiated services.

1.3.1 Establishing the Configuration TaskBefore configuring priority mapping, familiarize yourself with the applicable environment,complete the pre-configuration tasks, and obtain the data required for the configuration. Thiswill help you complete the configuration task quickly and accurately.

Applicable Environment

The AR150/200 can determine the queues that packets enter based on packet priorities (802.1p/DSCP priorities) or the priority of an interface. The AR150/200 can re-mark packet prioritiesso that the connected device can provide differentiated QoS based on precedence fields of the

packets.

l When the AR150/200 is configured to trust DSCP priorities of packets, it maps DSCP priorities to 802.1p priorities based on priority mapping. The AR150/200 also determinesthe queues that packets enter, and re-marks packet priorities.

l When the AR150/200 is configured to trust 802.1p priorities of packets or uses the priorityof an interface, it determines the queues that packets enter based on the 802.1p priority andre-marks packet priorities.

Pre-configuration Tasks

Before configuring priority mapping, complete the following tasks:l Configuring link layer attributes of interfaces to ensure that these interfaces work properlyl Configuring IP addresses and routing protocols for interfaces to ensure connectivity



12


20/111

Data Preparation

To configure priority mapping, you need the following data.

No. Data

1 Types and numbers of interfaces

2 Priority of an interface

3 Packet priority trusted by an interface

4 Input priority and output priority

1.3.2 Configuring the Packet Priority Trusted by an InterfaceAfter an interface is configured to trust the priority of packets, the AR150/200 performs mapping

based on the specified priority.

Context

The AR150/200 trusts the following priorities:l 802.1p priority

The AR150/200 determines the queues that tagged packets enter based on 802.1p priorities in the tagged packets and modifies packet priorities.

The AR150/200 determines the queues that untagged packets enter based on the default802.1p priority of an interface and modifies packet priorities.

l DSCP priority

The AR150/200 maps DSCP priorities of packets to 802.1p priorities, determines thequeues that packets enter, and modifies packet priorities based on the priority mappingtable.

l Priority of an interface

The AR150/200 determines the queues that packets enter based on the priority of theinterface, and modifies packet priorities based on the priority mapping table.

Procedure

Step 1 Run:system-view

The system view is displayed.

Step 2 Run:interface interface-type interface-number [. subinterface-number ]

The interface view or sub-interface view is displayed.

Step 3 Run:trust { 8021p | dscp } [ override ]

The packet priority trusted by the interface is configured.



13


21/111

By default, the priority of an interface is used.

----End

1.3.3 Setting the Default 802.1p Priority of an InterfaceThe default 802.1p priority of an interface can be set. By default, the default 802.1p priority of an interface is 0.

ContextThe 802.1p priority is determined by the 3-bit priority field contained in a VLAN tag. The 802.1p

priority is used to provide differentiated services.

The default 802.1p priority of an interface is used in the following situations presented in thefollowing table.

Interface Configuration Whether a ReceivedPacket Carries a VLANTag

Processing Method

Default configuration (the priority of an interface isused)

No The AR150/200 searches for the mapping table between802.1p priorities and other

priorities based on the priority of an interface,modifies the packet priority,and send the packet to aqueue based on the modified

802.1p priority.Default configuration (the

priority of an interface isused)

Yes The AR150/200 searches for the mapping table between802.1p priorities and other

priorities based on the priority of an interface,modifies the packet priority,and sends the packet to aqueue based on the modified802.1p priority.

trust 8021p override No The AR150/200 searches for the mapping table between802.1p priorities and other

priorities based on the priority of an interface,modifies the packet priority,and sends the packet to aqueue based on the 802.1p

priority.



14


22/111

Interface Configuration Whether a ReceivedPacket Carries a VLANTag

Processing Method

trust 8021p No The AR150/200 searches for the mapping table between802.1p priorities and other

priorities based on the priority of an interface,modifies the packet priority,and sends the packet to aqueue based on the 802.1p

priority.

ProcedureStep 1 Run:

system-view


Step 2 Run:interface interface-type interface-number

The interface view is displayed.

Step 3 Run: port priority priority-value

The priority of the interface is set.

By default, the priority of an interface is 0.

----End

1.3.4 Configuring a Priority Mapping TableA priority mapping table includes mappings between 802.1p priorities, DSCP priorities.

ContextThe AR150/200 performs priority mapping based on packet priorities or the default priority of an interface. Mappings between priorities can be configured in the priority mapping table.

Procedure



Step 2 Run:qos map-table { dot1p-dot1p | dot1p-dscp | dscp-dot1p | dscp-dscp }



15


23/111

The priority mapping table view is displayed.

Step 3 Run:input { input-value1 [ to input-value2 ] } & output output-value

Mappings are configured in the priority mapping table.

----End

1.3.5 Checking the ConfigurationAfter priority mapping is configured, you can view global priority mapping.

Prerequisites

All the priority mapping configurations are complete.

Procedurel Run the display qos map-table [ dot1p-dot1p | dot1p-dscp | dscp-dot1p | dscp-dscp ]

command to view the priority mapping.

----End

1.4 Configuring Traffic Policing The AR150/200 supports interface-based traffic policing and flow-based traffic policing.

1.4.1 Establishing the Configuration TaskBefore configuring traffic policing, familiarize yourself with the applicable environment,complete the pre-configuration tasks, and obtain the data required for the configuration. Thiswill help you complete the configuration task quickly and accurately.


A network is congested if traffic sent by users is not limited. To make use of limited network resources, limit the user traffic.l Interface-based traffic policing limits the incoming or outgoing service traffic on an

interface.NOTE

You can use interface-based traffic policing to:l Limit all the service traffic on an interface.l Limit the service traffic matching a specified ACL rule.l Limit the service traffic whose source and destination IP addresses within a specified range.

l Flow-based traffic policing limits incoming or outgoing service traffic matching trafficclassification rules.


Before configuring traffic policing, complete the following tasks:



16


24/111

l Configuring link layer attributes of interfaces to ensure that these interfaces work properlyl Configuring IP addresses and routing protocols for interfaces to ensure connectivity

Data PreparationTo configure interface-based traffic policing, you need the following data.

No. Data

1 Interface and direction to which the traffic policy is applied

2 CIR value and optional parameters including the PIR value, CBS value, PBS value,color, and coloring mode

To configure flow-based traffic policing, you need the following data.

No. Data

1 Traffic classifier name and related parameters

2 Traffic behavior name and committed access rate (CAR) parameters: CIR value andoptional parameters including the PIR value, CBS value, PBS value, color, andcoloring mode

3 Traffic policy name and interface and direction to which the traffic policy is applied

1.4.2 Configuring Traffic Policing on an InterfaceYou can configure CAR on an interface in the inbound or outbound direction to limit the rateof incoming or outgoing traffic on the interface.

ContextTo limit all the incoming or outgoing traffic on an interface, configure traffic policing on theinterface. If the rate of received or sent packets exceeds the rate limit, packets are discarded.

Procedure





Step 3 Run the following commands as required.l On the WAN side, run:

qos car { inbound | outbound } [ acl acl-number | { destination-ip-address range | source-ip-address range } start-ip-address to end-ip-address [ per-



17


25/111

address ] ] cir cir-value [ pir pir-value ] [ cbs cbs-value pbs pbs-value ][ green { discard | pass [ remark-8021p 8021p-value | remark-dscp dscp-value ] } ] [ yellow { discard | pass [ remark-8021p 8021p-value | remark-dscp dscp-value ] } ] [ red { discard | pass [ remark-8021p 8021p-value | remark-dscp dscp-value ] } ]

Traffic policing is configured on the WAN-side interface.By default, traffic policing is not performed on an interface.

NOTE

If the CBS and PBS values are not specified, the ir values are determined as follows:

l If the PIR value is not set or the PIR and CIR values are the same, the CBS value is 188 times theCIR value and the PBS value is 313 times the CIR value.

l If the PIR value is set and the PIR and CIR values are different, the CBS value is 125 times theCIR value and the PBS value is 125 times the PIR value.

When the CBS value is less than the number of bytes in a single packet of a service, packets of theservice are discarded.

l On the LAN side, run:qos car { inbound cir cir-value | { inbound | outbound } { acl acl-number |{ destination-ip-address range | source-ip-address range } start-ip-address to end-ip-address [ per-address ] } cir cir-value [ pir pir-value ] [ cbs cbs-value pbs pbs-value ] [ green { discard | pass [ remark-8021p 8021p-value |remark-dscp dscp-value ] } ] [ yellow { discard | pass [ remark-8021p 8021p-value | remark-dscp dscp-value ] } ] [ red { discard | pass [ remark-8021p 8021p-value | remark-dscp dscp-value ] } ] }

Traffic policing is configured on the LAN-side interface.

By default, traffic policing is not performed on an interface.

NOTE

LAN-side interfaces on the AR150/200 support only traffic policing in the inbound direction.

----End

1.4.3 Configuring Flow-based Traffic Policing After traffic policing is configured in a traffic policy, the traffic policy can be applied to multipleinterfaces to limit the rate of incoming or outgoing traffic on the interfaces.

Context

To limit the rate of incoming or outgoing traffic matching traffic classification rules, configureflow-based traffic policing. A traffic policy can be applied to multiple interfaces. When thetraffic rate exceeds the rate limit, traffic is discarded. Flow-based traffic policing can implementdifferentiated services using complex traffic classification.

Procedure

Step 1 Configure a traffic classifier.

The AR150/200 can classify traffic according to the ACL, Layer 2 information in packets, andLayer 3 information in packets. Configure a traffic classifier by selecting appropriate trafficclassification rules. For details, see 2.3 Configuring a Traffic Classifier .

Step 2 Configure a traffic behavior.



18


26/111

Create a traffic behavior and configure the CAR action in the traffic behavior. For details, see2.4.5 Configuring Traffic Policing .

Step 3 Configure a traffic policy.

Create a traffic policy, associate the traffic classifier and traffic behavior with the traffic policy,and apply the traffic policy to an interface. For details, see 2.5 Configuring a Traffic Policy .

----End

1.4.4 Checking the ConfigurationAfter traffic policing is configured, you can view the rate limit.

Prerequisites

All the traffic policing configurations are complete.

Procedurel Checking the interface-based traffic policing configuration

Run the display this command in the interface view to check the traffic policingconfiguration on the interface.

l Checking the flow-based traffic policing configuration Run the display traffic behavior { system-defined | user-defined } [ behavior-

name ] command to check the traffic behavior configuration. Run the display traffic classifier { system-defined | user-defined } [ classifier-

name ] command to check the traffic classifier configuration. Run the display traffic policy user-defined [ policy-name [ classifier classifier-

name ] ] command to check the traffic policy configuration. Run the display traffic-policy policy-name applied-record command to check the

specified traffic policy record.l Checking the packet statistics on the interface configured with traffic policing

Run the display qos car statistics interface interface-type interface-number { inbound | outbound }or display qos car statistics interface { virtual-template vt-number virtual-access va-number | dialer number } { inbound | outbound } commandto check the statistics on forwarded and discarded packets on the interface.

----End

1.5 Configuring Traffic Shaping Traffic shaping enables outgoing traffic to be sent out at an even rate and reduces the number of discarded packets whose rate exceeds the rate limit.

1.5.1 Establishing the Configuration TaskBefore configuring traffic shaping, familiarize yourself with the applicable environment,

complete the pre-configuration tasks, and obtain the data required for the configuration. Thiswill help you complete the configuration task quickly and accurately.



19


27/111


28/111

No. Data

1 Number of the interface and index of the queue to which traffic shaping is applied

2 Queue profile name

3 CIR value and (optional) CBS value

4 Number of the interface to which a queue profile is applied

To configure flow-based traffic policing, you need the following data.

No. Data


2 Traffic behavior name, and CIR value and optional parameters including the PIR value, PBS value, and queue length

3 Traffic policy name and interface to which the traffic policy is applied

1.5.2 Configuring Interface-based Traffic Shaping You can configure traffic shaping on an interface to limit the rate of data sent by the interface.

ContextTo limit all the outgoing traffic on an interface, configure traffic shaping on the interface. Trafficshaping buffers the packets whose rate is greater than the traffic shaping rate. When there aresufficient tokens in the token bucket, the buffered packets are forwarded at an even rate. If queuesare full, packets are discarded.

Procedure


The system view is displayed.Step 2 Run:

interface interface-type interface-number [. subinterface-number ]


Step 3 Run:qos gts cir cir-value [ cbs cbs-value ]

Traffic shaping is configured.

By default, traffic shaping is not performed on an interface. When you configure traffic shapingon an interface without specifying the CBS value, the CBS value is 25 times the CIR value.

----End



21


29/111

1.5.3 Configuring Queue-based Traffic Shaping This section describes how to configure a queue profile, set traffic shaping parameters, and applya traffic shaping profile to an interface.

Context

To shape packets in each queue on an interface, configure a queue profile and apply it to theinterface. The received packets enter different queues based on priority mapping. Differentiatedservices are provided by setting different traffic shaping parameters for queues with different

priorities.

Procedure



Step 2 Run:qos queue-profile queue-profile-name

A queue profile is created and the queue profile view is displayed.

Step 3 Run:queue { start-queue-index [ to end-queue-index ] } & length { bytes bytes-value | packets packets-value } *

The length of each queue is set.

By default, the length of a queue using PQ on the WAN side is 40960 bytes; the length of aqueue using WFQ on the WAN side is 131072 bytes; the length of a queue using PQ, DRR, or WRR on the LAN side is 5120 bytes.

NOTE

The layer 2 FE interface on the AR150/200 does not support the queue length command.

Step 4 Run:queue { start-queue-index [ to end-queue-index ] } & gts cir cir-value [ cbs cbs-value ]

Queue-based traffic shaping is configured.

By default, traffic shaping is not performed for queues. When you configure traffic shaping onan interface without specifying the CBS value, the CBS value is 25 times the CIR value.

Step 5 Run:quit

Exit from the queue profile view.






22


30/111

The queue profile is applied to the interface or sub-interface.

----End

1.5.4 Configuring Flow-based Traffic Shaping After traffic shaping is configured in a traffic policy, the traffic policy can be applied to multipleinterfaces to limit the rate of traffic of a specified type on the interfaces.

Context

To limit all the outgoing traffic of a specified type on an interface, configure flow-based trafficshaping. A traffic policy can be applied to different interfaces. When the rate of packets matchingthe traffic policy exceeds the rate limit, traffic shaping buffers the excess packets. When thereare sufficient tokens in the token bucket, the buffered packets are forwarded at an even rate.When the token bucket is full, the packets are discarded. Flow-based traffic shaping can

implement differentiated services using complex traffic classification.

Procedure


The AR150/200 can classify traffic according to the ACL, Layer 2 information in packets, andLayer 3 information in packets. Configure a traffic classifier by selecting appropriate trafficclassification rules. For details, see 2.3 Configuring a Traffic Classifier .


Create a traffic behavior and configure the GTS action in the traffic behavior. For details, see2.4.6 Configuring Traffic Shaping .



----End

1.5.5 Checking the ConfigurationAfter traffic shaping is configured, you can view the traffic shaping rate.

Prerequisites

All the traffic shaping configurations are complete.

Procedurel Check the traffic shaping configuration in the interface view.

Run the display this command in the interface view to check the traffic shapingconfiguration on the interface.

l Check the traffic shaping configuration in the queue profile view.

Run the display this command in the interface view to check the queue profile boundto the interface.



23


31/111

Run the display qos queue-profile [ queue-profile-name ] command to check the queue profile configuration.

l Check the traffic shaping configuration in the traffic behavior view. Run the display traffic behavior { system-defined | user-defined } [ behavior-




specified traffic policy record.

----End

1.6 Configuring Congestion ManagementIf congestion occurs on a network after congestion management is configured, the AR150/200determines the sequence at which packets are forwarded according to the defined scheduling

policy.

1.6.1 Establishing the Configuration TaskBefore configuring congestion management, familiarize yourself with the applicableenvironment, complete the pre-configuration tasks, and obtain the data required for the

configuration. This will help you complete the configuration task quickly and accurately.


When congestion occurs on a network, configure congestion management to implement thefollowing functions:l Smooth out the delay and jitter.l Preferentially process packets of delay-sensitive services, such as video services and voice

services.l Process packets with the same priority uniformly and process packets with different

priorities based on packet priorities among delay-insensitive services, for example, emailservices.

The AR150/200 supports queue-based congestion management and class-based congestionmanagement:l Queue-based congestion management: When packets enter queues on an interface based

on packet priorities, configure different scheduling modes for queues using a queue profileso that differentiated services are provided.

l Class-based congestion management: The AR150/200 provides EF, AF, and BE queuesand different scheduling modes for packets matching traffic classifiers so that differentiatedservices can be provided.



24


32/111

NOTE

Queue-based congestion management and class-based congestion management cannot be configuredsimultaneously.

Class-based congestion management can only be configured on the AR150/200 WAN-side interfaces but

not on the LAN-side interfaces.Congestion management can be configured on the logical interface, including the dialer interface, MP-Group interface, virtual template interface, virtual Ethernet interface, tunnel interface, or the physicalinterface corresponding to the virtual interface. If traffic shaping, congestion management, congestionavoidance, or a combination of them is configured on the virtual interface, the configuration on the virtualinterface takes effect. The configuration on the physical interface, however, does not take effect.

Pre-configuration TasksBefore configuring congestion management, complete the following tasks:l Configuring priority mappingl Configuring priority re-marking based on traffic classification

Data PreparationTo configure queue-based congestion management, you need the following data.

No. Data

1 Interface and queue to which congestion management is applied


3 Queue scheduling mode

4 (Optional) Queue length

5 (Optional) Queue weight

To configure class-based congestion management, you need the following data.

No. Data


2 Traffic behavior name and scheduling mode


1.6.2 Configuring Queue-based Congestion ManagementThe AR150/200 supports the following scheduling modes: PQ, DRR, WFQ, WRR, PQ+DRR,PQ+WFQ, and PQ+WRR.

Context

After packets enter queues on an interface based on priority mapping, they are scheduledaccording to rules. Interfaces on the AR150/200 support different scheduling modes. PQ queues



25


33/111

are scheduled first, and multiple queues are scheduled in descending order of priorities. After all the PQ queues are scheduled, the AR150/200 schedules DRR, WFQ, or WRR queues in turn.

Table 1-4 Scheduling mode supported by each interface

Interface Scheduling Mode

LAN-side interface l PQl DRR l WRR l PQ+DRR l PQ+WRR NOTE

Layer 2 FE interfaces on the AR150/200 do notsupport DRR scheduling. They support thefollowing scheduling modes:l PQl WRR l PQ+WRR

WAN-side interface l PQl WFQl PQ+WFQ

ProcedureStep 1 Run:

system-view



A queue profile is created and the queue profile view is displayed.

Step 3 Run the following commands as required.

l On the WAN-side interface, run:schedule { { pq start-queue-index [ to end-queue-index ] } | { wfq start-queue-index [ to end-queue-index ] } } *

A scheduling mode is configured for each queue on the WAN-side interface.l On the LAN-side interface, run:

schedule { { pq start-queue-index [ to end-queue-index ] } | { drr start-queue-index [ to end-queue-index ] } | { wrr start-queue-index [ to end-queue-index ] } } *

A scheduling mode is configured for each queue on the LAN-side interface.

By default, all the queues on the LAN side use WRR and all the queues on the WAN useWFQ.

Step 4 (Optional) Run:



26


34/111

queue { start-queue-index [ to end-queue-index ] } & length { bytes bytes-value | packets packets-value } *

The length of each queue is set on the interface.

By default, the length of a queue using PQ, DRR, or WRR on the LAN side is 5120 bytes; thelength of a queue using PQ on the WAN side is 40960 bytes; the length of a queue using WFQon the WAN side is 131072 bytes.

NOTE

The Layer 2 FE interface on the AR150/200 does not support the queue length command.

Step 5 (Optional) Run:queue { start-queue-index [ to end-queue-index ] } & weight weight-value

The weight value of each queue is set on the interface.

By default, the weight value of a queue is 10.

NOTE

The Layer 2 FE interface on the AR150/200 does not support the queue weight command.

Step 6 Run:quit






----End

1.6.3 Configuring Class-based Congestion ManagementAfter a traffic policy is configured, it can be applied to multiple interfaces so that the samescheduling mode can be applied to traffic of a specified type on multiple interfaces.

Context

The AR150/200 provides the following queues for data packets matching traffic classificationrules:l AF: ensures low drop probability of packets when the rate of outgoing service traffic does

not exceed the minimum bandwidth. It is applied to services of heavy traffic that need to be ensured.

l EF: is applied to services requiring a low delay, low drop probability, and assured bandwidth. EF is also applied to services occupying low bandwidth, for example, voice packets. After packets matching traffic classification rules enter EF queues, they are

scheduled in Strict Priority (SP) mode. Packets in other queues are scheduled only after allthe packets in EF queues are scheduled.



27


35/111

l BE: is used with the default traffic classifier. The remaining packets that do not enter AFor EF queues enter BE queues. BE queues use WFQ scheduling. When a greater number of queues are configured, WFQ allocates bandwidth more evenly but more resources areoccupied. WFQ is applied to the services insensitive to the delay and packet loss, for example, Internet access services.

Class-based congestion management, also called CBQ, on the main interface or sub-interface isexclusive with the queue profile or traffic shaping on the same main interface or sub-interface.

CBQ Configuration Whether the QueueProfile Can BeConfigured (qos queue-profile (interface view))

Whether Traffic Shaping Can Be Configured (qosgts)

Main interface Main interface: No Main interface: Yes

Sub-interface: No Sub-interface: No

Sub-interface Main interface: Yes Main interface: Yes

Sub-interface: No Sub-interface: Yes

Procedure


The AR150/200 can clas sify traffic according to the ACL, Layer 2 info rmation in packets, andLayer 3 information in packets. Configure a traffic classifier by selecting appropriate trafficclassification rules. For details, see 2.3 Configuring a Traffic Classifier .


Create a traffic behavior and configure flow-based queue scheduling in the traffic behavior. For details, see 2.4.7 Configuring Congestion Management .



----End

1.6.4 Checking the ConfigurationAfter congestion management is configured, you can view scheduling parameters of each queueon a specified interface.

Prerequisites

All the congestion management configurations are complete.

Procedurel Checking the queue-based congestion management configuration



28


36/111


Run the display qos queue-profile [ queue-profile-name ] command to check the queue profile configuration.

l Checking the class-based congestion management configuration Run the display traffic behavior { system-defined | user-defined } [ behavior-




specified traffic policy record.

----End

1.7 Configuring Congestion AvoidanceAfter congestion avoidance is complete, the AR150/200 discards the packets whose rate exceedsthe rate limit based on the WRED configuration.

1.7.1 Establishing the Configuration TaskBefore configuring congestion avoidance, familiarize yourself with the applicable environment,complete the pre-configuration tasks, and obtain the data required for the configuration. This

will help you complete the configuration task quickly and accurately.

Applicable EnvironmentBy default, the AR150/200 uses the tail drop method and discards data packets at the end of aqueue when congestion occurs. The tail drop method may cause global Transmission ControlProtocol (TCP) synchronization, and it reduces link usage. The Weighted Random EarlyDetection (WRED) can solve these problems.

The AR150/200 supports queue-based congestion avoidance and flow-based congestionavoidance:l Queue-based congestion avoidance: In a queue profile, different drop profiles are bound

to queues. Different WRED parameters in drop profiles take effect for queues with different priorities so that differentiated services are provided.

l Flow-based congestion avoidance: The AR150/200 provides EF, AF, and BE queues for packets matching traffic classification rules. EF queues can use only tail drop, and AF andBE queues can bind drop profiles to traffic behaviors so that differentiated services are

provided.



29


37/111

NOTE

Queue-based congestion avoidance and flow-based congestion avoidance are exclusive.

Congestion avoidance can only be configured on the AR150/200 WAN-side interfaces but not on the LAN-side interfaces.

Congestion avoidance can be configured on the logical interface, including the dialer interface, MP-Groupinterface, virtual template interface, virtual Ethernet interface, tunnel interface, or the physical interfacecorresponding to the virtual interface. If traffic shaping, congestion management, congestion avoidance,or a combination of them is configured on the virtual interface, the configuration on the virtual interfacetakes effect. The configuration on the physical interface, however, does not take effect.


Before configuring congestion avoidance, complete the following tasks:l Configuring priority mappingl Configuring priority re-marking based on traffic classification

l Configuring congestion management

Data Preparation

To configure queue-based congestion avoidance, you need the following data.

No. Data

1 Drop profile name and WRED parameters


3 Interface and queue to which congestion avoidance is applied

To configure flow-based congestion avoidance, you need the following data.

No. Data

1 Drop profile name and WRED parameters


3 Traffic behavior name


1.7.2 Configuring Queue-based WREDTo reduce congestion on a WAN-side interface, apply a drop profile to the interface.

Context

A drop profile defines WRED parameters and is used to implement congestion avoidance for queues bound to the drop profile.



30


38/111

After a drop profile is bound to a queue profile, bind the queue profile to an interface so thatWRED parameters in the drop profile take effect on the interface.

The AR150/200 supports WRED based on DSCP priorities or IP priorities:l The value of an IP precedence ranges from 0 to 7.l The value of a DSCP priority ranges from 0 to 63.l Eight DSCP priorities correspond to one IP precedence.

WRED based on DSCP priorities differentiates services in a more refined manner.

NOTE

Drop profiles can be bound to only queues using WFQ on the AR150/200 WAN interfaces.

Procedure

Step 1 Configuring a drop profile1. Run:

system-view


2. Run:drop-profile drop-profile-name

A drop profile is created and the drop profile view is displayed.

3. (Optional) Run: wred { dscp | ip-precedence }

A WRED drop profile based on DSCP or IP priorities is configured.4. Run the following commands as required.

l Run:dscp { dscp-value1 [ to dscp-value2 ] } & low-limit low-limit-

percentage high-limit high-limit-percentage discard-percentage discard- percentage

WRED parameters based on DSCP priorities are set.l Run:

ip-precedence { ip-precedence-value1 [ to ip-precedence-value2 ] } &low-limit low-limit-percentage high-limit high-limit-percentage discard-

percentage discard-percentage

WRED parameters based on IP priorities are set.5. Run:

quit

Exit from the drop profile view.

Step 2 Applying the drop profile1. Run:

qos queue-profile queue-profile-name

The queue profile view is displayed.

The drop profile can be an existing drop profile or a new drop profile. You can set thescheduling mode, queue weight, queue length, and queue shaping in the queue profile.



31


39/111

2. Run:schedule wfq start-queue-index [ to end-queue-index ]

WFQ is specified for the specified queue in the queue profile.

3. Run:queue { start-queue-index [ to end-queue-index ] } & drop-profile drop-

profile-name

A drop profile is bound to a queue in a queue profile.

By default, no queue is bound to a drop profile. All queues use tail drop.

4. Run:quit


5. Run:interface interface-type interface-number [. subinterface-number ]


6. Run:qos queue-profile queue-profile-name


----End

1.7.3 Configuring Flow-based WREDTo reduce congestion in AF and BE queues, bind a drop profile to a traffic behavior of a traffic

policy.

PrerequisitesClass-based congestion management has been configured.

ContextA drop profile defines WRED parameters and is used to implement congestion avoidance for queues bound to the drop profile.

After a drop profile is bound to a traffic behavior, bind the traffic behavior and traffic classifier to a traffic policy and apply the traffic policy to an interface so that WRED parameters in the

drop profile take effect on the interface.The AR150/200 supports WRED based on DSCP priorities or IP priorities:l The value of an IP precedence ranges from 0 to 7.l The value of a DSCP priority ranges from 0 to 63.l Eight DSCP priorities correspond to one IP precedence.

WRED based on DSCP priorities differentiates services in a more refined manner.

NOTE

Congestion avoidance can only be configured on the AR150/200 WAN-side interfaces but not on the LAN-side interfaces.

A drop profile takes effect for only AF and BE queues; therefore, class-based congestion management musthave been configured before you configure flow-based congestion avoidance.



32


40/111

Procedure

Step 1 Configuring a drop profile

1. Run:

system-view


2. Run:drop-profile drop-profile-name

A drop profile is created and the drop profile view is displayed.

3. (Optional) Run: wred { dscp | ip-preceden ce }

A WRED drop profile based on DSCP or IP priorities is configured.

4. Run the following commands as required.l Run:

dscp { dscp-value1 [ to dscp-value2 ] } & low-limit low-limit- percentage high-limit high-limit-percentage discard-percentage discard- percentage

WRED parameters based on DSCP priorities are set .l Run:

ip-precedence { ip-precedence-value1 [ to ip-precedence-value2 ] } &low-limit low-limit-percentage high-limit high-limit-percentage discard-

percentage discard-percentage

WRED parameters based on IP priorities are set.

5. Run:quit

Exit from the drop profile view.

Step 2 Applying the drop profile

1. Configure a traffic classifier.

The AR150/200 can classify traffic according to the ACL, Layer 2 information in packets,and Layer 3 information in packets. Configure a traffic classifier by selecting appropriatetraffic classification rules. For details, see 2.3 Configuring a Traffic Classifier .

2. Configure a traffic behavior.

Create a traffic behavior and configure flow-based congestion avoidance in the traffic behavior. For details, see 2.4.8 Configuring Congestion Avoidance .

3. Configure a traffic policy.

Create a traffic policy, associate the traffic classifier and traffic behavior with the traffic policy, and apply the traffic policy to an interface. For details, see 2.5 Configuring aTraffic Policy .

----End

1.7.4 Checking the ConfigurationAfter congestion avoidance is configured, you can view WRED parameters in a drop profile.



33


41/111

Prerequisites

All the congestion avoidance configurations are complete.

Procedurel Checking the queue-based congestion avoidance configuration


Run the display this command in the queue profile view to check the drop profile boundto the queue profile.

Run the display drop-profile [ drop-profile-name ] command to check the drop profileconfiguration.

l Checking the flow-based congestion avoidance configuration Run the display traffic behavior { system-defined | user-defined } [ behavior-



name ] ] command to check the traffic policy configuration. Run the display traffic-policy policy-name applied-record command to c

Documents

AR150 - Configuration Guide - QoS(V200R002C00_02)