iuwvn v1.0 student guide_vol1

IUWVN

Implementing Cisco Unified Wireless Voice Networks

Volume 1

Version 1.0

Student Guide

Text Part Number: 97-2791-03

www.CareerCert.info

http://www.careercert.info

Student Guide © 2009 Cisco Systems, Inc. All Rights Reserved.

DISCLAIMER WARRANTY: THIS CONTENT IS BEING PROVIDED “AS IS.” CISCO MAKES AND YOU RECEIVE NO WARRANTIES IN

CONNECTION WITH THE CONTENT PROVIDED HEREUNDER, EXPRESS, IMPLIED, STATUTORY OR IN ANY OTHER PROVISION OF

THIS CONTENT OR COMMUNICATION BETWEEN CISCO AND YOU. CISCO SPECIFICALLY DISCLAIMS ALL IMPLIED

WARRANTIES, INCLUDING WARRANTIES OF MERCHANTABILITY, NON-INFRINGEMENT AND FITNESS FOR A PARTICULAR

PURPOSE, OR ARISING FROM A COURSE OF DEALING, USAGE OR TRADE PRACTICE. This learning product may contain early release

content, and while Cisco believes it to be accurate, it falls subject to the disclaimer above.

www.CareerCert.info


Students, this letter describes important course evaluation access information!

Welcome to Cisco Systems Learning. Through the Cisco Learning Partner Program, Cisco Systems is committed to bringing you the highest-quality training in the industry. Cisco learning products are designed to advance your professional goals and give you the expertise you need to build and maintain strategic networks.

Cisco relies on customer feedback to guide business decisions; therefore, your valuable input will help shape future Cisco course curricula, products, and training offerings. We would appreciate a few minutes of your time to complete a brief Cisco online course evaluation of your instructor and the course materials in this student kit. On the final day of class, your instructor will provide you with a URL directing you to a short post-course evaluation. If there is no Internet access in the classroom, please complete the evaluation within the next 48 hours or as soon as you can access the web.

On behalf of Cisco, thank you for choosing Cisco Learning Partners for your Internet technology training.

Sincerely,

Cisco Systems Learning

www.CareerCert.info


www.CareerCert.info


Table of Contents Volume 1 Course Introduction 1

Overview 1 Learner Skills and Knowledge 1

Course Goal and Objectives 2 Course Flow 4 Additional References 5

Cisco Glossary of Terms 5 Your Training Curriculum 6

Implementation of QoS for Wireless Applications 1-1

Overview 1-1 Module Objectives 1-1

Identifying General Considerations for Wired and Wireless QoS 1-3 Overview 1-3

Objectives 1-3 QoS Overview 1-4 Classification and Marking 1-6 Trust Boundary 1-11 Congestion Management 1-13 Congestion Avoidance 1-15 Policing and Shaping 1-17 Link Efficiency Mechanisms 1-19 QoS in the Network 1-21 Congestion in Wireless Cells 1-22 802.11 DCF 1-24 Describing the 802.11e Protocol 1-36 Describing WMM Implementations 1-39 802.11e, 802.1p, and DSCP Mapping 1-56 Summary 1-58

References 1-59 Describing Wireless QoS Deployment Schemes 1-61

Overview 1-61 Objectives 1-61

QoS Parameters 1-62 Upstream and Downstream QoS 1-63 QoS and Network Performance 1-73 Summary 1-74

Configuring the Controller and Cisco WCS for QoS 1-75 Overview 1-75

Objectives 1-75 Assign a Cisco WLC QoS Profile to a WLAN on a Controller 1-76 Configure Cisco WLC QoS Profiles 1-80 Configure Cisco WLC Voice Parameters 1-83 Configure Cisco WLC EDCA Support 1-86 Configure Cisco WLC QoS Roles 1-88 Configure QoS Using Cisco WCS 1-94 Summary 1-98

References 1-98

www.CareerCert.info


ii Implementing Cisco Unified Voice Networks (IUWVN) v1.0 © 2009 Cisco Systems, Inc.

Configuring the Wired Infrastructure for QoS 1-99 Overview 1-99

Objectives 1-99 Modular QoS CLI 1-100 Modular QoS CLI Components 1-101

Example: Configuring MQC 1-101 Class Maps 1-102 Configure and Monitor Class Maps 1-103

Example: Class Map Configuration 1-103 Example: Using the match Command 1-106 Example: Nested Traffic Class to Combine match-any and match-all Characteristics in One Traffic Class 1-106

Policy Maps 1-109 Configure and Monitor Policy Maps 1-110

Example: Policy Map Example 1-111 Example: Policy Map 1-112

Service Policy 1-114 Attach Service Policies to Interfaces 1-115

Example: Complete MQC Configuration 1-116 Policy Map Examples 1-118 QoS on a Switch 1-122 Configure QoS on a Switch 1-126

Example Default Standard 2960 QoS Configuration 1-126 Monitor QoS on a Switch 1-133 Summary 1-135

Understanding Current Best-Practice Guidelines 1-137 Overview 1-137

Objectives 1-137 Throughput 1-138 Switch QoS Configuration 1-139 CAPWAP Traffic Classification 1-141 CAPWAP Traffic Volumes 1-143 CAPWAP Marking Manipulation 1-148 Summary 1-151 Module Summary 1-152

References 1-153 Module Self-Check 1-154

Module Self-Check Answer Key 1-159

Voice over Wireless Architecture 2-1

Overview 2-1 Module Objectives 2-1

Describing the Evolution of Voice Architecture 2-3 Overview 2-3

Objectives 2-3 Traditional Voice Network 2-4 VoIP Network 2-8 Summary 2-24

Describing VoWLAN Call Flow 2-25 Overview 2-25

Objectives 2-25 Call Setup and Data Flow 2-26 VoIP over Wireless Protocols 2-28 Infrastructure Hardware and Software Components 2-43

Cisco CAPWAP APs 2-45 Wireless IP Phones 2-51 Summary 2-61

www.CareerCert.info


© 2009 Cisco Systems, Inc. Implementing Cisco Unified Voice Networks (IUWVN) v1.0 iii

Designing Wireless for Voice 2-63 Overview 2-63

Objectives 2-63 General Site Survey Guidelines 2-64 RF Design Guidelines 2-80 Combinations of WLAN Services 2-90 Voice over WLAN Security 2-97 Voice over WLAN Roaming 2-100 Cisco Compatible Extensions for VoWLANs 2-104 WLAN Controller Configuration and Design 2-113 Campus Network Design 2-118 Voice Support in Mesh Environments 2-127 Summary 2-129

Verifying Voice Readiness 2-131 Overview 2-131

Objectives 2-131 Cisco WCS Voice Readiness Tool 2-132 Postdeployment Site Survey 2-134 AirMagnet VoFi Analyzer 2-138 Summary 2-141 Module Summary 2-142 Module Self-Check 2-143

Module Self-Check Answer Key 2-147

www.CareerCert.info


iv Implementing Cisco Unified Voice Networks (IUWVN) v1.0 © 2009 Cisco Systems, Inc.

www.CareerCert.info


IUWVN

Course Introduction

Overview Implementing Cisco Unified Wireless Voice Networks (IUWVN) version1.0 is an instructor-led course presented by Cisco training partners to their end-user customers. This five-day course is designed to help students prepare for the CCNP-Wireless certification, a professional-level certification specializing in the wireless field. The IUWVN course is a component of the CCNP-Wireless curriculum.

The IUWVN course is designed to give students a firm understanding of how to integrate voice-over-WLAN (VoWLAN) services into the WLAN and be able to implement quality of service (QoS) and high-bandwidth applications into the wireless network.

Learner Skills and Knowledge This subtopic lists the skills and knowledge that learners must possess to benefit fully from the course. The subtopic also includes recommended Cisco learning offerings that learners should first complete to benefit fully from this course.

© 2009 Cisco Systems, Inc. All rights reserved. IUWVN v1.0—3

Learner Skills and Knowledge

� Students considered for this training will have attended the following classes or obtained equivalent level:

– ICND1 Interconnecting Cisco Network Devices part 1 v1.0

– ICND2 Interconnecting Cisco Network Devices part 2 v1.0

– IUWNE Implementing Cisco Unified Wireless Networking Essentials v1.0

� Knowledge of the Cisco Lifecycle Services deployment

� Basic knowledge of wireless standards (IEEE), wireless regulatorenvironment (FCC, ETSI, etc.), and wireless certification organization (WiFi Alliance)

www.CareerCert.info


2 Implementing Cisco Unified Voice Networks (IUWVN) v1.0 © 2009 Cisco Systems, Inc.

Course Goal and Objectives This topic describes the course goal and objectives.


Implementing Cisco Unified Wireless Voice Networks v1.0

“Acquire a firm understanding of how to integrate VoWLAN services into the WLAN and be able toimplement QoS and high-bandwidth applications into the wireless network”

Course Goal

Upon completing this course, you will be able to meet these objectives:

� Implement QoS for wireless applications using the best-practices guidelines

� Describe evolution of voice from traditional through VoIP architecture and finally into VoWLAN.

� Implement a VoWLAN network infrastructure

� Implement multicast in a wireless network

� Configure the wireless infrastructure for video and high-bandwidth applications

www.CareerCert.info


© 2009 Cisco Systems, Inc. Course Introduction 3


General Administration

Class-Related� Sign-in sheet

� Length and times

� Break and lunch room locations

� Attire

Facilities-Related� Course materials

� Site emergency procedures

� Rest rooms

� Telephones and faxes

The instructor will discuss the following administrative issues so that you know exactly what to expect from the class:

� Sign-in process

� Start and anticipated end times of each class day

� Class break and lunch facilities

� Appropriate attire during class

� Materials you can expect to receive during class

� What to do in the event of an emergency

� Location of the rest rooms

� How to send and receive telephone and fax messages

www.CareerCert.info



Course Flow This topic presents the suggested flow of the course materials.


Course Flow

AM

PM

Day 1 Day 2 Day 3 Day 4 Day 5

Lunch

QoS forWireless

Applications

Introduction

QoS for WirelessApplications

(Cont.)

Multicast overWireless

Implementation

VoWLAN Implementation

QoS for WirelessApplications

(Cont.)

Voice overWireless

Architecture

Voice overWireless

Architecture(Cont.)

Voice overWireless

Architecture(Cont.)

VoWLAN Implementation

(Cont.)

Multicast overWireless

Implementation(Cont.)

Wireless andHigh Bandwidth

Applications(Cont.)

Wireless andHigh Bandwidth

Applications

The schedule reflects the recommended structure for this course. This structure allows enough time for the instructor to present the course information and for you to work through the lab activities. The exact timing of the subject materials and labs depends on the pace of your specific class.

www.CareerCert.info



Additional References This topic presents the Cisco icons and symbols that are used in this course, as well as information on where to find additional technical references.


Cisco Icons and Symbols

Router

Workgroup Switch

PC

Server

Tablet

Dual-ModeAccess Point

Single-ModeAccess Point

WLANController

Wireless Bridge

Wi-Fi Tag

Wireless Location Appliance/ MSE

WirelessRouter

WiSMPC Card

Scanner

WirelessClient

WirelessLink

IP PhoneWi-FiIP Phone

Cisco Glossary of Terms For additional information on Cisco terminology, refer to the Cisco Internetworking Terms and Acronyms glossary of terms at http://www.cisco.com/en/US/docs/internetworking/terms_acronyms/ita.html.

www.CareerCert.info



Your Training Curriculum This topic presents the training curriculum for this course.


Cisco Career Certifications:

Expand Your Professional Options, Advance Your Career

Cisco Certified Network Professional WirelessRecommended Training Through Cisco Learning Partners

Conducting Cisco Unified Wireless Site Survey

Implementing Advanced Cisco Unified Wireless Security

Implementing Cisco Unified Wireless Mobility Services

Implementing Cisco Unified Wireless Voice Networks

CCNA Wireless CCNA

CCNP Wireless

CCIE

Professional

Associate

Expert

www.cisco.com/go/certifications

You are encouraged to join the Cisco Certification Community, a discussion forum open to anyone holding a valid Cisco Career Certification (such as Cisco CCIE®, CCNA®, CCDA®, CCNP®, CCDP®, CCIP®, CCVP™, or CCSP®). It provides a gathering place for Cisco certified professionals to share questions, suggestions, and information about Cisco Career Certification programs and other certification-related topics. For more information, visit www.cisco.com/go/certifications.

www.CareerCert.info




Learner Introductions

� Your name

� Your company

� Job responsibilities

� Skills and knowledge

� Brief history

� Objective

Prepare to share this information:

� Your name

� Your company

� Your job responsibilities

� The prerequisite skills that you have

� A profile of your experience

� What you would like to learn from this course

www.CareerCert.info



www.CareerCert.info


Module 1

Implementation of QoS for Wireless Applications

Overview Deploying voice-over-WLAN (VoWLAN) is far more than just deploying wireless IP phones. The wireless network needs to be designed with voice in mind, with the appropriate cell size, throughput, and overlap. Another key element to a successful VoWLAN deployment is bandwidth management. When congestion occurs, voice quality degrades before becoming unacceptable. Wireless Call Admission Control (CAC) can be implemented on controllers to limit the number of voice devices competing for bandwidth in the cell, and another CAC can be used on the wired side of the network to take LAN and WAN bandwidth information into consideration.

Keep in mind that quality of service (QoS) is an end-to-end problem. Voice must be given a steady pace and consistent bandwidth all the way between both IP phones. QoS cannot be thought of as a wireless-only issue. The wireless side of the network is often seen as a weak area because of the half-duplex nature of the RF environment. It is true that special care must be taken to make sure that cells are built to provide the best usage efficiency for voice devices. Deploying VoWLAN still implies implementing a global QoS strategy, on both the wireless and the wired side. This module will show you what QoS mechanisms can be put in place on the wireless side and how they can be extended to the wired side. It will also give you example configurations and the best practices to build an efficient VoWLAN network.

Module Objectives Upon completing this module, you will be able to implement QoS for wireless applications. This ability includes being able to meet these objectives:

� Identify general considerations for wired and wireless QoS

� Describe wireless QoS deployment schemes

� Configure the WLC and WCS for QoS

� Configure the wired infrastructure for QoS

� Understand best practices for wireless QoS deployments

www.CareerCert.info


1-2 Implementing Cisco Unified Wireless Voice Networks (IUWVN) v1.0 © 2009 Cisco Systems, Inc.

www.CareerCert.info


Lesson 1

Identifying General Considerations for Wired and Wireless QoS

Overview The first step to deploy quality of service (QoS) is to understand what mechanisms it brings to

manage the shared network resources. QoS is, of course, different on a wired cable than in the

wireless space. This lesson will show you the different families of QoS mechanisms used on

LANs, WANs, and in the wireless space with the IEEE 802.11e protocol and the Wi-Fi

Multimedia (WMM) specification.

Objectives

Upon completing this lesson, you will be able to identify general considerations for end-to-end

QoS. This ability includes being able to meet these objectives:

Describe QoS

Describe classification and marking

Describe trust boundary

Describe congestion management

Describe congestion avoidance

Describe policing and shaping

Describe link efficiency mechanisms

Describe where each QoS feature is used

Describe congestion in the wireless space

Describe the 802.11 DCF mechanism

Describe 802.11e

Describe WMM

Describe 802.11e, 802.1p, and DSCP mapping

www.CareerCert.info



QoS Overview This topic describes QoS principles as they apply to wired and wireless traffic.

© 2009 Cisco Systems, Inc. All rights reserved. IUWVN v1.0—1-2

QoS Overview

Integrates data, voice, and video services into a single, packet-based infrastructure using IP

Provides differentiated service to selected network traffic

Services model includes best effort, integrated services, and differentiated services

Relies on several components: CAC, classification and marking, congestion avoidance, policing and shaping, congestion management queuing tools, and link efficiency

Network infrastructure has migrated from the separate circuit-switching networks for each

application to a common infrastructure based mainly on IP. The issue becomes the ability to

ensure different service levels for each type of application (voice, video, or data) on the

common network infrastructure.

QoS technologies refer to the set of tools and techniques used to manage network resources and

are considered the key enabling technology for network convergence. The objective of QoS

technologies is to make voice, video, and data convergence appear transparent to end users.

QoS technologies allow different types of traffic to contend inequitably for network resources.

Network devices can grant priority or preferential services to voice, video, and critical data

applications so that the quality of these strategic applications does not degrade to the point of

being unusable. Therefore, QoS is a critical, intrinsic element for successful network

convergence.

QoS tools are not only useful in protecting desirable traffic, but also in providing deferential

services to undesirable traffic such as the exponential propagation of worms. In other words,

QoS refers to the capability of a network to provide differentiated service to selected network

traffic over various network technologies. This implies that QoS is an end-to-end process: You

must apply it throughout the entire network so that critical traffic is prioritized from the source

to the destination point.

QoS can be applied using three different methods:

Best effort: This is simply the lack of any features or tools other than bandwidth and

standard devices buffers. When buffers are filled up, incoming packets are dropped,

regardless of the type of traffic they carry.

www.CareerCert.info


© 2009 Cisco Systems, Inc. Implementation of QoS for Wireless Applications 1-5

Integrated service: This is a pre-qualifying method that checks for resources before

sending packets into the network through out-of-band control messages. Resource

Reservation Protocol (RSVP) and H.323 are both examples of integrated-service QoS

methods. This method is sometimes referred to as ―hard‖ QoS. A limitation of processes

like RSVP is that they suppose an out-of-band communication mechanism between

endpoints, throughout the whole network. They also require each network component to be

compatible with RSVP and configured with consistent parameters. In an IP network where

different sections of the WAN belong to different organizations, RSVP is often challenging

to implement.

Differentiated service: This method, abbreviated DiffServ, is the most common method. It

uses in-band marking of the packets themselves. Each packet gets a level of priority

specified by a number. DiffServ-enabled devices read this number and apply a

prioritization accordingly. This technique is sometimes referred to as ―soft‖ QoS.

There are several ways of enhancing the use of the available bandwidth through QoS. They are

classified in different techniques:

Classification and marking: When traffic enters the network, you must recognize it to

prioritize it accordingly.

Congestion management queuing tools: When packets gather in the network buffer, you

can organize prioritization in several ways.

Congestion Avoidance: If you can anticipate congestion, you might be able to prevent it.

Call Admission Control (CAC) mechanism: If there is no space for a voice call, it might

be better not to allow it to occur.

Traffic Conditioners (Policing or Shaping): On a WAN, you can organize the available

bandwidth to limit the amount used by each type of traffic.

Link Efficiency: On some links, you can compress or simplify packets to send the same

amount of information with less bandwidth use.

www.CareerCert.info



Classification and Marking This topic describes the process of classification and marking.


Classification and Marking

The first step in implementing QoS is identifying the traffic on the network and determining

QoS requirements for the traffic. Define the service levels required by different traffic classes

in terms of response time and availability. What is the impact on business if a transaction is

delayed by two or three seconds? Can file transfers wait until the network is quiescent?

Voice traffic has extremely stringent QoS requirements. Voice traffic generally generates a

smooth demand on bandwidth and has minimal impact on other traffic as long as voice traffic is

managed. While voice packets are typically small (60 to 120 bytes), they cannot tolerate delay

or drops. The result of delays and drops are poor—and often unacceptable—voice quality.

After the majority of network traffic has been identified, network administrators use classes to

group traffic flows of the same type into categories. Because of its stringent QoS requirements,

voice traffic will usually exist in a class by itself. The other applications are grouped based on

their bandwidth and delay requirements.

A typical enterprise might define five traffic classes as follows:

Voice: Absolute priority for VoIP traffic

Mission critical: Small set of locally defined critical business applications

Transactional: Database access, transaction services, interactive traffic, preferred data

services

Best effort: Internet, email

Scavenger (less-than-best-effort): Napster, Kazaa, and other point-to-point applications

www.CareerCert.info



Marking a packet implies categorizing it by writing a category number on the header. You can

mark at Layer 2 or at Layer 3. On an Ethernet segment, Layer 2 is 802.3. When using 802.1Q,

802.3 headers add a 4-B field where information such as the VLAN tag can be written. This 4-

B section dedicates 3 bits for a class of service (CoS) value. Values can range from 000

(decimal 0) to 111 (decimal 7): the higher the number, the higher its priority. Voice is

commonly set as 5 or 6.

The 802.1Q 4-B section only exists in segments such as trunks. For non-Ethernet links or

segments where 802.1Q is not used, you can also mark at Layer 3. IP headers are preserved end

to end when IP packets are transported across a network, while data-link-layer headers are not

preserved. This means that the IP layer is the most logical place to mark packets for end-to-end

QoS. However, there are edge devices that can only mark frames at the data-link layer, and

many other network devices operate only at the data-link layer. To provide true end-to-end

QoS, the ability to map QoS marking between the data-link layer and the network layer is

essential.

At Layer 3, the IP header has a 1-B segment called type of service (ToS1). You can use this

field for QoS marking in two different ways:

It takes three bits to mark the QoS value. This is the IP precedence system. These bits are

the bits of highest weight (bits on the left). The values are the same as the Layer 2 CoS,

from 000 to 111 (decimal 0 to decimal 7). There can be a one-to-one match between the

Layer 2 tag and the Layer 3 tag if both are used in the same frame. You can use the

remaining five bits of the ToS field to mark the type of service, with bits giving

information about delay, throughput, reliability, and cost. They are usually left to zero. The

last bit is always unused and set to zero (must be zero [MBZ]).

1 Keep in mind that CoS is Layer 2, ToS is Layer 3. A way to remember it is to say that ToS is on Top, while CoS is

Closer to the cable.

www.CareerCert.info




Differentiated Services Code Point

Another system, differentiated services code point (DSCP) extends the number of

possibilities. DSCP uses the six bits on the left, thus creating 64 different QoS values. The

last two bits mark explicit congestion notification (ECN), thus informing the destination

point about congestion on the link.

DSCP is a bit more complex than IP precedence, as it further subdivides the 6 bits into

three sections. The first 3 bits (bits of highest weight) create a general priority. The next

two bits determine a drop probability; that is to say, compare the packet priority within the

same general class. The last bit, on the right, is usually set to 0 when DSCP is used in full.

The resulting tag is a combination of letters and numbered values, the letters representing

the general priority group and the numbers the value inside that group:

— When the general priority group (three bits on the left) is set to 000, the group is

called best-effort service; no specific priority is marked.

— When the general priority group ranges from 001 to 100 (001, 010, 011, and 100),

the traffic is said to belong to the Assured Forwarding (AF) class.

— When the general priority group is set to 101, traffic belongs to the Expedited

Forwarding (EF) class. This value is decimal 5, the same as the one used for voice

with IP precedence. In the EF class, for voice traffic, the drop probability bits are set

to 11, which gives a resulting tag of 101110, or 46. ―EF‖ or ―46‖ represent the same

priority value with DSCP.

The marking convention is a bit more difficult in the AF class: The higher the number in the

class, the higher the priority. Therefore, traffic tagged 011 will have a higher priority than

traffic tagged 001. The next two bits are expressing the drop probability: The higher this

number, the higher the probability for the packet to be dropped if congestion occurs and

packets have to be dropped. This means that among the three used values (01, 10, and 11), the

packet tagged 11 has more chances to be dropped than a packet tagged 01. Packet 00 is unused

and kept to represent ―no drop probability defined.‖

www.CareerCert.info



When using DSCP, you can see tags named AF31 or AF22. These tags mark the general

priority group, the number that follows the subclass (3 or 2 in this example), and the drop

probability (1 or 2 in this example). This is a naming convention. The value 31 or 22 itself, in

DSCP context, shows that these values are from the AF class, but AF is still commonly written

as well.

AF31 has a higher priority than AF22, as 3 (011) is higher than 2 (010). If a packet must be

dropped, AF22 will be dropped before AF31. Within one class, keep in mind the meaning of

―drop probability‖: AF31 has a higher priority than AF32 or AF33, because AF32 has a higher

drop probability than AF31, and AF33 an even higher drop probability. In other words, the

higher the number on the right, the lower the priority.

The following table summarizes the AF values.

AF Class Drop Probability DSCP Value

AF Class 1 Low 001 01 0

Medium 001 10 0

High 001 11 0


Medium 010 10 0

High 010 11 0


Medium 011 10 0

High 011 11 0


Medium 100 10 0

High 100 11 0

DSCP tends to be used more than IP precedence. Marking IP precedence can be seen as

marking just the DSCP main priority group without the drop probability. Some networking

devices mark at Layer 2, some mark at Layer 3, some use both markings. It is, therefore,

critical to have a consistent QoS policy throughout the enterprise network about which type of

traffic should receive which marking.

To preserve backward-compatibility with any IP precedence scheme, DiffServ has defined a

DSCP value in the form xxx000, where x is either 0 or 1. These DSCP values are called Class-

Selector Code Points (or CS). They represent a DSCP value without any drop probability. For

example, packets with a DSCP value of 110000 (the equivalent of the IP precedence-based

value of 110, which is IP precedence 6) will be marked CS6 and will have preferential

forwarding treatment (for scheduling, queuing, and so on), as compared to packets with a

DSCP value of 100000 (CS4, the equivalent of the IP precedence-based value of 100, or IP

precedence 4). Packets with a DSCP value of 100010 will not be marked ―CS,‖ because they

also contain a drop probability value. They belong to the AF class and are AF41. IP

precedence-only devices will still see them as 100 only, which is IP precedence 4, but DSCP

devices will know that there is more than just a CS value.

www.CareerCert.info




QoS Packet Marking Translations

DSCP Per Hop

BehaviorDSCP Tag Value

IP Precedence Tag

IP Precedence Label

CS7 111000 (56) 7 Network Control

CS6 110000 (48) 6 Internetwork Control

EF 101110 (46) — —

CS5 101000 (40) 5 Critical

AF41 100010 (34) — —

CS4 100000 (32) 4 Flash Override

AF31 011010 (26) — —

CS3 011000 (24) 3 Flash

AF21 010010 (18) — —

CS2 010000 (16) 2 Immediate

AF11 001010 (10) — —

CS1 001000 (08) 1 Priority

Default 000000 (00) 0 Routine

The table provides the IP precedence equivalent to the main DSCP classes. The CS classes

were built for partial backward compatibility with the IP precedence values, and a direct

translation is possible. The AF and EF classes create subcategories that are not available in the

IP precedence limited set of QoS values. An IP precedence-based system would be able to read

the first three bits and would assimilate any AF class to the CS category immediately below it.

IP precedence-based systems would be able to prioritize packets belonging to different classes,

thus reading that AF41 class should be given a higher priority than AF31. Because IP

precedence-based systems would be unable to interpret the Drop Probability section, they

would treat AF31 and AF33 packets as having the same priority level.

www.CareerCert.info



Trust Boundary This topic describes the concept of trust boundary.


Trust Boundary

Cisco QoS model assumes that the CoS carried in a frame may or may not be trusted by the network device.

For scalability, classification should be done as close to the edge as possible.

End hosts generally cannot be trusted to tag a packet priority correctly.

The outermost trusted devices represent the trust boundary.

1 and 2 are optimal, 3 is acceptable (if access switch cannot perform classification).

The concept of trust is important and integral to deploying QoS. The QoS tag can be set

anywhere in the network. End devices, if they are QoS-enabled, can tag their own traffic before

sending it. After the end devices have set CoS or ToS values, the switch has the option of

trusting them. If the switch trusts the values, it does not need to reclassify; if the switch does

not trust the values, then it must perform reclassification for the appropriate QoS.

The policy to trust or to re-mark is dictated by the type of devices and the environment. If a

switch has an IP phone connected to a port, it might be safe to trust the CoS or ToS value

received on that port. If a PC is connected behind the phone, there is a risk that some users

might try to give a high-priority tag to their personal traffic, and the administrator might decide

not to trust the received value and reclassify and mark all incoming traffic.

The notion of trusting or not trusting forms the basis for the trust boundary. Ideally,

classification should be done as close to the source as possible. If the end device is capable of

performing this function, the trust boundary for the network is at the end device. If the device is

not capable of performing this function, or the wiring closet switch does not trust the

classification done by the end device, the trust boundary might shift.

How this shift happens depends on the capabilities of the switch in the wiring closet. If the

switch can reclassify the packets, the trust boundary is in the wiring closet. If the switch cannot

perform this function, the task falls to other devices in the network, going toward the backbone.

If possible, try to avoid performing this function in the core of the network and, instead,

perform classification at the network edge.

www.CareerCert.info



If the network-edge switch is QoS able, re-marking is easy. For example, consider the campus

network containing IP telephony and host endpoints. Frames can be marked as important by

using link-layer CoS settings or the IP precedence or DSCP bits in the ToS and DiffServ field

in the IP version 4 (IPv4) header. Cisco IP phones can mark voice packets as high priority

using CoS as well as ToS. By default, the IP phone sends 802.1p-tagged packets with the CoS

and ToS set to a value of 5 for its voice packets. Because most PCs do not have an 802.1Q-

capable network interface card (NIC), they send packets untagged. This means that the frames

do not have an 802.1p field. Also, unless the applications running on the PC send packets with

a specific CoS value, this field is zero.

Note A special case exists where the TCP/IP stack in the PC has been modified to send all

packets with a ToS value other than zero. Typically this does not happen, and the ToS value

is zero.

Even if the PC is sending tagged frames with a specific CoS value, Cisco IP phones can zero

out this value before sending the frames to the switch. This is the default behavior. Voice

frames coming from the IP phone have a CoS of 5 and data frames coming from the PC have a

CoS of 0. If the DSCP is set, then the IP phone cannot re-mark the DSCP. This is why it is a

common practice on switches to trust the CoS coming from the phone, trust the CoS coming

from the PC if the phone re-marks it, or distrust the CoS coming from the PC if the phone does

not re-mark. The switch then applies a CoS-to-DSCP map to change the incoming DSCP values

of the packets according to the CoS policy.

www.CareerCert.info



Congestion Management This topic describes how QoS principles can be used for congestion management.


Congestion Management

Congestion management uses the marking on each packet to determine in which queue to place packets.

Congestion management uses FIFO or sophisticated queuing technologies such as priority queuing (PQ), custom queuing (CQ),weighted fair queuing (WFQ), class-based weighted fair queuing (CBWFQ), or low-latency queuing (LLQ) to prioritize important traffic.

After packets are marked for a QoS category, they are forwarded without change until

congestion occurs. Congestion management mechanisms (queuing algorithms) use the marking

on each packet to determine in which queue to place packets. Different queues are given

different treatment by the queuing algorithm, based on the class of packets in the queue.

Generally, queues with higher-priority packets receive preferential treatment.

Congestion management is implemented on all output interfaces in a QoS-enabled network by

using queuing mechanisms to manage the outflow of traffic. Each queuing algorithm solves a

specific network traffic problem and has a particular effect on network performance.

The Cisco IOS Software features for congestion management or queuing include FIFO, priority

queuing (PQ), custom queuing (CQ), weighted fair queuing (WFQ), class-based weighted fair

queuing (CBWFQ), and low latency queuing (LLQ).

www.CareerCert.info




Low Latency Queuing Example

Among the queuing mechanisms, FIFO is the most basic. When buffers are full, packets

entering the queue are dropped. PQ guarantees a strict service of a queue to the starvation of all

others. The prioritized queue is served until it has no packet left to send. CQ, on the other hand,

provides a guarantee for some level of service to each queue. Each queue can send a

configurable number of packets in turn. For example, four queues are defined, weighted 50, 20,

20, and 10. Fifty packets of the first queue are sent, then 20 packets of the second queue, then

20 packets of the third queue, then 10 packets of the first queue before going back to the first

queue.2 CQ can create delays for voice traffic.

WFQ is like dynamic CQ based on traffic flow types to allow fair bandwidth allocation to

large- and small-volume traffic. ―W,‖ for weighted, means awareness of ToS and service

accordingly. In other words, traffic that does not use a lot of bandwidth, or is infrequent, gets a

higher prioritization so that it does not drown under traffic that uses a lot of bandwidth.

WFQ is a great mechanism, but it serves all flows. Under conditions where the number of flows

or queues becomes large, even the higher-priority traffic, like voice, begins to experience

unacceptable delays and lack of bandwidth. This leads to CBWFQ, which is a variation of

WFQ where you can specify a minimum bandwidth per flow. You define classes and the

amount of bandwidth that they should get.

CBWFQ is ideal in many situations, but voice, by its nature, still needs a specific treatment.

LLQ allows you to add PQ on top of CBWFQ. With LLQ, delay-sensitive traffic such as voice

gets priority service followed by the rest of traffic that CBWFQ services. LLQ is the ideal

congestion mechanism for VoIP.

2 This example will help you understand the general principle: In a real router buffer, the calculation related to the

number of packets to send for each queue is also based on time units. The proportion described is respected, but if

during a time unit, only 10 packets can be sent, 5, 2, 2, and 1 packets of each queue, respectively, will be sent. The time

size is usually configurable.

www.CareerCert.info



Congestion Avoidance This topic describes how congestion can be partly avoided with weighted random early

detection (WRED) and Call Admission Control (CAC).


Congestion Avoidance and CAC

Congestion-avoidance mechanisms monitor network traffic loads in an effort to anticipate and

avoid congestion at common network bottlenecks. Congestion avoidance is achieved through

packet dropping.

Congestion avoidance mechanisms are typically implemented on output interfaces wherever a

high-speed link or set of links feeds into a lower-speed link (such as a LAN feeding into a

slower WAN link.) This ensures that the WAN is not instantly congested by LAN traffic.

WRED is a Cisco primary congestion-avoidance technique. WRED drops low-priority packets

rather than high-priority packets to avoid congestion. Low priority traffic is commonly TCP

based. TCP traffic is typically bursty. When all network devices are sending TCP streams to the

edge router at LAN speed, and the WAN link is far slower, the router buffers fill up and TCP

packets start being dropped massively. Many TCP sessions then simultaneously go into slow

start. Consequently, traffic temporarily slows down to the extreme, and then all flows slow-start

again. This activity creates a condition called global synchronization.

Global synchronization occurs as waves of congestion crest, only to be followed by troughs

during which the transmission link is not fully used. Global synchronization of TCP hosts can

occur because packets are dropped all at once. Global synchronization occurs when multiple

TCP hosts reduce their transmission rates in response to packet dropping. When congestion is

reduced, their transmission rates are increased. The most important point is that the waves of

transmission known as global synchronization result in significant link underutilization.

www.CareerCert.info



By anticipating this situation, WRED starts dropping packets in the incoming queue before the

buffers fill up. This allows the TCP mechanisms to resend and slow down before all sessions

are interrupted. WRED allows the bandwidth use to be close to its maximum without the global

synchronization effect.

WRED is not recommended for voice queues. Do not design a network to drop voice packets.

Voice streams are based on User Datagram Protocol (UDP) and cannot be re-sent. When

congestion occurs, the best mechanism to preserve bandwidth is to limit the number of calls.

This mechanism is the CAC, by which controllers can refuse extra calls in the wireless cell if

there is not enough bandwidth left. Controllers have an understanding only of the RF

environment conditions. This is why you can also implement CAC at the CUCM level, to take

into account the bandwidth of the wired side of the network and allow or forbid new calls

accordingly.

www.CareerCert.info



Policing and Shaping This topic describes policing and shaping.


Policing and Shaping

WRED is very efficient for TCP traffic and congestion avoidance, but it does not prioritize

traffic. It prevents the buffers from filling up. On edge routers that are connected on one side to

a fast LAN and on the other side to a slow WAN, WRED needs to be complemented by other

techniques to share the bandwidth of the slow side. Two common mechanisms are policing and

shaping.

Policing or shaping mechanisms are often used to condition traffic before transmitting traffic to

a network or receiving traffic from a network.

Policing is the ability to control bursts and conforming traffic to ensure that certain types of

traffic get a certain amount of bandwidth. Policing drops or marks packets when they reach

predefined limits. You can set policing mechanisms to first drop traffic classes that have lower

QoS priority markings.

You can use policing mechanisms at either input or output interfaces. These mechanisms

typically control the flow into a network device from a high-speed link by dropping excess low-

priority packets. A good example would be the use of policing by a service provider to throttle

a high-speed inflow from a customer that was in excess of the service agreement. In a TCP

environment, this policing would cause the sender to slow its packet transmission.

Shaping helps smooth out speed mismatches in the network and limits transmission rates.

www.CareerCert.info



Shaping mechanisms are used on output interfaces. These mechanisms typically limit the flow

from a high-speed link to a lower-speed link to ensure that the lower-speed link does not

become overrun with traffic. You could also use shaping to manage the flow of traffic at a point

in the network where multiple flows are aggregated. Service providers use shaping to manage

the flow of traffic to and from customers to ensure that the flows conform to service

agreements between the customer and provider.

A major difference between policing and shaping is that policing typically drops packets when

predefined limits are reached, whereas shaping simply buffers these packets. Later, if

bandwidth use gets lower, these buffered packets can be sent. Here again, shaping is not well

adapted for voice traffic, as resending voice packets later is pointless. Commonly, the voice

portion of the traffic is policed, while TCP traffic is shaped.

www.CareerCert.info



Link Efficiency Mechanisms This topic describes some link efficiency mechanisms used to maximize link usage on

relatively slower point-to-point links below E1 data rates.


Link Efficiency Mechanisms

Another way to mitigate the impact of congestion when an edge router connects a fast LAN to a

slower WAN is to use link efficiency mechanisms, to make better usage of the available

bandwidth on the WAN side. There are two major techniques:

Header compression

Link fragmentation and interleaving (LFI)

You can use these techniques in combination with other mechanisms such as shaping and

policing or congestion avoidance.

WAN links are often point-to-point on a given segment. They can take advantage of the fact

that voice traffic relies on the Real-Time Transport Protocol (RTP). RTP is a host-to-host

protocol that is used for carrying converged traffic (voice, but also video) over an IP network.

RTP provides end-to-end network transport functions intended for applications that transmit

real-time requirements.

A voice packet typically carries a 20-B IP header, an 8-B UDP header, and a 12-B RTP header

to carry 20 bytes of voice payload. By using compressed RTP (cRTP), the three headers with a

combined 40 bytes are compressed to 2 or 4 bytes, depending on whether the cyclic redundancy

check (CRC) is transmitted. This compression can dramatically improve the performance of a

link. Compression is used typically on WAN links between sites to improve bandwidth

efficiency. These links need to be point-to-point links. cRTP saves bandwidth, but adds delay

header compression on one end and decompression on the other end. Add this delay to the

overall forwarding delay to decide if you should implement compression. This efficiency

www.CareerCert.info



decreases as link speed increases. For this reason, compression is only recommended on links

of less than 2.048 Mb/s speed (E1 types of lines and slower).

LFI is another mechanism that is classified in the link efficiency category. Interactive traffic,

such as Telnet and VoIP, is susceptible to increased latency and jitter when the network

processes large packets, such as LAN-to-LAN FTP Telnet transfers traversing a WAN link.

This susceptibility increases as the traffic is queued on slower links.

LFI can reduce delay and jitter of smaller packets on slower-speed links by breaking up

preceding large packets and interleaving low-delay traffic packets with the resulting smaller

packets. Typically, you would use LFI on WAN links between sites to ensure minimal delay for

voice and video traffic.

With LFI, administrators decide the maximum size of each datagram to be sent on the WAN,

based on the speed of the link. This allows each packet to take no more than a predefined

amount of time to be sent. One downside of LFI is that after each large datagram is segmented

into smaller sections, each section has to be sent individually, thus creating the need for

individual Layer 3 and Layer 2 headers, which, in turn, can slow down the overall throughput

of the WAN side. For this reason, use LFI only when the speed of the link is so low that a large

packet can take too long to be transmitted; thereby, causing a voice conversation on the same

link to lose an acceptable quality. LFI is usually not recommended on links faster than 768

kb/s.

www.CareerCert.info



QoS in the Network This topic describes where in the network each type of QoS mechanism is used.


QoS Features: Applied Everywhere

Each QoS feature has its own purpose and fits into a global QoS strategy.

At the wireless cell level, you can use wireless QoS to prioritize voice traffic over the other

traffic. Wireless CAC is added to limit the number of concurrent calls in the cell. Between the

access point (AP) and the controller, all traffic is encapsulated into Control and Provisioning of

Wireless Access Points (CAPWAPs). The ToS and CoS visible on the outside header of the

packet can be trusted, because it has been configured on the controller and depends on the QoS

policy that is associated to each wireless LAN (WLAN).

After packets leave the controller to reach the first switch, you can re-mark to verify the level

of QoS that the wireless client originally requested. Use admission control also on Cisco

Unified Communications Manager for voice traffic. You can add policing to limit the allocated

bandwidth. For TCP traffic, congestion avoidance will optimize network resources utilization.

As you can see, most QoS tools are applied at the edge of the network at ingress to avoid

problems at the distribution or core of the network before packets reach this point.

You can apply some common tools such as congestion avoidance throughout the entire

network. If the traffic flow involves traversing from LAN-WAN-LAN, you might need

additional tools such as fragmentation and compression.

www.CareerCert.info



Congestion in Wireless Cells This topic describes congestion issues in wireless environments.


Congestion in the Wireless World

= 54-Mb/s link/radio

Congestion in wireless cells presents slightly different challenges than on a wired network. You

can configure several service set identifiers (SSIDs) for the same APs. Regardless of the

number of configured SSIDs, the AP has one or two radios. This physical limitation makes that

wireless clients that seem to be isolated in different SSIDs might still have to share the same RF

space. In this shared environment, clients do not have any awareness of each other’s traffic

requirements. Clients can detect that some other stations are sending in the cell, but cannot

analyze each other’s bandwidth needs. They compete to gain access to the wireless medium on

a per-packet basis. As clients share the same RF environment, collisions are likely to occur, and

you need a first mechanism to manage these collisions.

Depending on their position in the cell, clients can get up to 54 Mb/s in a classical 802.11a/g

network and up to 300 Mb/s in an 802.11n network. These values have to be understood as ―per

radio.‖ In either case, the available bandwidth in the wireless space is different from the

available bandwidth on the wired link through which the AP connects to the enterprise switch.

Congestion management has to occur to ensure that traffic coming from either side will not be

dropped due to congestion issues. The same phenomenon occurs when traffic coming from

many APs is sent to one controller. To take an extreme example, a Cisco Catalyst 6500 Series

Wireless Services Module (WiSM) with 8-Gbps link to the switch can manage up to 300

802.11n APs with two radios each. Congestion can occur at the controller port level.

www.CareerCert.info



Although congestion on the wired side of the network is an issue, the main preoccupation of a

wireless network designer is congestion in the wireless cell itself. In an infrastructure type of

wireless network, all packets coming from wireless clients need to be first sent to the AP before

being forwarded to another wireless client of the wired network. The bandwidth available in the

wireless cell usually dictates the AP bandwidth consumption on the wired side. An AP offering

24 Mb/s on the wireless side will usually consume the same bandwidth on the wired side,

simply because the link of lowest bandwidth, in this case the wireless space, dictates the overall

bandwidth consumption throughout the whole link.

www.CareerCert.info



802.11 DCF This topic describes the 802.11 Distributed Coordination Function (DCF).


Sending a Frame

.

.

Wireless devices are half duplex: they can send and receive, but cannot do both at the same

time. This is true for client adapters and for APs. Only one device can transmit at a time on a

channel in a given area. If two frames are sent at the same time, a collision occurs and both

frames have to be discarded. Therefore, each device has to send in turn.

No device centralizes the turns to decide who will send next. This commonly used wireless

medium access method is called DCF. The DCF coordination is distributed, which allows each

device to take care of itself. In some rare cases, the coordination can be done by the AP, which

is called Point Coordination Function (PCF).

To avoid collisions, the devices in the cell use Carrier Sense Multiple Access with Collision

Avoidance (CSMA/CA) as opposed to the 802.3 Ethernet method of CSMA/CD. When a

device needs to send a data frame, it starts by picking a random number. It will then count

down from that number.

www.CareerCert.info




Sending a Frame (Cont.)

.

While counting down, the station listens to the media. Every time it detects a signal, it stops the

countdown for the duration of the transmission. The total amount of time waited (backoff time

plus time waited during transmissions) is the ―contention window‖ (the time during which the

station refrains from sending). The machine does not need to keep listening to the detected

frame. The 802.11 header contains a duration field that expresses how long it will take to

transmit it. This duration, which is a reservation of the medium, is the Network Allocation

Vector (NAV). The station would add the NAV value to the countdown number, which it was

at when hearing the other transmission starting, and would carry on counting from the new total

instead of listening to the air.

www.CareerCert.info




Sending a Frame (Cont.)

.

.

My countdown reached 0.

When the counter reaches zero, the station will send its frame, assuming that the media is free.

This means that wireless devices have two ways of determining if the media is free:

The first is logical and based on the NAV that recalculates the time to wait (depending on

the other station’s signal) to avoid being ready to send when another device is occupying

the medium.

The second one is physical. When a station is ready to send, it listens to the media to verify

that nothing else is sending. This is Clear Channel Assessment (CCA). If the media is free,

it sends its wave.

If the transmission fails, the device will pick up a new random number, this time between 0 and

127 (0 and 255 the third time, then 0 to 511, and 0 to 1023 for all the following attempts).

Wireless devices cannot send and receive at the same time. This means that while sending, the

station has no idea if another machine is sending at the same time, so there is no way to know if

the message reached the recipient in good condition. Wireless networks rely on a system of

acknowledgements confirming that the frame was received; for each frame sent, the recipient

will return an acknowledgement frame.

www.CareerCert.info




After a Frame Is Sent

.

.

.

A mechanism is in place for a short interframe space (SIFS) to ensure that the

acknowledgements will always be sent with a higher priority than any other frame. Otherwise,

another station, having reached zero, might be transmitting before the acknowledgment (ACK)

is sent, and the first station might deduce that the frame was lost just because the ACK was

delayed.

This links to the notion of silence, or space, between frames, defined before. To avoid any late

reflections, there is always a moment of silence between frames. (The NAV actually reserves

the time needed by the frame and the subsequent silence.) This silence is the distributed

interframe space (DIFS), and is the normal silence time.

Instead of waiting for a full DIFS, the receiving station sends its ACK after a shorter amount of

time, called short interframe space (SIFS), to be sure to have priority over any other sender.3

When a station sends, it actually reserves the medium for the duration of its frame, a SIFS, and

the duration of the expected ACK.

The receiver reads the frame, waits a SIFS, and sends the ACK back, the duration of which will

be set to zero (it is an empty frame) to indicate the end of the transaction.

After the medium is free, any other station wishing to send and having its backoff timer at 0

will wait a DIFS and then transmit.

3 There is a third interframe space, PCF interframe space (PIFS), that is used during point coordination function (PCF);

its length is intermediate between SIFS and DIFS. There is also an extended interframe space (EIFS). As for durations,

PIFS = SIFS + slotTtime, and DIFS = SIFS +slotTtime + slotTtime.

www.CareerCert.info




QoS for Wireless: Interframe Spaces

Protocol SIFS (µs) Slot time (µs) PIFS (µs) DIFS (µs)

802.11b 10 20 30 50

802.11g 10 9 19 28

802.11a 16 9 25 34

After a station is ready to send, multiple timers help to determine, how long it should actually

wait before sending. A silence between frames always allows the multipath issues to clear. If

the frame to send has a high priority, the station will wait a SIFS. If the packet has a low

priority, it will wait a DIFS, the normal timer in Damage Cleanup Services (DCS) networks.

The value of these interframe spaces is as follows:

Short interframe space (SIFS)—the reference from which the others are built

PCF interframe space (PIFS)—SIFS + 1 x slot time

DCF interframe space (DIFS)—SIFS + 2 x slot time

The value in microseconds of the slot time and the SIFS depends on the protocol. The

following table gives you the values for the main protocol families:

Protocol SIFS (µs) Slot Time (µs) PIFS (µs) DIFS (µs)

802.11b 10 20 30 50

802.11g 10 9 19 28

802.11a 16 9 25 34

The interframe spaces (SIFS, PIFS, and DIFS) allow 802.11 to control which traffic gets first

access to the channel after carrier sense declares the channel free. Generally, 802.11

management frames and frames not expecting contention (a frame that is part of a sequence of

frames) use SIFS, and normal data frames use DIFS. If PCF were used, stations would wait a

PIFS instead of a DIFS before sending their frames. Protocol 802.11n introduces a new

interframe space, the reduced interframe space (RIFS), which is shorter than the SIFS value.

www.CareerCert.info




DCF Example

When a station using DCF is ready to send a frame, it goes through the following steps:

1. It generates a random backoff number between a value called minimum contention

window (CWmin) and a value called maximum contention window (CWmax). If it is a

first attempt to send a given frame, the station usually uses the CWmin value itself.

2. It listens to the medium. If the medium is busy, it waits until the channel is free for a

DIFS interval.

3. When the channel is free for a DIFS, the station begins to decrement the random

backoff number for every slot time that the channel remains free.

4. If the channel becomes busy, such as another station getting to 0 before your station,

the decrement stops and repeats Steps 2 through 4.

5. If the channel remains free until the random backoff number reaches 0, you can send

the frame.

This DCF mechanism is illustrated in the figure as follows:

1. Station A successfully sends a frame; three other stations also want to send frames but

must defer to Station A traffic.

2. After Station A completes the transmission, all the stations must still defer to the DIFS.

When the DIFS is complete, stations waiting to send a frame can begin to decrement

the backoff counter, once every slot time, and can send their frame.

3. The backoff counter of Station B reaches zero before Stations C and D, and therefore

Station B begins transmitting its frame.

www.CareerCert.info



4. When Station C and D detect that Station B is transmitting, they must stop

decrementing the backoff counters and defer until the frame is transmitted and a DIFS

has passed.

5. During the time that Station B is transmitting a frame, Station E receives a frame to

transmit, but because Station B is sending a frame, it must defer in the same manner as

Stations C and D.

6. When Station B completes transmission and the DIFS has passed, stations with frames

to send begin to decrement the backoff counters. In this case, the Station D backoff

counter reaches zero first, and it begins transmission of its frame.

7. The process continues as traffic arrives on different stations.

www.CareerCert.info




Frame Duration Values

Management Frames duration field typically set to ―0‖

Stations reserve the medium for duration of frame + SIFS + ACK

Fragmented frames reserve the medium until ACK of next fragment

This process requires that a station that needs to send a frame must wait for the previous station

to send its frame, and then wait at least a DIFS before starting to send. Protocol 802.11 is built

on a half-duplex environment, and each frame needs to be acknowledged. Upon receiving a

frame, the recipient waits a SIFS before sending an acknowledgement. This ensures that the

ACK is sent before any other station has a chance to start sending a new frame.

Each frame header contains a frame control section in which you can read a frame duration

value. This duration reserves the medium for the actual duration of the frame plus the length of

an SFIS and the acknowledgement.

This process is true for all data frames. When a frame is fragmented and the datagram sent in

several consecutive frames, the duration field of the initial frame encompasses the duration of

all the frames to follow, with each relevant SIFS and ACK. When RTS/CTS is used, the

medium is reserved for the duration of the RTS/CTS sequence and the frame, SIFS, and ACK

that follow.

The intent of this process is to inform the other stations that a frame is to be sent, so that the

other stations can increase their NAV by the duration value.4 They carry on counting down,

from the new increased value. They do not attempt to transmit during the time interval learned

through the duration information, as they know that the medium is busy.

4 The Network Allocation Vector is the total amount of time a station waits before sending. It is built from the backoff

timer, to which all the extra waiting times are added. To take an analogy from sports, if a soccer game is supposed to

last 90 minutes, 90 minutes is the backoff timer. Every time the clock is stopped, each break, increases the total duration

of the event. The total time is the NAV.

www.CareerCert.info




Contention Windows

Backoff timer is a slot number starting from aCWmin

Increases each time that the packet collides

Cannot exceed aCWmax

DCF uses a contention window (CW) to control the size of the random backoff. The contention

window is defined by two parameters:

aCWmin

aCWmax

The aCWmin and aCWmax are minimum and maximum contention windows values defined by

the protocol and confirmed in each device MIB. The random number used in the random

backoff is initially aCWmin. When the initial random backoff expires, the station tries to send

the frame.

If the station fails to receive an ACK for the frame, it deduces that the frame collided and

increments the retry counter. The increment is usually a fixed value that doubles for each new

attempt and is added to the aCWmin.

This doubling and addition to aCWmin continues until CW size equals aCWmax. All the

subsequent attempts will have the aCWmax value. The retries continue until the maximum

retries or time-to-live (TTL) is reached.

This process of doubling the backoff window is often referred to as a binary exponential

backoff and is illustrated in the figure, where the aCWmin is 31, and it increases up to the

aCWmax value of 1023.

www.CareerCert.info



You can see how collisions severely affect voice traffic. If too many stations and phones are in

the same cell, chances increase that two wireless clients get to 0 at the same time and send at

the same time, thus colliding. Both packets have to be resent, and both with increased delay.

This means that when packets collide in the cell, the overall throughput of the cell decreases

because packets are increasingly delayed before being sent. This protection mechanism is

efficient for the cell itself. If collisions occur because too many devices need to send at the

same time, resending them all after the same interval just reproduces the problem. Delaying

some of them reduces the collision risk.

For voice packets, though, this mechanism can result in the packet being delayed for more than

30 ms and, therefore, being dropped by the sending phone. Avoiding congestion by sizing the

number of devices is a key element of voice-over WLAN (VoWLAN) QoS.

Another important point is to avoid mixing fast packets with slow packets. A 200-B-long frame

at 1 Mb/s takes longer to send than a 2346-B-long frame sent at 54 Mb/s. During the sending

time, all other stations interrupt their countdown, and this process repeats itself as many times

as stations are sending before the phone reaches 0. This is why VoWLAN users should try to

avoid mixing data and voice clients, limit the number of active clients on a given channel, and

set the minimum speed to 24 Mb/s at the edge of the cell.

www.CareerCert.info




Priority with PCF

In an attempt to prioritize traffic in the wireless space, the 802.11 protocol defines an

alternative to DCF, named point coordination function (PCF). The AP is usually the point of

coordination and starts the PCF mode. The AP indicates that it starts using PCF mode by

sending a beacon with a duration field value set to 32768. All the stations understand that this

duration means that the AP is taking control of the transmission process.

From that moment, no station will individually take the initiative of sending a frame without

being instructed to do so by the AP. This period is the contention-free period (CFP), because

the stations need not refrain from sending by counting down, but merely need to wait for the

AP instructions. The normal DCF period is called, by contrast, the contention period (CP).

During the CFP, the AP polls the stations, asking each of them to send. The intent of the AP is

to poll only the stations that have urgent traffic to send, such as a wireless phone during a call;

thus giving it absolute priority over the other devices in the cell.

If the polled stations do not have any packets left to be sent, the AP can release the cell by

sending another beacon with a duration field set to 0: the cell goes back to DCF mode, until the

AP sends the next special beacon.

During a phone conversation, the cell can alternate many times between CP and CFP,

depending on the needs of individual stations in the cell. When the phone needs to send

buffered frames, the AP switches to PCF, then releases the cell when all packets are sent (the

user is listening instead of talking, or the phone buffer is emptied).

When stations authenticate and associate to the AP, they can signal if they are ―pollable,‖ that

is to say if they support PCF and want to be polled during the PCF period. The AP can still

decide to poll only some of the pollable stations and not all of them, depending on their

individual needs at the beginning of the CFP.

www.CareerCert.info



PCF still does not address a number of issues. It does not define how a phone should signal to

the AP that it has urgent traffic to send. If you send this message with the same contention

mechanism as a normal frame, the phone might as well send its voice packet directly, instead of

waiting during a countdown, just to signal that it needs to send packets without waiting. This is

an important consideration because this process is supposed to recur every time the phone has

new frames to send.

Also, if another phone needs to start a conversation flow during a contention period, it cannot

signal itself to the AP, because it has to wait for the AP to poll it. If the phone is a softphone on

a PC, PCF does not define any real QoS mechanism to classify and mark traffic. Such a PC

would be given high priority for voice traffic, but all other traffic sent by this PC would be

prioritized at the same time. PCF does not define any process to filter which traffic can be

considered urgent; this definition is supposed to be determined by other mechanisms.

For all these reasons, few, if any, vendors have implemented PCF, but this original attempt is

the foundation of a new definition of QoS for wireless known as 802.11e.

www.CareerCert.info



Describing the 802.11e Protocol This topic describes the 802.11e protocol.


Created in 2005 to create QoS for 802.11.

802.11e-enabled cells are backward-compatible with classical 802.11 cells.

802.11e creates HCF:

– EDCA

– HCCA

– General QoS features:

Blocks acknowledgements

APSD

NoACK frames

APs are HC, and cells are QBSS.

IEEE 802.11e

To compensate for the limitations of both DCF and PCF, which were not efficient enough to

accommodate for voice traffic, the IEEE 802.11 working group created a committee to work on

a protocol allowing QoS features specifically for the 802.11 devices. In 2005, this committee

published the 802.11e protocol and later integrated it into the 802.11-2007 general protocol.

In 802.11e, stations that support 802.11e are called QoS stations (QSTA). Devices that operate

in the cell while not supporting 802.11e themselves are non-QoS stations (nQSTA). This new

mode is called hybrid coordination function (HCF), because the cell brings QoS features while

allowing nQSTA to operate normally. This new mode enhances the performances of

compatible stations, but any 802.11 station can join the cell. The AP is called a hybrid

coordinator (HC) and its cell a QoS Basic Service Set (QBSS).

The 802.11e protocol brings several modifications of station behaviors to include prioritization

functions. These modifications can be grouped in three families:

Enhanced Distributed Channel Access (EDCA), which is close to DCF. EDCA performs

Enhanced Distributed Coordination Function (EDCF).

HCF Controlled Channel Access (HCCA), which is close to PCF.

General QoS features, which are general enhancements of station behaviors, such as block

acknowledgements (BAs), automatic power-save delivery (APSD) or frames of NoACK

type (not requiring acknowledgement).

www.CareerCert.info




802.11e Traffic Categories

802.11e has four ACs, each having two TCs or UPs.

Priority 802.1p Priority 802.11e DesignationAccess

Category Designation

Highest 7 NCAC_VO Voice

6 VO

5 VIAC_VI Video

4 CL

3 EEAC_BE Best Effort

0 BE

2 -AC_BK Background

Lowest 1 BK

A major improvement created with 802.11e is the possibility for individual stations to create

internal queues. With DCF, when a station wants to send a packet, it picks up a backoff number

and counts down from it. With 802.11e, the same station starts by classifying the packet to send

into a category. Up to eight categories can exist. This behavior is very close to the classification

and marking activity performed on internetworking devices in classical QoS.

The 802.11e protocol actually created equivalents to the 802.1p CoS categories, thus creating

queues 0 to 7, 7 being the most urgent. These categories are called traffic categories (TCs).

They are grouped into four access categories (ACs), each AC having two TCs or user priorities

(UPs). This allows tuning priority if two streams are of the same category within one station.

As you see in the figure, each AC has an optional designation name representing the type of

traffic that it is supposed to carry.5 The TCs are usually abbreviated as follows:

Access category background is AC_BK. It contains the traffic category background (BK). The

UP with 802.1P equivalent of 2 is unused and kept for spare.

Access category best effort is AC_BE. It contains the traffic categories best effort (BE) and

excellent effort (EE).

Access category video is AC_VI. It contains the traffic categories controller load (CL) and

video (VI).

Access category voice is AC_VO. It contains the traffic categories voice (VO) and network

control (NC).

5 The IEEE 802.11e classification is different from the classifications recommended and used in the Cisco network,

which are based on IETF recommendations.

www.CareerCert.info




802.11e HCCA

HCCA is close to PCF, with enhanced features such as free starts and free ends.

The HCCA period is called a superframe.

HCCA occurs within EDCA.

The 802.11e standard defines two modes, Enhanced Distributed Channel Access (EDCA) and

hybrid coordination function (HCF) controlled channel access (HCCA), which is close to PCF.

In this mode, a random access protocol that allows fast collision resolution is defined. The AP

requires information that has to be updated by the stations from time to time. This information

contains the identity of stations that need to be polled, at which times, and for which duration.

The controlled contention mechanism (CCM) allows stations to request the allocation of

―transmission opportunities‖ (TXOPs) by sending resource requests. Each instance of

controlled contention occurs during the controlled contention interval, which starts when the

AP sends a specific frame.

The main difference with PCF is that HCCA allows the CFP period to start at any time. With

PCF, the AP would send a beacon to start the CFP, and another one to end it. With HCCA, any

station can require a CFP from the AP, and if the station does not need it anymore, the AP ends

it immediately.

When the AP starts the contention period, the initial control frame defines a number of

controlled contention opportunities (that is, short intervals separated by SIFS) and a filtering

mask containing the traffic categories for which resource reservation can be requested. Each

station with queued traffic for a traffic category matching the mask chooses one opportunity

interval and transmits a resource request frame. This request identifies the traffic to be sent and

its duration or size. The AP acknowledges the reception of request by generating a control

frame with a feedback field so that the requesting stations can detect collisions during

controlled contention.

The result is a ―superframe,‖ where a station can send a burst of several packets acknowledged

by the AP. The total duration of the series must be less than the TXOP given to the station.

www.CareerCert.info



Describing WMM Implementations This topic describes wireless multimedia implementations.


Wi-Fi Multimedia

WMM is a Wi-Fi Alliance certification on partial implementation of 802.11e.

Ensures compatibility between vendors implementing the same 802.11e features.

Eight traffic categories (TCs) become four queues.

Does not retain HCF Controlled Channel Access (HCCA).

Maintains EDCA with Block Acknowledgement (BA) and NoACK frames.

Adds WMM Power Save certification to integrate Automatic Power Save Delivery (APSD).

The Wi-Fi Alliance is a certification organization that ensures compatibility between 802.11

devices built by various vendors. This effort was very close in its spirit to the Wi-Fi Protected

Access (WPA) certification, based on the 802.11i draft. During the 802.11i protocol

development, the WiFi Alliance released a certification based on the first draft, and called

WPA.

Similarly for QoS, in mid 2003, the IEEE released a first certification proposal known as

Wireless Multimedia Extensions, based on an early draft of QoS implementation for wireless.

Shortly after that, the IEEE released a new draft for the future 802.11e protocol, and the Wi-Fi

Alliance redrew the certification to implement part of this new draft. This new effort also

removed features that were not retained by the 80211e draft, even though some vendors wanted

to retain these features, and included them in the WME proposal.

This new certification, based on the 802.11e draft, is Wi-Fi Multimedia (WMM). This

certification applies the EDCA part of 802.11e, with only four queues instead of eight, one

traffic category in each access category. This does not mean that eight queues are forbidden or

that HCCA cannot be implemented. To be certified for WMM, a wireless device must be able

to support four queues and EDCA, and be able to communicate evenly with other devices

certified the same way. Extra features are possible but are not part of the certification.

In 2006, the Wi-Fi Alliance released a new version of the WMM certification, called WMM

Power Save, which adds automatic power-save delivery (APSD) to the WMM certification.

APSD is a power-saving feature present in the final 802.11e protocol, but which was not

described in the draft from which the first WMM certification was derived. To be fully

classified as WMM, a device has to support WMM with APSD.

www.CareerCert.info




WMM—EDCA Queues

AIFS = arbitrated interframe space. CW = contention window. PF = persistence factor

When a WMM-enabled station needs to send a packet, it starts by classifying the packet into

one of the four categories. After the packet is classified, a backoff timer is picked up. The

backoff timer is again a random value, just like in DCF, but the contention window minimum

value depends on the packet, or traffic category (TC), instead of being the default aCWmin

defined by the MIB.

For this reason, the contention window minimum value is called CWmin instead of aCWmin. If

the packet belongs to an urgent category, CWmin is lower than if the packet is best effort. For

example, if the station has to send a voice packet, the backoff timer could be 3 slots. If the

packet were a normal (best-effort) packet, the backoff timer would be 31 slots. With this

system, urgent packets are more likely to pick up lower numbers than less urgent packets.

A single station can implement up to eight categories as per 802.11e, but usually implements

only four to adhere to the WMM requirements, each of them associated with one transmission

queue. These queues create virtual stations inside a station, with QoS parameters that determine

their priorities. Each has a backoff timer, and their countdown runs in parallel. The first to

reach zero is sent. If the counters of two or more parallel queues in a single station reach zero at

the same time, a scheduler inside the station avoids the virtual collision by giving priority to the

queue with the highest QoS or urgency level.

It is possible that the transmitted frame can collide at the wireless medium with a frame

transmitted by other stations, but more urgent traffic is prioritized inside the stations

themselves.

www.CareerCert.info




WMM—Interframe Space

In DCF, after a backoff timer reaches 0, the station listens to the media, and, if the media is

free, waits another DIFS before sending. In EDCA, the station waits what is called an AIFS

(arbitrated interframe space). The AIFS is defined as having to be at least equal to DIFS.

In other words, QoS-enabled stations (QSTAs) are already prioritized by having the possibility

to use a shorter backoff timer for time-sensitive traffic. When they reach zero, the non-QoS

stations (nQSTAs) must have a chance to send. So AIFS, the interframe space for QSTAs, must

be the same as, if not larger than, DIFS, the interframe space for nQSTAs.

Outside of this rule, the AIFS value can be fixed or variable. Its variation can depend upon the

TC or other parameters such as the packet loss rate of the wireless environment. In other words,

it is possible to have a shorter AIFS for urgent traffic than for default traffic.

Suppose that two QSTAs get to zero at the same time, one after having counted down from 46

for an FTP packet; the other after having counted down from 12 for a voice packet. Without

QoS, they would both wait a DIFS and start sending at the same instant, resulting in a

collision.6 With different AIFS, the station that was attempting to send a voice packet would

wait a shorter amount of time and would start sending before the station that was sending FTP

traffic. With this system, voice traffic is not only prioritized inside the station queues, but also

between stations.

6 In reality, it is very unlikely that they would get to 0 exactly at the same time. This situation is more likely to occur

when both stations get to 0 at slightly different moments and both have to wait because another station starts sending

before the end of the DIFS. Both stations wait for the frame to finish being sent and wait the same DIFS before sending

at the same time and colliding.

www.CareerCert.info



Another key feature of 802.11e WMM is the transmission opportunity (TXOP). A TXOP is an

interval of time when a station has the right to initiate transmissions, defined by a starting time

and a maximum duration. The duration of an EDCF-TXOP is limited by a cell-wide TXOP

limit that is distributed in beacon frames. If a frame is too large to transmit in a single TXOP,

you should fragment it into smaller frames.

The use of TXOPs reduces the problem of stations gaining a large amount of channel time by

sending large frames at low data rate. The AP publishes the available bandwidth in beacons,

and this bandwidth affects the TXOP value. The clients can check the available bandwidth

before adding more traffic in the network that cannot be entertained.

Stations can also interact dynamically with the AP by informing the cell about the TXOPs for

the next period by sending the specifications of the traffic that they want to send. If this traffic

is compatible with the TXOPs for this cell, the AP can reserve the corresponding bandwidth for

that client.

www.CareerCert.info




EDCA Example

Voice traffic is queued in TC 6, while data traffic is queued in TC 0

For voice, AIFS = 2 slots and CWmin = 3 slots

For best effort, AIFS = 3 slots and CWmin = 5 slots

The figure illustrates the EDCA process.

While Station X is transmitting its frame, three other stations determine that they must send a

frame. Each station defers because a frame is already being transmitted, and each station

generates a random backoff.

Because the voice station has a traffic classification of voice, it has an arbitrated interframe

space (AIFS) of two, and uses an initial CWmin of three. Voice traffic must defer the

countdown of its random backoff for two slot times and has a short random-backoff value.

Best effort has an AIFS of three and a longer random-backoff time, because its CWmin value is

five.

Voice has the shortest random-backoff time and, therefore, starts transmitting first. When voice

starts transmitting, all other stations defer.

After the voice station finishes transmitting, all stations wait their AIFS, and then begin to

decrement the random-backoff counters again.

Best effort then completes decrementing its random-backoff counter and begins transmission.

All other stations defer. This can happen even though there might be a voice station waiting to

transmit. This shows that best-effort traffic is not starved by voice traffic because the random-

backoff decrementing process eventually brings the best-effort backoff down to similar sizes as

high-priority traffic. On occasion, the random process might generate a small random-backoff

number for best-effort traffic.

www.CareerCert.info




WMM CW Values

Access Category

CWmin CWmax AIFS TXOP Limit (802.11b)

TXOP Limit (802.11a/g)

AC_BK aCWmin aCWmax 7 0 0

AC_BE aCWmin 4*(aCWmin+1)-1 3 0 0

AC_VI (aCWmin+1)/2-1 aCWmin 2 6.016 ms 3.008 ms

AC_VO (aCWmin+1)/4-1 (aCWmin+1)/2-1 2 3.264 ms 1.504 ms

Protocol 802.11e does not define the CWmin value. The AP transmits the CWmin values in its

beacons in an EDCA parameter set information element. Vendors have difficulty determining

which value of the CWmin for each category is best for both QSTAs and nQSTAs to be fairly

treated in the same cell. This setting depends very much on the specific needs of the QSTAs. A

Cisco controller, on the other hand, lets you change the EDCA parameters for 802.11a and

802.11b/g to provide best performance for all devices that are present in the cell.

Another efficient mechanism is the calculation of CWmin and CWmax. When collision occurs

in DCF, the station picks up a new backoff timer, which is an increment to the original backoff

timer value. If the second packet fails to get an acknowledgement, the same process recurs. For

example, if the first backoff timer was 31 and the increment 8, the second will be 39 (31+8), the

third 47 (31+16), the fourth 43 (31+32), and so on, the maximum aCWmax being 1023.

A negative effect of this mechanism is interference occurring for a short time. Several

consecutive packets sent by different stations fail to reach their destination, and all stations

increase their backoff timer in a large proportion, because doubling the previous increment

value is the only option. When the interference stops, the cell can be idle for a long time7 before

a station gets to 0.

With EDCA and the four-queue system, each queue defines a CWmax. A lower CWmax rather

than best-effort traffic can affect urgent traffic. The new value picked up when no ACK is

received is not necessarily twice the previous increment value either. An algorithm determines

the new backoff timer based on a value called the persistence factor (PF). The formula is as

follows:

NewCW = ((OldCW – 1) x PF) -1.

NewCW must be less than CWmax.

7 ―Long time‖ is a relative value here, as slots are in microseconds. Long time is at the scale of the 802.11 stations; not

at the human scale.

www.CareerCert.info



By affecting a different PF value for each queue, it is possible to retransmit faster critical traffic

and to delay less-time-sensitive traffic.

With the implementation of the WMM certification, some values such as aCWmin and

aCWmax had to be fixed among vendors. The following tables summarized the common values

used by Cisco and many other vendors:

Table 1 WMM Client Parameters

AC CWmin CWmax AIFS

TXOP Limit

(802.11b

TXOP Limit

(802.11a/g)





Table 2 WMM AP Parameters

Access

Category aCWmin aCWmax AIFS

TXOP Limit

(802.11b

TXOP Limit

(802.11a/g)





These values mean that, for background traffic coming from an AP, the CWmin is the default

aCWmin for the spectrum.8 The aCWmax is the maximum allowed for the spectrum.9 The

AIFS is 7 times a DIFS, and only one frame can be sent per TXOP.

For voice traffic, still taking the AP as an example, CWmin is (aCWmin+1)/4-1, which is 7 for

802.11b and 3 for 802.11g and 802.11a, 7 being allowed for 802.11g for backward

compatibility with 802.11b. Voice packets will get a number statistically closer to zero than

background packets.

If they have to be resent, the maximum slot number, CWmax, is (aCWmin+1)/2 -1, which is 15

for 802.11b and 7 for 802.11a and 802.11g. The AIFS is twice the DIFS. This seems to be a lot

more, but given the shorter slot time, this extended time is necessary to give to the nQSTAs a

chance to send their packets from time to time when a WMM voice device is in the cell. The

TXOP is expressed in milliseconds, which allows the stations to send several consecutive

frames with one block acknowledgement if they fit within this interval.

8 For 802.11b, the aCWmin defined by the protocol is 31, which means that a station needing to send a frame for the

first time picks up 31 slots. For 802.11a, aCWmin is 15. For 802.11g, aCWmin is 15, but 31 is also supported for

backward compatibility with 802.11b stations. 9 A station cannot pick up a number larger than 1023. This is true for 802.11b, 802.11g, and 802.11a.

www.CareerCert.info




QoS BSS Information Element

Sent by WMM APs in beacons and probe responses

Helps clients decide which AP to associate or roam to

No real interaction between client and AP

A key element of QoS in the wireless cell is the ability of the stations and the APs to make

informed decisions based on the available bandwidth and the competition among stations to

gain access to the wireless medium. There are two main ways for the stations to exchange

information about traffic conditions with the AP: QBSS and traffic specification (TSpec).

QoS Basic Service Set Information Element (QBSS IE) is a new element in the AP beacons and

probes responses when WMM is in place. It provides information to stations to help them make

decisions on where to associate and when to roam to another AP. This Information Element

contains three fields that inform about the station population and bandwidth consumption:

Station Count: This field is a simple 16-bit value that indicates the number of stations

associated to the AP. This field mentions all the stations; it does not distinguish between

idle stations and stations that are actively sending and receiving. It is still a good indicator

of the ―load potential‖ of the cell. An AP with more stations will be more likely to present a

heavy load level than an AP with only a few associated stations.

Channel Utilization: This field indicates the percentage of time, on average, that the AP

sensed the medium as busy. Depending on vendor implementation, the busy value only

takes into consideration the 802.11 frames (logical carrier sensing) or all RF activity

affecting transmissions (physical carrier sensing). The Channel Utilization is a value

ranging from 0 to 255 (0 = no activity, 255 = 100 percent utilization).

Available Admission Capacity: This 16-bit field indicates the number of 32-microsecond

units that were available over the last second. This gives an indication of the available

bandwidth on the AP over the last second. This field does not offer any guarantee about the

future, but is an indicator to complement the other two values about the potential to add

extra frames to the cell.

www.CareerCert.info




Used by WMM client to signal traffic requirements to the AP

Contains up to 16 fields describing all aspects of traffic to send

Can be a request to use a traffic category (TC) or access category (AC) in EDCA with or without specific TSpec request packets:

– Add TSpec (ADDTS) action frame at the beginning of the specific flow

– Reassociation to the AP mentioning the TSpec requirements

AP can accept or refuse

More interactive and detailed than QBSS IE

Traffic Specification (TSpec)

Stations can use QBSS IE to evaluate which AP offers the best availability, but QBSS does not

offer any direct interaction between the stations and the AP. After a station decides that an AP

offers the best available bandwidth and joins the cell, a second mechanism would be needed so

that the AP is made aware of the station bandwidth requirements and reserves the

corresponding bandwidth.

TSpec allows a WMM client to signal its traffic requirements to the AP. In the 802.11e MAC

definition, two mechanisms provide prioritized access. These are the contention-based EDCA

option and the controlled access option provided by the transmit opportunity (TXOP). When

describing TSpec features where a client can specify its traffic characteristics, you might

assume that this would automatically result in the use of the controlled access mechanism and

grant the client a specific TXOP to match the TSpec request.

However, this does not have to be the case; you can use a TSpec request to control the use of

the various access categories (ACs) in EDCA. Before a client can send traffic of a certain

priority type, it must have requested to do so via the TSpec mechanism. For example, a WLAN

client device wanting to use the voice AC must first make a request for use of that AC. Whether

or not TSpec has to be used for voice and video is generally configurable. Best-effort and

background ACs are normally used without TSpec control.

Note Unlike the Cisco Unified Wireless IP Phone7921G, which does have support for TSpec, the

Wireless IP Phone 7920 WVoIP handset does not support TSpec admission control. It uses

the QBSS load information element that is sent in the AP beacons to learn the load of the

AP. The QBSS load information element is part of the 802.11e protocol and was

implemented by Cisco into the Wireless IP Phone 7920 as a vendor pre-protocol feature.

The Add Traffic Stream (ADDTS) function is how a WLAN client performs an admission

request to an AP. Signaling its TSpec request to the AP, an admission request is in one of two

forms:

www.CareerCert.info



ADDTS action frame: This happens when a phone call is originated or terminated by a

client that is associated to the AP. The ADDTS contains TSpec and might contain a traffic

stream rate set (TSRS) IE (Cisco Compatible Extensions v4 clients).

Association and re-association message: The association message might contain one or

more TSpecs and one TSRS IE if the STA wants to establish the traffic stream as part of

the association. The re-association message might contain one or more TSpecs and one

TSRS IE if a STA roams to another AP.

The ADDTS contains the TSpec element that describes the traffic request. Apart from key data

describing the traffic requirements, such as data rates and frame sizes, the TSpec element also

tells the AP the minimum physical rate that the client device will use. This element allows the

calculation of how much time that station can potentially consume in sending and receiving in

this TSpec, and, therefore, allowing the AP to calculate whether it has the resources to meet the

TSpec. TSpec admission control is used by the WLAN client (target clients are VoIP handsets)

when a call is initiated and during a roam request. During a roam, the TSpec request is

appended to the re-association request.

With the information contained in the TSpec messages sent from a station, the AP can evaluate

the bandwidth requirements of this client and accept or refuse the client traffic. When the traffic

is accepted, the AP deduces the client requirements from the available bandwidth and uses this

information to determine if it will accept or refuse other TSpec requests from other clients. If

the AP refuses the client traffic, the client cannot send the frames into the requested traffic

category. The client can then choose to try another category or cancel its traffic. The effect of

this last action is easily visible on a wireless VoIP phone: the phone associates to the cell, but a

network busy sign appears when the phone tries to place a call.

www.CareerCert.info




Power Save Feature: PS-Poll vs. U-APSD

Traditional 802.11 networks use PS-Poll.

Networks with WMM capability use the more efficient U-APSD.

WMM and 802.11e also enhance the way stations can save power. A wireless phone is by

essence not connected to a power socket and battery duration is an important issue. Most

wireless phones tend to save power by going to sleep mode as often as possible. Two

mechanisms are possible:

With a traditional 802.11 network, the phone sends a null function frame with the ―sleep‖

bit set to 1 to the AP. The AP knows that the phone is sleeping and keeps the frames that it

might receive in its buffer. The phone wakes up for each beacon and listens to the Traffic

Indication Map (TIM) field at the end of the beacon where are listed all the devices for

which the AP has buffered traffic. If the phone sees its identifier listed, it sends a PS-Poll

(power-save poll) message to the AP by which it informs the AP that it is awake and asks

for the buffered packets. Because this PS-Poll is a normal frame, the phone uses the normal

process of backoff timer and countdown before sending it. The AP follows the same

process to forward the first packet to the phone. The phone acknowledges it, waits for a

normal backoff timer, then sends a new frame to ask for the next packet. The process is

repeated as long as there are remaining buffered packets on the AP.

With WMM, another mechanism, called unscheduled automatic power-save delivery (U-

APSD) is in use. With APSD, a station gets to sleeping mode as with traditional 802.11 by

sending a null frame where a ―sleep bit‖ is set to 1. Any application within the station can

wake it up (for example, when a call is initiated), and the station starts sending. The AP

learns from the packets that the station does is no longer asleep and sends packets to it as if

it had never fallen asleep. The station can still sleep until the next AP beacon and send a

frame (any frame) to the AP to inform it that it is awake. The AP can then send the buffered

frames without the need to be polled. When the phone receives the first buffered packet, the

acknowledgement frame contains a ―more‖ field, which makes the AP send immediately

the next frame. This burst system allows the buffer to be emptied in a very short delay, with

far less impact on the cell than the traditional PS-Poll system. Whenever possible, that is if

both the AP and the wireless phone have WMM capability, U-APSD is preferred over the

traditional sleep mode feature.

www.CareerCert.info




S-APSD vs. U-APSD

U-APSD is actually one of the two possible power-save enhancement mechanisms in 802.11e

and WMM. The other is scheduled automatic power-save delivery (S-APSD). S-APSD is

available in both EDCA and HCCA modes, whereas U-APSD is only available in EDCA. U-

APSD works best for variable bit rate traffic, while S-APSD is more efficient for moderate to

heavily loaded networks with predictable traffic.

For both mechanisms, the time during which a station is awake and receiving traffic is the

service period (SP). The SP can be started by the station or the AP, depending on the power-

save mechanism and is ended by a frame that contains an end of service period (EOSP) flag in

its QoS field.

With S-APSD, periods during which stations are sleeping are scheduled. When a station sends

an ADDTS to send its traffic specifications to the AP, it can indicate in the APSD and Schedule

subfields (both set to 1) that it wants to use S-APSD. The AP, upon accepting the requesting

traffic, will return service start time (SST) and a service interval (SI) values that will indicate

when the station will have to wake up for the first time (SST) and how often the station should

wake up afterward to receive the traffic that was buffered at the AP.

The AP is responsible for calculating the SST and SI values based on the traffic specifications.

The station will follow the AP instructions and wake up at the predefined intervals. With this

system, there is no trigger mechanism. The station wakes up in due time and waits. The AP

knows that the station should have awakened and started sending the buffered traffic.

The station stays awake until it receives a frame with the EOSP bit set to 1. It then falls back

asleep until the next SI. At any time during the frame exchange, the AP can update the SI to

adapt to the received traffic specific volume. You can define the SST and SI values on a per-

AC basis. During its sleeping period, a station can always wake up and start sending traffic

without any specific notice.

www.CareerCert.info



With U-APSD, sleep periods are freely determined by each station. An exchange with the AP

does not determine scheduling. At any time, a station can send to the AP a QoS null frame with

the power management bit set to 1 to start dozing. The station can wake up at any time and

send a trigger frame to inform the AP about its awakened state. This frame can be any frame.

Upon receiving this trigger frame, the AP starts sending the buffered frames without requiring

further information from the awakened station. The last frame contains the EOSP bit. If the AP

does not have any buffered frames, it responds to the trigger frame with a QoS null frame

containing the EOSP bit set to 1. Immediately, the station can start a new sleeping period.

www.CareerCert.info




APSD and Traffic Categories

APSD can coexist with PS-Poll on the same cell or in the same station.

APSD can be set on a per-traffic category basis.

Another strength of APSD, in both U-APSD and S-APSD modes, is that traffic belongs to

traffic categories (TCs). TCs are grouped in pairs to access categories (AC). The power-save

behavior can then be decided depending on the AC. For example, data traffic can obey the

classical power-save rule, while voice traffic can awaken the station in mid-sleep and trigger a

faster exchange with the AP. Not having to wait for the next beacon reduces latency. This

allows a voice application to doze several times during a beacon interval, thus saving power

while still providing low-latency service.

APSD, which was added to the WMM certification in 2006, is a general feature that is

described in 802.11e. It can occur in both EDCA and HCCA. All wireless devices certified as

WMM after 2006 must support APSD. Some devices might have been certified before 2006

and might not support APSD while being WMM certified. This is why it is common to see the

APSD support as a configurable feature in wireless clients (connecting to potentially non-

APSD supporting APs) or infrastructure devices (connecting potentially non-APSD supporting

clients).

www.CareerCert.info




802.11e Block Acknowledgement

BA allows sending several frames in a burst with a global ACK.

Global ACK marks which frames were received and missed within the block.

BA is negotiated with ADDBA and delete block acknowledgement (DelBA) frames marking the beginning and the end of the BA.

In DCF and PCF, each frame must be acknowledged. Although an ACK does not take long to

send, it reduces the overall throughput of the cell. To improve on this mechanism, 802.11e and

WMM create the notion of block acknowledgement (BA). The BA mechanism aggregates

several acknowledgments into one frame.

To send several frames, a station first checks the TXOP to determine how many frames it can

send within the TXOP. It then sends an add block acknowledgement (ADDBA) request to the

recipient. This request also defines the traffic category (TC) for which the request is made. The

receiving QSTA, which is the intended peer, has the option of accepting or rejecting the

request. When the receiving QSTA accepts, then a BA agreement exists between the originator

and recipient. When the receiving QSTA accepts, it indicates the type of BA and the number of

buffers that it will allocate for the support of this block. The BA negotiation allows the QSTA

to accept blocks only for some types of traffic.

After the block is sent, the emitter sends a last frame that contains a BA request. The receiving

station just needs to send an acknowledgement frame (BlockACK) to validate the entire block.

If the sender does not need to send another block, it sends a delete block acknowledgement

(DelBA) request, which is simply acknowledged by the receiver. Several blocks can be sent if

one block is larger than the TXOP. After one block is sent, the sender waits for the next TXOP

without deleting it.

The BlockACK frame contains the list of the frames received in the block. If one frame is

missing, the sender merely resends that specific frame without having to resend the entire

block. This process increases the efficiency of the cell by reducing the time wasted to

acknowledge frames in cells where the RF environment is good.

To enhance this behavior further, 802.11e and WMM define some frames that need not be

acknowledged, such as multicast frames. These frames are sent with a NoAck bit set to one,

which informs receivers that acknowledgement is unnecessary. If the transmission fails, the

frame is not re-sent. A typical usage of this feature is a keepalive message sent at regular

intervals. Losing one message is not critical and acknowledging each of them is unnecessary.

www.CareerCert.info



The 802.11n protocol further enhances the BA mechanism by allowing several frames to be

aggregated into a larger frame, which can contain one single physical header (MAC Service

Data Unit [MSDU] aggregation) or be an aggregate of several entire frames (Message Protocol

Data Unit [MPDU] aggregation). You can also send frames in burst, as with the 802.11e

mechanism.

In all cases, you can use a single BA frame to determine which frames were received. The

802.11 BA basically acknowledges each frame within one large ACK frame. You can view this

as an aggregated ACK, within which each frame is acknowledged individually. The 802.11n

BA is shorter than the 802.11e BA (8 bytes instead of 128 bytes) and, therefore, is more

efficient as far as bandwidth conservation is concerned.

www.CareerCert.info




WMM Frame Format

QoS Information is added at the end of the frame.

Non-WMM stations see normal header plus data.

When implementing WMM, the frame must represent the 802.11 AC and information such as

NoACK. The WMM-compatible frame has a different header than the original 802.11 frame.

When a station registers to an AP, it mentions its WMM capability in the QoS control field. A

great advantage of this technique is that the nQSTAs assume that the header ends after the

fourth MAC address. They assume that what is behind is part of the body.

Although they would not be able to read the QoS control field, they are not disturbed by other

stations sending QoS-enabled frames. Nevertheless, both the AP and the station sending or

receiving QoS-enabled frames must be WMM compliant to accept the frame. For an nQSTA,

the frame shows a frame check sequence (FCS) error, as the FCS is computed also on a part of

the frame that is not seen as the header.

In the QoS section, bits 0 to 2 show the traffic category. Bit 3 is set to 0. It is reserved for future

use if more traffic categories are to be defined. Bit 4 expresses if the frame marks the end of the

QoS period in the cell.10 Bits 5 to 6 specify if ACK is required or not. Bits 7 to 15 are reserved

for future use.

10 Because 802.11e is backward compatible with DCF, being hybrid (this is the meaning of Hybrid in Hybrid

Coordination Function), the QoS support starts as soon as a QSTA and its QoS enabled AP (QAP) communicate.

During that period of time, the AP can still communicate normally with nQSTAs.

www.CareerCert.info



802.11e, 802.1p, and DSCP Mapping This topic describes the process of mapping classifications to and from DSCP to class of

service.


WMM to 802.1p Mapping

Queue Name Application802.1p Default Mapping

AVVID 802.1p

Platinum Voice 6 5

Gold Video 5 4

Silver Best-Effort Traffic 3 0

BronzeOther or Background

1 1

The WMM QoS information gives a traffic category number. Where 802.11e defined eight

TCs, WMM kept only four TCs, one per access category: Platinum for the Voice AC, Gold for

the Video AC, Silver for the Best-Effort AC and Bronze for the Background AC. The other

four TCs defined in 802.11e are not kept in WMM. Just like the 802.11e protocol, WMM maps

each TC to an 802.1p CoS. The main difference is that, because there are four queues instead of

eight, some mappings are not available by default.

The following mapping is applied:

Queue Name Application 802.1p Default Mapping

Platinum Voice 6

Gold Video 5

Silver Best-Effort Traffic 3

Bronze Other or Background 1

This chart means that when a voice packet, classified WMM Platinum, is sent from a station,

the WMM certification, following the 802.11e protocol, recommends to map it to 802.1p CoS

6. When an 802.1p CoS of 6 is received, it can be translated to the Platinum queue.

Cisco implementation follows the IETF recommendations instead of the WMM certification

and recommends to map Platinum to CoS 5 and Gold to CoS 4. This allows reserving 6 to

Layer 3 network control traffic.

www.CareerCert.info



To help you understand this difference, the following table shows the recommended mapping

in a Cisco Architecture for Voice, Video, and Integrated Data and is compared to the default

WMM mapping:

Cisco AVVID

802.1p UP-

Based Traffic

Type

Cisco

AVVID IP

DSCP

Cisco AVVID

802.1p UP

IEEE 802.11e

UP

WMM Notes

Network Control

- 7 - - Reserved for network control only

Inter-Network Control

48 6 7 (AC_VO) - CAPWAP control

Voice 46 (EF) 5 6 (AC_VO) AC_VO Controller: Platinum QoS profile

Video 34 (AF41)

4 5 (AC_VI) AC_VI Controller: Gold QoS profile

Voice Control 24 (CS3) 3 4 (AC_VI) - -

Best Effort 0 (BE) 0 3 (AC_BE)

0 (AC_BE)

AC_BE Controller: Silver QoS profile

-

Background (Cisco AVVID Gold Background)

18 (AF21)

2 2 (AC_BK) - -

Background (Cisco AVVID Silver Background)

10 (AF11)

1 1 (AC_BK) AC_BK Controller: Bronze QoS profile

By default, no mapping is done on the controller. Mapping must be enabled on the controller

and the CoS equivalent can then be changed. It displays the default WMM to 802.1p mapping.

If your network infrastructure is based on AVVID, you should change it to reflect the AVVID

recommendations.

www.CareerCert.info



Summary This topic summarizes the key points that were discussed in this lesson.


Summary

QoS technologies refer to the set of tools and techniques that are used to manage and share network resources.

The first step in implementing QoS is to identify traffic, which is done with classification and marking.

Devices can tag their own traffic: a trust boundary must be created to determine which tag is trusted.

When congestion occurs, various queuing mechanisms allow congestion management: Low latency queuing is best adapted for voice traffic.

Some mechanisms can anticipate congestion and selectively drop packets in advance.

Policing and shaping can be added to limit bandwidth use based on traffic type.

Link efficiency mechanisms can enhance WAN link utilization.


Summary (Cont.)

Each QoS mechanism is used at key points of the network to create a global QoS policy.

In the wireless space, congestion has to take into account the shared medium and half-duplex nature of the RF environment.

In the wireless space, DCF is the legacy mechanism by which each station sends its traffic whenever the medium is available and a random countdown timer has expired.

802.11e creates QoS possibilities in the cell by prioritizing critical traffic within the stations and by allowing shorter timers for some flows.

WMM is a certification for vendors implementing a selected part of the 802.11e protocol.

With the 802.11e protocol, voice traffic is tagged 802.1p 6; Cisco follows the IETF recommendation to tag voice 5 instead.

www.CareerCert.info



References

For additional information, refer to these resources:

Cisco Enterprise Mobility 4.1 Design Guide:

http://www.cisco.com/en/US/docs/solutions/Enterprise/Mobility/emob41dg/emob41dg-

wrapper.html

IEEE 802.11e protocol:

http://standards.ieee.org/getieee802/

WMM:

http://www.wi-fi.org/white_papers/whitepaper-090104-wmm

www.CareerCert.info

http://www.cisco.com/en/US/docs/solutions/Enterprise/Mobility/emob41dg/emob41dg-wrapper.html

http://www.cisco.com/en/US/docs/solutions/Enterprise/Mobility/emob41dg/emob41dg-wrapper.html

http://standards.ieee.org/getieee802/



www.CareerCert.info


Lesson 2

Describing Wireless QoS Deployment Schemes

Overview An important element of quality of service (QoS) in the wireless space is the Control and

Provisioning of Wireless Access Points (CAPWAP) encapsulation. Wireless frames received at

the access point (AP) level are encapsulated into CAPWAP before being sent to the controller.

Frames received by the controller from the wired side are encapsulated into CAPWAP before

being forwarded to the AP. QoS levels can be configured at the wireless LAN (WLAN) level.

This lesson will show you how these different QoS elements interact with each other and how

QoS levels can be managed from the controller.

Objectives

Upon completing this lesson, you will be able to describe wireless QoS deployment schemes.

This ability includes being able to meet these objectives:

Describe QoS parameters for the wireless network

Describe the differences between upstream and downstream QoS

Describe how QoS affects network performance

www.CareerCert.info



QoS Parameters This topic describes QoS parameters for wireless networks.


QoS Parameters

Day 1 Day 2 Day 5 Some wireless devices are WMM and request level of QoS; some devices are not QoS aware.

Some applications need QoS both ways; some applications just need downstream QoS.

In the past, WLANs were used mainly to transport low-bandwidth, data-application traffic.

Currently, with the expansion of WLANs into vertical (such as retail, finance, and education)

and enterprise environments, WLANs transport high-bandwidth data applications in

conjunction with time-sensitive multimedia applications. This requirement led to the necessity

for wireless QoS. Several vendors, including Cisco, support proprietary wireless QoS schemes

for voice applications. To speed up the rate of QoS adoption and to support multivendor time-

sensitive applications, a unified approach to wireless QoS is necessary.

The IEEE 802.11e working group has completed the standard definition, but there are many

optional components. Just as occurred with 802.11 security in 802.11i, industry groups such as

the Wi-Fi Alliance and industry leaders such as Cisco are defining the key requirements in

WLAN QoS through their Wi-Fi Multimedia (WMM) and Cisco Compatible Extensions

programs, ensuring the delivery of key features and interoperation through their certification

programs. Cisco Unified Wireless products support WMM and WMM power save, as well as

admission control.

You need not apply QoS the same way everywhere. Some devices, such as wireless phones,

may need QoS support for both upstream and downstream flows. Some other devices, such as

portable video readers, may need QoS support only for the downstream. Other devices, such as

laptops used for data communication, may not need QoS support at all. You must deploy QoS

as a network-wide strategy. Prioritization is given to devices and applications needing it.

Devices not needing QoS support still need to be defined in the global strategy and are

classified in the best-effort category. .Traffic between controllers and APs use CAPWAP in

controller code release 5.2 and later and Lightweight Access Point Protocol (LWAPP) with

earlier code releases. This traffic also presents specific QoS characteristics.

www.CareerCert.info



Upstream and Downstream QoS This topic describes upstream and downstream QoS.


QoS Between Controller and APs

Ensures that packets receive the proper QoS handling end-to-end.

Makes sure that the packet will maintain QoS information as it traverses the network.

Policing of 802.11e user priority and 802.1p and IP DSCP values ensures that endpoints conform to network QoS policies.

Uses Cisco AVVID packet marking mappings and IEEE mappings as appropriate.

Supported on all Cisco controllers and all Cisco controller-based APs (calendar year 2009).

To configure an efficient QoS policy for the wireless network, you need to understand how

QoS tagging occurs between the AP, the controller, and the network.

Each WLAN is associated to a QoS profile matching one of the WMM classes. You must deal

with unique issues to maintain the appropriate settings. First, because the traffic from the AP to

the controller, and vice versa, is encapsulated in CAPWAP, the QoS settings need to be

maintained as well as used, even when the packet is encapsulated. WMM values are copied

from the client traffic to the CAPWAP headers.

The other key feature is the ability to support priority on the wireless network. These two

elements may be contradictory. A wireless client associated to a WLAN for which QoS level is

set to Bronze may require a Platinum QoS level. The controller must decide if this high level is

allowed or if the QoS level must be capped to Bronze.

The general rule is that a wireless client QoS level, between the AP and the controller in the

cell, cannot exceed the maximum QoS that is defined for the entire WLAN and configured on

the controller. The only exception is identity-based networking (IBN), where a RADIUS can

override the controller settings for a particular client.

www.CareerCert.info




QoS Downstream—Wired Client to WLC AP

Suppose that the network wired infrastructure sends a data packet to the controller and then to

the AP. This packet may have a differentiated services code point (DSCP) field or a class of

service (CoS) value when leaving the wired source. Configure the network infrastructure to

ensure that the CoS value is consistent with the DSCP value if both exist.

As packets are encapsulated between controllers and APs, they have both an inner header and

an outer header. The inner header contains the DSCP information received from the wired

network, while the outer header contain QoS information used to transfer the packet from the

controller to the AP and back.

When the packet reaches the controller, the original DSCP value is read and kept in the inner

DSCP field. If the packet does not contain a DSCP value, CoS is used instead. The packet is

transformed into an 802.11 frame and encapsulated into CAPWAP. During the encapsulation

process, the QoS value read in the DSCP field is compared to the QoS value applied to the

WLAN. Several cases can occur:

The controller WLAN has no QoS mapping: In that case, the outer header does not carry

any tag.

The controller WLAN has a QoS mapping that is higher than the DSCP or CoS value

in the received packet: For example, the WLAN is associated to Silver, 802.1p 3, and the

CoS in the packet is 2. In that case, the DSCP or CoS value requested in the packet is

transferred to the outer header. In this example, the outer header carries the CoS value 2

and the corresponding DSCP value.

The controller WLAN has a QoS mapping that is lower than the DSCP or CoS value

in the packet: For example, the WLAN is associated to Silver, 802.1p 3, and the CoS in

the packet is 5. In that case, the DSCP or CoS value in the outer header is capped to the

WLAN maximum. In this example, the outer header carries the CoS value 3 and the

corresponding DSCP value.

www.CareerCert.info



You can see that in any case, the outer QoS level does not exceed the maximum defined for the

WLAN. The packet is transferred then to the switch. Because the link between the controller

and the switch is a trunk port, it supports 802.1Q, and therefore 802.1p. The outer header

contains both the DSCP information and the 802.1p information.

CAPWAP control traffic is different from encapsulated data traffic. CAPWAP control traffic

between the controller and the AP is tagged with DSCP 48, which is IP precedence 6, or CoS 6.

www.CareerCert.info




QoS Downstream - AP to Wireless Client

If the AP is connected to a trunk port of a switch, which is the minority of cases, the frame is

sent “as is” down to the AP. Note that the outer tag might be altered by internetworking devices

between the controller and the AP.

If the AP is connected to an access port of a switch, which is common for APs operating in

local mode, keeping the 802.1p tag is not possible because access ports do not accept tagged

frames. Before sending the frame to the AP, the access switch removes the 802.1p tag, and the

AP receives the frame with only the DSCP information on the outside header.

In either case, the priority assigned to the frame between the controller and the AP is capped by

the controller WLAN QoS mapping. After the packet has arrived at the AP, the inner packet is

retrieved and distributed to the cell. Two cases can occur:

The client has no WMM support: The packet is placed in the default transmit (Tx) queue

for the WLAN, which is the Distributed Coordination Function (DCF) queue, without any

WMM prioritization.

The client has WMM support: The packet is placed in the appropriate queue for the

802.11e traffic category (TC) value derived from the CAPWAP packet outer DSCP value.

This mapping is set for the entire WLAN. IBN allows RADIUS servers to override WLAN

parameters for specific clients. If the frame were to be sent to a client for whom IBN was used

in conjunction with a specific QoS tagging limit defined by the RADIUS server, the controller

would cap the DSCP to the client-specific limit instead of the WLAN general limit, both for the

outer and the inner headers. If the cap limit defined by the RADIUS is higher than the general

limit for the WLAN, the client might receive a higher priority than the general rule for the

WLAN would allow.

www.CareerCert.info




QoS Downstream—QoS Parameters

Day 1 Day 2 Day 5

Wired SourceWLAN

ConfigurationWLC

AP (WMM Client)

AP (Non-WMM Client)

DSCP or 802.1p present

or both

802.1p mapping present

Outer header capped, inner

DSCP maintained

Outer DSCP becomes

802.11e, WMM queue applies

QoS values are dropped, DCF queue applies

DSCP or 802.1p present

or both

802.1p mapping not present

No outer QoS tagging, inner

DSCP maintained

DCF queue applies

DCF queue applies

DSCP or 802.1p not

present

802.1p mapping present or not

No outer QoS tagging (*)

DCF queue applies

DCF queue applies

(*) The results using v5.2 is not consistent with VoWLAN DG using v4.1, May 6, 2008. Therefore, check the latest documentation.

The table summarizes the QoS tagging translation that occurs when a packet is received from a

wired source and is to be transmitted to a wireless client. Understand that if a tag is present in

the incoming data frame, the outer header shows a tag only if there is an 802.1p mapping for

the WLAN. This outer tag is capped with the value defined for the WLAN. Then you can place

the frame into a WMM queue or send it through DCF if the client is not WMM-compliant. If no

incoming QoS tag is present, the frame is untagged between the controller and the AP and sent

through DCF in the cell.

You can observe this behavior when using controller code release v5.2. On former code

releases and as indicated in the Cisco voice over wireless LAN (VoWLAN) design guide dated

May 6, 2008, and based on controller code release v4.1, the behavior is different. With this

earlier code, if the controller is configured for 802.1p mapping, when untagged packets arrive

at the controller interface, the controller encapsulates the packet into LWAPP and sends the

packet with an outside header bearing the 802.1p mapping max value.

For example, with an 802.1p mapping set to 5, an untagged packet would have an 802.1p outer

CoS tag set to 5 between the controller and the AP. Verify on the latest controller and

VoWLAN documentation which behavior applies to the version of controller code that you are

using.

www.CareerCert.info




QoS Upstream—Wireless Client to AP

Packets sent from wireless clients present a different case. The client might not support WMM,

in which case no 802.11e field is set on the 802.11 frame. If the client supports WMM, the

802.11e field displays a traffic category. After the AP receives the 802.11 frame, two cases can

occur:

The controller WLAN has no QoS mapping. The frame is encapsulated and sent to the

controller without any QoS information in the outer header. As the inner frame is the

802.11 frame, it contains the 802.11e information if the client supports WMM.

The controller WLAN has a QoS mapping. The behavior depends on the client WWM

support:

— For clients not supporting WMM, the AP tags the outer header with the QoS value

mapped to the WLAN that is indicated by the controller.

— For clients supporting WMM, the AP tags the outer header with the QoS value

requested by the client, capped to the controller WLAN QoS maximum. For

example, suppose that the WLAN is set to Silver, CoS 3 and DSCP 26. A client

requesting Bronze, which is CoS 1 or DSCP 10, will get the QoS level it requests. If

a client requests more than Silver, such as Platinum or Gold, it will get Silver, which

is the maximum allowed by the AP.

Given that the AP is normally on an access port1, it cannot tag 802.1p information on the 802.3

outer header. Therefore, the AP tags only the DSCP information on the outer header. In any

case, the inner header contains both the original 802.11e and DSCP header information.

1 The AP might be on a trunk port in some cases, but the AP has no way of knowing the configuration of the switch

port.

www.CareerCert.info



Here again, the general rule is that the client QoS level does not exceed the maximum defined

for the WLAN, at least between the AP and the controller. IBN allows the RADIUS to override

the WLAN settings also in this case. With IBN, the DSCP maximum value matches the DSCP

value sent by the RADIUS server, and not the common maximum for the WLAN. If the DSCP

value sent by the RADIUS is higher, the client might have a higher priority level than the

WLAN would allow. If the DSCP value sent by the RADIUS is lower than the WLAN

maximum, the IBN client will be capped at the maximum defined by the RADIUS, and not the

maximum defined for the WLAN.

Note APs carry out Enhanced Distributed Coordination Function (EDCF)-like queuing on the radio

egress port only. APs do FIFO queuing only on the Ethernet egress port.

www.CareerCert.info




QoS Upstream—AP to Controller

On the path to the controller, internetworking devices might alter the outer header. For

example, the switch to which the controller is connected might map the outer DSCP to a CoS

value and tag the frame sent to the controller with 802.1p accordingly.

After the controller receives the frame, it retrieves the inner frame that contains the 802.11e and

DSCP values requested by the client. The controller converts the 802.11 frame into an 802.3

frame without altering the QoS request. In other words, the controller does not override the

QoS level requested by the client, but converts the 802.11e value into the corresponding DSCP

and CoS values. The frame leaves the controller with the QoS level originally requested by the

client. Apart from the IBN case, the outer header is capped to the WLAN maximum QoS level,

while the inner frame carries the original QoS level requested by the client.

The wired infrastructure is supposed to have a QoS policy configured to decide if this requested

value is kept as it is and trusted, or if it should be remapped to another value.

www.CareerCert.info




QoS Upstream—QoS Parameters

Day 1 Day 2 Day 5

Wireless Source

WLAN Configuration

AP WLC

WMM802.1p mapping

present

Outer header capped, inner 802.11e and DSCP

maintained

Inner DSCP maintained, outer DSCP becomes

802.1p

WMM802.1p mapping

not present

No outer QoS tagging, inner 802.11e and DSCP

maintained

Inner DSCP maintained, no 802.1p

Non-WMM802.1p mapping present or not

No outer QoS tagging (*)No DSCP or 802.1p

tagging

(*) The results using v5.2 are not consistent with VoWLAN DG using v4.1, May 6, 2008. Therefore, check the latest documentation.

This table summarizes the QoS tagging translation when a packet is received from a wireless

client and sent to a wired destination beyond the controller. The key point to understand is that

if the client does request QoS through WMM and DSCP, the AP tags the outer header only if

an 802.1p mapping has been defined for this WLAN.

In this last case, the outer header QoS value is capped with the 802.1p mapping defined for this

WLAN. If the client is not WMM-compliant and does not require any QoS level, the AP cannot

tag the outer header, and the packet is forwarded to the controller untagged. The packet is also

untagged when leaving the controller.

You can observe this behavior when using controller code release v5.2. On former code

releases and as indicated in the Cisco voice over wireless LAN (VoWLAN) design guide dated

May 6, 2008, and based on controller code release v4.1, the behavior is different. With this

earlier code, if the controller is configured for 802.1p mapping, when untagged packets arrive

at the controller interface, the controller encapsulates the packet into LWAPP and sends the

packet with an outside header bearing the 802.1p mapping max value.

For example, with an 802.1p mapping set to 5, an untagged packet would have an 802.1p outer

CoS tag set to 5 between the AP and the controller. Verify on the latest controller and

VoWLAN documentation which behavior applies to the version of controller code that you are

using.

www.CareerCert.info




QoS on H-REAP

The Hybrid Remote-Edge Access Points (H-REAPs) present a special case, because they are

separated from the controller via a WAN and may lose connectivity in case of WAN link

failure. H-REAPs are configured to locally switch the traffic or forward it to the controller.

For WLANs that have data traffic forwarded to the controller, the behavior is same as regular

local-mode APs. For locally-switched WLANs with WMM traffic, the AP marks the dot1p

value in the dot1q VLAN tag for upstream traffic. This occurs only on tagged VLANs; that is,

not native VLANs. For downstream traffic, the H-REAP uses the incoming dot1q tag from the

Ethernet side and uses this to queue and mark the WMM values on the radio on which the

packet is to be sent.

The WLAN QoS profile is applied both for upstream and downstream packets. For

downstream, if an IEEE 802.1p value is higher than the default WLAN value, the default

WLAN value is used. For upstream, if the client sends a WMM value that is higher than the

default WLAN value, the default WLAN value is used. For non-WMM traffic, there is no CoS

marking on the client frames from the AP.

www.CareerCert.info



QoS and Network Performance This topic describes how QoS impacts network performance.


QoS and Network Performance

The application of QoS features might not be detected easily on a lightly loaded network. If

latency, jitter, and loss are noticeable when the media is lightly loaded, it indicates either a

system fault, poor network design, or that the latency, jitter, and loss requirements of the

application are not a good match for the network. QoS features start to be applied to application

performance as the load on the network increases. QoS works to keep latency, jitter, and loss

for selected traffic types within acceptable boundaries.

Radio QoS can be applied both upstream and downstream if clients support WMM. Radio QoS

can be downstream only if clients do not support WMM. When providing only radio

downstream QoS from the AP, radio upstream client traffic is treated as best effort. A client

must compete with other clients for upstream transmission as well as competing with best-

effort transmission from the AP. Under certain load conditions, a client can experience

upstream congestion, and the performance of QoS-sensitive applications might be unacceptable

despite the QoS features on the AP. Ideally, upstream and downstream QoS should be operated

either by using WMM on both the AP and WLAN client, or by using WMM and a client-

proprietary implementation.

Note WLAN client support for WMM does not mean that the client traffic automatically benefits

from WMM. The applications looking for the benefits of WMM assign an appropriate priority

classification to their traffic, and the operating system needs to pass that classification to the

WLAN interface. In purpose-built devices, such as VoWLAN handsets, this is part of the

design. However, if implementing on a general-purpose platform such as a PC, application

traffic classification and OS support must be implemented before the WMM features can be

used to good effect.

www.CareerCert.info





Summary

Some clients request QoS, some others are not QoS aware; in either case, the AP and the controller have to decide which QoS level to assign to each data stream.

As much as possible, the QoS level requested by the clients is maintained in the outer header between the AP and the controller; the WLAN configuration for QoS determines the maximum QoS level for traffic to and from that QoS.

The WLAN configuration for QoS allows the administrator to determine a QoS level for clients in a WLAN, even if these clients are not able by themselves to request QoS.

www.CareerCert.info


Lesson 3

Configuring the Controller and Cisco WCS for QoS

Overview With the diversity of clients and needs, you can configure many quality of service (QoS)-related parameters on the controller, and most of them can be deployed as a template from the Cisco Wireless Control System (Cisco WCS). This lesson will guide you through these different parameters to show you how they are configured to best support voice deployments.

Objectives Upon completing this lesson, you will be able to configure the controller and Cisco WCS for QoS. This ability includes being able to meet these objectives:

� Assign a QoS profile to a WLAN on a controller

� Configure QoS profiles on a controller

� Configure voice parameters on the controller

� Configure EDCA support on the controller

� Configure QoS roles on a controller

� Configure QoS templates on the Cisco WCS

www.CareerCert.info



Assign a Cisco WLC QoS Profile to a WLAN on a Controller

This topic describes the procedure to assign a QoS profile to a wireless LAN (WLAN) on a controller.


Cisco WLC—WLANs > Edit > QoS

You can configure the WLAN with various default QoS profiles. Each of the profiles (platinum, gold, silver, or bronze) is annotated with its typical use. In addition, through authentication, authorization, and accounting (AAA) you can assign a client a QoS profile based on its identity. For a typical enterprise, WLAN deployment parameters, such as per-user bandwidth contracts and over-the-air QoS, should be left at their default values, and standard QoS tools, such as Wi-Fi Multimedia (WMM) and wired QoS, should be used to provide optimum QoS to clients.

Using the controller GUI, follow these steps to assign a QoS profile to a WLAN.

Step 1 Click WLANs to open the WLANs page.

Step 2 Click the name of the WLAN to which you want to assign a QoS profile.

Step 3 When the WLANs > Edit page appears, click the QoS tab.

www.CareerCert.info



Step 4 From the Quality of Service (QoS) drop-down box, choose one of the following:

� Platinum (voice)

� Gold (video)

� Silver (best effort): (the default value)

� Bronze (background).

Step 5 Click Apply to commit your changes. Click Save Configuration to save your changes.

www.CareerCert.info




� QBSS Load IE is an 802.11e feature, enabled with WMM.� APs use it to send their load level in beacons and probe responses.� On the 802.11b/g band, this information is used for phones to choose the

best AP.� 802.11a QoS uses TSpec instead of QBSS.� Possible combinations:

– WMM allowed or required -> 802.11e standard QBSS IE– 7920 Client CAC -> Prestandard QBSS IE; incompatible with WMM– 7920 AP CAC -> New CCX QBSS IE; compatible with WMM

Cisco WLC—WLANs > Edit > QoS: WMM and QBSS

The QoS Basic Service Set (QBSS) Load Information Element (IE) is defined in the IEEE 802.11e protocol. This information element is sent in the AP beacons and provides information about the number of stations currently associated, a measure of the channel utilization and a measure of the additional load that the AP can accept. This information is very important because it enables a phone to determine if the AP is the best candidate or if the phone should join another AP. Both the Cisco Unified Wireless IP Phone 7920 and 7921G can use the QBSS Load IE on the 802.11b/g band.

Because this information element is critical, Cisco has been supporting it since the first drafts of the 802.11e protocol. There are actually three QBSS IEs that need to be supported in certain situations:

1. Old QBSS Load IE (Draft 6 (pre-standard))

2. New QBSS Load IE (Draft 13 802.11e (standard))

3. New distributed Call Admission Control (CAC) load IE (a Cisco proprietary information element)

The QBSS Load IE used depends on which boxes are checked in the illustration. The wireless LAN controller (WLC) provides Wireless IP Phone 7920 support through the client CAC limit or AP CAC limit. The various combinations of WMM, client CAC limit, and AP CAC limit result in different QBSS IEs being sent:

� WMM Policy: If you enable WMM only, the new 802.11e standard QBSS Load IE is sent out in the beacons and probe responses. This setting is adapted for non-proprietary Cisco phones.

www.CareerCert.info



� Client CAC limit: If you check this box, the old prestandard QBSS IE is sent out in the beacons and probe responses on the b/g radios. The Wireless IP Phone 7920 uses a CAC setting that is set on the client. This supports legacy Wireless IP Phone 7920 prior to version 2.01. The client listens passively to the AP QBSS IE and makes its own decision about remaining space on the AP. The Wireless IP Phone 7921G does not use this older feature. Use this prestandard feature only with Wireless IP 7920 phones with old firmware. It is not compatible with WMM. You cannot enable WMM and also check the Client CAC limit box.

� AP CAC limit: If this box is checked, the number Cisco Proprietary New distributed CAC load IE is sent in the beacons and probe responses for b/g radios. The Wireless IP Phone 7920 uses CAC settings learned from WLAN advertisement. This is valid for Wireless IP Phone 7920 with firmware newer than 2.01 and the Wireless IP 7921G phones.

You can enable WMM, thus sending the 802.11e standard QBSS Load IEs for nonproprietary Cisco phones and check the AP CAC limit box to replace it with the Cisco IE for the Wireless IP Phone 7920 and 7921G phones. The various QBSS IEs use the same ID, and, therefore, the three QBSSs are mutually exclusive. For example, the beacons and probe responses can contain only one QBSS IE. Also, QBSS is not supported when using 802.11a. The right configuration depends on you clients and the bands on which the WLAN is enabled:

� If your WLAN is enabled only for the 802.11b/g band and you have recent Wireless IP 7920 and 7921G phones, enable WMM and use the AP CAC limit feature. Only the Cisco IE will be sent for the QBSS information, but all other WMM features will be enabled.

� If your WLAN is enabled for both 802.11a and 802.11b/g, enable WMM and AP CAC limit. WMM will be available for both 802.11a and 802.11b/g channels. The Cisco QBSS IE will be sent on the 802.11b/g band. Traffic specification (TSpec) will be used in the 802.11a band.

� If your WLAN is enabled for only the 802.11a band, enable WMM.

WMM defines many other features beyond the QBSS information element. On the 802.11a band, the Wireless IP Phone 7921G uses a WMM feature called TSpec to specify what type of traffic it is about to send and how much bandwidth this traffic requires. The AP can accept (or refuse) the traffic stream and reserve bandwidth for it. Wireless IP 7920 phones only use the 802.11b/g band. Wireless IP 7921G phones use the QBSS element in the 802.11b/g band and TSpec in the 802.11a band.

www.CareerCert.info



Configure Cisco WLC QoS Profiles This topic describes the procedure to configure QoS profiles.


Cisco WLC—Wireless > QoS > Profiles > Edit:Wired QoS Mapping

� Protocol Type = None: Outer header on frames between AP and controller are untagged (this is the default).

� Cisco and the IETF recommend using 5 for voice.

When a wireless packet reaches the AP, it is encapsulated into Control and Provisioning of Wireless Access Points (CAPWAP) and forwarded to the controller. By default, and regardless of the wireless traffic category (TC) used, the AP does not tag the outer header. If you want to see the outer header tagged on the wired side, you need to enable wireless-to-wired QoS (802.1p) mapping for the default wireless priority associated to the WLAN.

Suppose that WMM voice traffic is arriving with a TC of Platinum at the AP. If a wireless-to-wired QoS mapping was configured, the AP automatically performs a TC to class of service (CoS), and then CoS-to-DSCP mapping for this traffic based on the maximum 802.1p mapping configured in the profile.

If the CoS value in the WLC configuration is set to a value less than the automatic mapping result, this changed value is used by the WLAN QoS profile at the AP to set the maximum CoS marking used and, therefore, set which WMM access category (AC) to use. If Platinum is requested and the maximum 802.1p mapping set to 5, as recommended by Cisco, the AP maps the Platinum level requested by the client to CoS 6 as per the WMM specification, then 5 as per the configuration. The CoS 5 is then mapped to differentiated services code point (DSCP) 46 Expedited Forwarding (EF) and the packet is tagged accordingly. The key point is that with the Cisco Unified Wireless Network, you should always think in terms of IEEE 802.11e classifications and allow the Cisco Unified Wireless Network Solution to take responsibility for converting between IEEE classification and the Cisco QoS baseline.

www.CareerCert.info




Cisco WLC—Wireless > QoS > Profiles > Edit

You can select a QoS profile from a list of four (bronze, silver, gold, platinum) for a given WLAN. There is a controller-wide option to change the characteristics of each profile. You can access it from Wireless > QoS > Profiles > Edit Profile.

To configure QoS profiles using the controller GUI, first disable the 802.11a and 802.11b/g networks. To disable the radio networks, click Wireless > 802.11a/n or 802.11b/g/n > Network, uncheck the 802.11a (or 802.11b/g) Network Status check box, and click Apply. Then click Wireless > QoS > Profiles to open the QoS Profiles page. Click the name of the profile that you want to configure to open the Edit QoS Profile page.

You can change the description of the profile by modifying the contents of the Description field.

It is generally recommended that the Per-User Bandwidth Contracts settings be left at their default values and that the 802.11 WMM features be used to provide differentiated services. When you set the Per-User Bandwidth Contracts parameters to 0 (Off), the traffic allowed is unlimited and is restricted only by other 802.11 limitations. The possible parameters are as follows:

� Average Data Rate: Operator-defined average data rate (kb/s) for non-User Datagram Protocol (UDP) traffic. Valid values are from 0 to 60,000. A value of 0 imposes no bandwidth restriction on the profile.

� Burst Data Rate: Operator-defined peak data rate (kb/s) for non-UDP traffic. Valid values are from 0 to 60,000. A value of 0 imposes no bandwidth restriction on the profile.

Note The Burst Data Rate should be greater than or equal to the Average Data Rate. Otherwise,

the QoS policy might block traffic to and from the wireless client.

� Average Real-Time Rate: Operator-defined average data rate (kb/s) for UDP traffic. Valid values are from 0 to 60,000; the default value is 0 (Off).

www.CareerCert.info



� Burst Real-Time Rate: Operator-defined peak data rate (kb/s) for UDP traffic. Valid values are from 0 to 60,000; the default value is 0 (Off).

Note The Burst Real-Time Rate should be greater than or equal to the Average Real-Time Rate.

Otherwise, the QoS policy might block traffic to and from the wireless client.

The over-the-air QoS values define, for each profile:

� The maximum RF usage per AP: Valid values are from 1 to 100 (default). For example, if you set 50 percent for Bronze QoS, all the Bronze WLAN users combined will not get more than 50 percent of the available RF bandwidth. Actual throughput could be less than 50 percent, but it will never be more than 50 percent.

� The queue depth: Causes packets with a greater value to be dropped at the AP. Valid values are from 50 to 500. Nominally, they are 100 for Bronze, 150 for Silver, 255 for Gold, and 255 for Platinum. For a typical enterprise, you should leave WLAN deployment parameters, such as per-user bandwidth contracts and over-the-air QoS, at their default values and use standard QoS tools, such as WMM and wired QoS, to provide optimum QoS to clients.

www.CareerCert.info



Configure Cisco WLC Voice Parameters This topic describes the procedure to configure voice parameters on the controller.


Cisco WLC—Wireless > 802.11a/n > Voice: Wireless CAC

CAC enables an AP to maintain controlled QoS when the wireless LAN experiences congestion. CAC enables the client to specify how much bandwidth or shared-medium time would be required to accept a new call and, in turn, it enables the AP to determine if it is capable of accommodating this particular call. The AP rejects the call if necessary to maintain the maximum allowed number of calls with acceptable quality.

To use CAC with voice applications, first do the following:

� Configure the WLAN for Platinum QoS.

� Enable WMM protocol for the WLAN.

Then check the Admission Control (ACM) checkbox. This enables admission control, based upon the AP capacity, but does not take into account the possible channel-loading impact of other APs in the area. To include this "channel load" in capacity calculations, check the Load-based AC checkbox as well as the Admission Control (ACM) checkbox. Load-based CAC incorporates a measurement scheme that takes into account the bandwidth consumed by all traffic types from itself, from cochannel APs, and by colocated channel interference.

Load-based CAC also covers the additional bandwidth consumption resulting from (physical layer) PHY and channel impairment. In load-based CAC, the AP periodically measures and updates the use of the RF channel, channel interference, and the additional calls that the AP can admit. The AP admits a new call only if the channel has enough unused bandwidth to support that call. By periodic measuring and updating, load-based CAC prevents over-subscription of the channel and maintains QoS under all conditions of WLAN loading and interference. Load-based CAC is preferred over simple CAC.

www.CareerCert.info




Cisco WLC—Wireless > 802.11a/n > Voice: Expedited Bandwidth

� Used for special bandwidth reservation requests such as emergency calls.

� Used in combination with CAC.

� Load-based CAC allows higher prioritization than ACM,

� Best-effort logic: If more than 90% bandwidth is used, expedited bandwidth is not guaranteed.

The expedited bandwidth request feature enables Cisco Compatible Extensions v.5 clients to indicate the urgency of a WMM traffic specification (TSpec) request (for example, an emergency 911 call) to the WLAN. When the controller receives this request, the controller attempts to facilitate the urgency of the call in any way possible without potentially altering the quality of other TSpec calls that are in progress.

Expedited bandwidth requests are disabled by default. If you configure the WLAN in such a way that it does not support Cisco Compatible Extensions v.5, or if you disable expedited bandwidth requests, the controller ignores all expedited requests and processes TSpec requests as normal TSpec requests.

See the following table for examples of TSpec request handling for normal TSpec requests and expedited bandwidth requests.

www.CareerCert.info



CAC Mode Reserved Bandwidth for Voice Calls

Usage Normal TSpec Request

TSpec with Expedited Bandwidth Request

Less than 75% Admitted Admitted

Between 75% and 90% (reserved bandwidth for voice calls exhausted)

Rejected Admitted

Bandwidth-

based CAC

More than 90% Rejected Rejected

Less than 75% Admitted Admitted

Between 75% and 90% (reserved bandwidth for voice calls exhausted)

Rejected Admitted

Load-based

CAC

75% (default setting)

More than 90% Rejected Admitted if the voice traffic load is light relative to the data traffic load. Otherwise, rejected.

www.CareerCert.info



Configure Cisco WLC EDCA Support This topic describes the procedure to tune Enhanced Distributed Channel Access (EDCA) support on the controller.


Cisco WLC—Wireless > 802.11a/n > Voice > EDCA Parameters

After you enable WMM, EDCA is active in the cell between the AP and clients supporting WMM. Different vendors have different requirements for minimum contention window (CWmin), maximum contention window (CWmax), and arbitrated interframe space (AIFS). Instead of having you tune these parameters manually individually, Cisco offers several thoroughly tested and optimized profiles that encompass major vendor solutions and requirements. The default is “WMM” and fits most situations with a mix of data, voice, and video clients. You can change it if the main clients in the cell are specific, as follows:

1. To change EDCA parameters, disable the corresponding radio and disable the WLANs that use WMM. To disable the radio network, click Wireless and then Network under 802.11a/n or 802.11b/g/n, uncheck the 802.11a (or 802.11b/g) Network Status check box, and click Apply.

2. Click EDCA Parameters under 802.11a/n or 802.11b/g/n. The 802.11a (or 802.11b/g) > EDCA Parameters page appears.

3. Choose one of the following options from the EDCA Profile drop-down box:

� WMM: Enables the Wi-Fi Multimedia default parameters. This is the default value. Choose this option when voice or video services are not deployed on your network.

� SpectraLink Voice Priority: Enables SpectraLink voice priority parameters. Choose this option if SpectraLink phones are deployed on your network to improve the quality of calls.

www.CareerCert.info



� Voice Optimized: Enables EDCA voice-optimized profile parameters. Choose this option when voice services other than SpectraLink are deployed on your network.

� Voice & Video Optimized: Enables EDCA voice- and video-optimized profile parameters. Choose this option when both voice and video services are deployed on your network.

Note If you deploy video services, you must disable Admission Control (ACM). ACM is called

wireless CAC (Call Admission Control) for voice, and simply ACM for video. You can find it in Wireless > 802.11a/n or 802.11b/g/n > Video.

1. To enable MAC optimization for voice, check the Enable Low Latency MAC check box. Otherwise, leave this check box unchecked, which is the default value. This feature enhances voice performance by controlling packet retransmits and appropriately aging out voice packets on lightweight APs, thereby improving the number of voice calls serviced per AP. With low latency MAC, the AP keeps track of the voice packets timestamp and does not forward packets coming from the wired side to the cell if their timestamp is too old. The phone would drop such packets anyway, so not sending them saves time and cell bandwidth.

Note You should enable low latency MAC only if the WLAN allows WMM clients. If you enable

WMM, then you can use low latency MAC with any of the EDCA profiles.

2. Click Apply to commit your changes.

3. To re-enable the radio network, click Network under 802.11a/n or 802.11b/g/n, check the 802.11a (or 802.11b/g) Network Status check box, and click Apply.

4. Click Save Configuration to save your changes.

www.CareerCert.info



Configure Cisco WLC QoS Roles This topic describes the procedure to configure QoS roles.


Cisco WLC—QoS Roles

� Used to assign specific bandwidth to temporary web-authenticated guest-user WLANs.

� Defined in three stages: creating the role, configuring the bandwidth for the role, creating users with guest roles.

After you configure a QoS profile and apply it to a WLAN, it limits the bandwidth level of clients associated to that WLAN. You can map multiple WLANs to the same QoS profile, which can result in bandwidth contention between regular users (such as employees) and guest users. To prevent guest users from using the same level of bandwidth as regular users, you can create QoS roles with different (and presumably lower) bandwidth contracts and assign them to guest users. The QoS Roles for Guest Users feature applies to web-authentication-based WLANs.

You can use the controller GUI or command-line interface (CLI) to configure up to ten QoS roles for guest users. The QoS role is defined in three stages: configure the QoS role, configure the bandwidth for the role, and then create users with guest roles.

Follow these steps to configure QoS roles using the controller GUI.

1. Click Wireless > QoS > Roles to open the QoS Roles for Guest Users page.

2. To create a new QoS role, click New. The QoS Role Name > New page appears.

3. In the Role Name field, enter a name for the new QoS role. The name should uniquely identify the role of the QoS user (such as Contractor, Vendor, and so on).


www.CareerCert.info




Cisco WLC—QoS Roles: Bandwidth

After you create it, you must edit the QoS role to define its settings:

1. To edit the bandwidth of a QoS role, click the name of the QoS role. The Edit QoS Role Data Rates page appears.

Note The values that you configure for the per-user bandwidth contracts affect only the amount of

bandwidth going downstream (from the AP to the wireless client). They do not affect the bandwidth for upstream traffic (from the client to the AP). They do not affect traffic on the wired side.

2. To define the average data rate for TCP traffic on a per-user basis, enter the rate in kb/s in the Average Data Rate field. You can enter a value between 0 and 60,000 kb/s (inclusive). A value of 0 imposes no bandwidth restriction on the QoS role.

3. To define the peak data rate for TCP traffic on a per-user basis, enter the rate in kb/s in the Burst Data Rate field. You can enter a value between 0 and 60,000 kb/s (inclusive). A value of 0 imposes no bandwidth restriction on the QoS role.

Note The Burst Data Rate should be greater than or equal to the Average Data Rate. Otherwise,

the QoS policy may block traffic to and from the wireless client.

4. To define the average real-time rate for UDP traffic on a per-user basis, enter the rate in kb/s in the Average Real-Time Rate field. You can enter a value between 0 and 60,000 kb/s (inclusive). A value of 0 imposes no bandwidth restriction on the QoS role.

5. To define the peak real-time rate for UDP traffic on a per-user basis, enter the rate in kb/s in the Burst Real-Time Rate field. You can enter a value between 0 and 60,000 kb/s (inclusive). A value of 0 imposes no bandwidth restriction on the QoS role.

www.CareerCert.info



Note The Burst Real-Time Rate should be greater than or equal to the Average Real-Time Rate.

Otherwise, the QoS policy may block traffic to and from the wireless client.


7. Click Save Configuration to save your changes.

www.CareerCert.info




Cisco WLC—New Guest Users

The guest user to whom the bandwidth restriction applies is authenticated using Web Authentication. The user can be authenticated via RADIUS or be a local network user defined on the controller. To create a Local Net User with a QoS Role on the controller, follow theses steps:

Click Security > AAA > Local Net Users to open the Local Net Users page. This page lists any local network users that have already been configured. It also specifies any guest users and the QoS role to which they are assigned (if applicable).

Perform one of the following:

� To edit an existing local network user, click the username for that user. The Local Net Users > Edit page appears. You can change some items, such as password or associated WLAN.

� To add a local network user, click New. The Local Net Users > New page appears.

www.CareerCert.info




Cisco WLC—Assigning Guest Roles to Guest Users

If you are adding a new user, enter a username for the local user in the User Name field. You can enter up to 24 alphanumeric characters.

Note Local network usernames must be unique because they are all stored in the same database.

In the Password and Confirm Password fields, enter a password for the local user. You can enter up to 24 alphanumeric characters.

If you are adding a new user, check the Guest User check box if you want to limit the amount of time that the user has access to the local network. The default setting is unchecked.

If you are adding a new user and you checked the Guest User check box, enter in the Lifetime field the amount of time (in seconds) that the guest user account is to remain active. The valid range is 60 to 2,592,000 seconds (30 days) inclusive, and the default setting is 86,400 seconds.

If you want to assign a QoS role to this guest user, check the Guest User Role check box. The default setting is unchecked.

Note If you do not assign a QoS role to a guest user, the bandwidth contracts for this user are

defined in the QoS profile for the WLAN.

If you are adding a new user and you checked the Guest User Role check box, choose the QoS role that you want to assign to this guest user from the Role drop-down box.

From the WLAN Profile drop-down box, choose the name of the WLAN that the local user will access. If you choose Any WLAN, which is the default setting, the user can access any of the configured WLANs. These WLANs need to rely on Web Authentication.

www.CareerCert.info



In the Description field, enter a descriptive title for the local user (such as "User 1").

Click Apply to commit your changes.

Click Save Configuration to save your changes.

Note You can create an entry on the RADIUS server for a guest user and enable RADIUS authentication for the WLAN on which web authentication is performed. Rather than adding a guest user to the local user database from the controller, you need to assign the QoS role

on the RADIUS server itself.

Note To do so, add a "guest-role" Airespace Cisco attribute (vendor ID 14179) on the RADIUS server with a data type of "string" and a return value of "11." This attribute is sent to the controller when authentication occurs. If a role with the name returned from the RADIUS server is found configured on the controller, the bandwidth associated to that role is enforced for the guest user after authentication completes successfully.

www.CareerCert.info



Configure QoS Using Cisco WCS This topic describes the procedure to configure QoS templates using Cisco WCS.


Cisco WCS—QoS Profiles Templates

You can change the QoS profiles modified on the controller via a WCS template. This feature lets you apply the profile change to several controllers with one click.

Follow these steps to make modifications to the quality of service profiles.

1. Choose Configure > Controller Templates.

2. On the left sidebar menu, choose System > QoS Profiles. The QoS Template window appears, and the number of controllers to which the template is applied automatically populates.

3. Set the following values in the Per-User Bandwidth Contracts portion of the window. All have a default of 0 (Off).

— Average Data Rate: The average data rate for non-UDP traffic

— Burst Data Rate: The peak data rate for non-UDP traffic

— Average Real-Time Rate: The average data rate for UDP traffic.

— Burst Real-Time Rate: The peak data rate for UDP traffic

4. Set the following values for the Over the Air QoS portion of the window.

— Maximum Rf Usage Per AP: The maximum air bandwidth available to clients. The default is 100 percent.

— Queue Depth: The depth of queue for a class of client. The packets with a greater value are dropped at the AP.

www.CareerCert.info



5. Set the following values in the Wired QoS Protocol portion of the window.

— Protocol: Choose 802.1P to activate 802.1p priority tags or None to deactivate 802.1p priority flags.

— 802.1P Tag: Choose 802.1p priority tag for a wired connection from 0 to 7. This tag is used for traffic and CAPWAP packets.

6. Click Save.

www.CareerCert.info




Cisco WCS—TSM Templates

The Cisco WCS also lets you configure Traffic Stream Metrics (TSM) templates. APs and clients measure the metrics; APs collect the measurements and then send them to the controller. The APs update the controller with traffic stream metric information every 90 seconds, and 10 minutes of data is stored at one time. Cisco WCS queries the controller for the metrics and displays them in the Traffic Stream Metrics QoS Status report. These metrics are compared to threshold values to determine their status level. If any of the statistics are displaying a status level of fair (yellow) or degraded (red), the administrator can investigate the QoS of the wireless LAN.

For the APs to collect measurement values, you must enable TSMs on the controller. The TSM template allows the administrator to modify default thresholds and influence the way the Cisco WCS reports data collected from the controllers. To configure a TSM template, proceed as follows:

1. Choose Configure > Controller Templates.

2. On the left sidebar menu, choose System > Traffic Stream Metrics QoS. The Traffic Stream Metrics QoS Status Configuration window appears

The Traffic Stream Metrics QoS Status Configuration window shows several QoS values. An administrator can monitor voice and video quality of the following:

� Upstream delay

� Upstream packet loss rate

� Roaming time

� Downstream packet loss rate

Packet loss rate (PLR) affects the intelligibility of voice. Packet delay can affect both the intelligibility and conversational quality of the connection. Excessive roaming time produces undesired gaps in audio.

www.CareerCert.info



There are three levels of measurement:

� Normal: Normal QoS (green)

� Fair: Fair QoS (yellow)

� Degraded: Degraded QoS (red)

System administrators should employ some judgment when setting the green, yellow, and red alarm levels. Some factors to consider are:

� Environmental factors including interference and radio coverage can affect PLR.

� End-user expectations and system administrator requirements for audio quality on mobile devices (which have lower audio quality) can permit greater PLR.

� Different codec types used by the phones have different tolerance for packet loss.

� Not all calls will be mobile-to-mobile; therefore, some have less stringent PLR requirements for the wireless LAN.

www.CareerCert.info





Summary

� A QoS profile can be assigned to WLANs from the Configure WLAN QoS tab.

� Each QoS profile is defined by default following the 802.11e andWMM specifications; it can be changed if needed.

� Specific voice parameters such as wireless CAC, expedited bandwidth, and TSM can be globally configured for all WLANs.

� Adapted EDCA parameters can be set for specific client types.

� Temporary WLAN users can be assigned a specific bandwidth through the guest user role.

� Most QoS parameters can be deployed via templates from the Cisco WCS.

References For additional information, refer to these resources:

� Cisco WLC configuration guide: http://www.cisco.com/en/US/products/ps6366/products_installation_and_configuration_guides_list.html

� Cisco WCS configuration guide: http://www.cisco.com/en/US/products/ps6305/products_installation_and_configuration_guides_list.html

www.CareerCert.info


Lesson 4

Configuring the Wired Infrastructure for QoS

Overview After quality of service (QoS) mechanisms are applied to the wireless infrastructure, it is crucial that the wired infrastructure follow the same logic and global QoS strategy. Complex deployments require QoS expertise; however, a voice-over-WLAN (VoWLAN) specialist should be able to recognize the main queuing mechanisms and verify some basic principles. This lesson will give you the basic tools needed to configure QoS parameters on Cisco infrastructure devices.

Objectives Upon completing this lesson, you will be able to configure the wired infrastructure for QoS. This ability includes being able to meet these objectives:

� Use the modular QoS CLI

� Describe the modular QoS CLI components

� Describe class maps

� Configure and monitor class maps

� Configure and monitor policy maps

� Attach service policies to interfaces

� Configure typical QoS queuing mechanisms

� Describe QoS on a switch

� Configure QoS on a switch

� Monitor QoS on a switch

www.CareerCert.info



Modular QoS CLI This topic describes how to implement a given QoS policy using Modular QoS CLI (MQC).


Modular QoS CLI

� The MQC provides a modular approach to configuration of QoS mechanisms.

� First, build modules defining classes of traffic.

� Then, build modules defining QoS policies and assign classes to policies.

� Finally, assign the policy modules to interfaces.

The MQC was introduced to allow the configuring of any supported classification with any QoS mechanism using the standard Cisco IOS command-line interface (CLI).

The separation of classification from the QoS mechanism allows newer Cisco IOS versions to introduce new QoS mechanisms and reuse all available classification options. On the other hand, old QoS mechanisms can benefit from new classification options.

Another important benefit of the MQC is the reusability of configuration. MQC lets you apply the same QoS policy to multiple interfaces. The MQC, therefore, is a consolidation of all the QoS mechanisms that have so far been available only as standalone mechanisms.

www.CareerCert.info



Modular QoS CLI Components This topic describes the three steps involved in implementing a QoS policy using MQC and differentiates between class maps, policy maps, and service policies.


Modular QoS CLI Components

Implementing QoS by using the MQC consists of three steps:

Step 1 Configure classification by using the class-map command.

Step 2 Configure traffic policy by associating the traffic class with one or more QoS features by using the policy-map command.

Step 3 Attach the traffic policy to inbound or outbound traffic on interfaces, subinterfaces, or virtual circuits by using the service-policy command.

Example: Configuring MQC Consider the following example of configuring MQC on a network with voice telephony:

Step 1 Classify traffic as voice, high priority, low priority, and browser by using class maps.

Step 2 Build a single policy map that defines three different traffic policies (different bandwidth and delay requirements for each traffic class): “NoDelay,” “BestService,” and “Whenever,” and assign the already defined classes of traffic to the policies. Voice is assigned to “NoDelay.” High-priority traffic is assigned to “BestService.” Both low-priority and browser traffic is assigned to “Whenever.”

Step 3 Assign the policy map to selected router and switch interfaces.

www.CareerCert.info



Class Maps This topic describes how a class map is used to define a class of traffic.


Class Maps

� “What traffic do we care about?”

� Each class is identified using a class map.

� A traffic class contains three major elements:

– A case-sensitive name

– A series of match commands

– If more than one match command exists in the traffic class, an instruction on how to evaluate these match commands

� Class maps can operate in two modes:

– Match all: All conditions must succeed

– Match any: At least one condition must succeed

� The default mode is match all.

� Multiple traffic classes can be configured as a single traffic class (nested).

Use class maps to create classification templates that are later used in policy maps where QoS mechanisms are bound to the traffic classes.

You can configure routers with a large number of class maps (currently limited to 256).

Create a class map by using the class-map global configuration command. Class maps are identified by case-sensitive names. Each class map contains one or more conditions that determine if the packet belongs to the class.

There are two ways of processing conditions when there is more than one condition in a class map:

� Match all: All conditions have to be met to bind a packet to the class.

� Match any: At least one condition has to be met to bind the packet to the class.

The default match strategy of class maps is match all.

www.CareerCert.info



Configure and Monitor Class Maps This topic describes the Cisco IOS MQC commands required to configure and monitor a class map.


Configuring Class Maps

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

router(config)#

match condition

router(config-cmap)#

description description


� Enter the class-map configuration mode.

� Specify the matching strategy.

� Match-all is the default matching strategy.

� Use at least one condition to match packets.

� You should use descriptions in large and complex configurations.

� The description has no operational meaning.

Use the class-map global configuration command to create a class map and enter the class-map configuration mode. A class map is identified by a case-sensitive name; therefore, all subsequent references to the class map must use exactly the same name.

At least one match command should be used within the class-map configuration mode (match none is the default).

Use the description command for documenting a comment about the class map.

Example: Class Map Configuration The following example shows a traffic class configured with the class-map match-all command:

Router(config)# class-map match-all cisco1 Router(config-cmap)# match protocol ip Router(config-cmap)# match qos-group 4 Router(config-cmap)# match access-group 101

If a packet arrives on a router with a traffic class called cisco1 configured on the interface, the packet is evaluated to determine if it matches the IP protocol, qos-group 4, and access-group 101. If all three of these match criteria are met, the packet matches traffic class cisco1.

www.CareerCert.info




Configuring Classification UsingSpecial Options

match not condition


match class-map class-map-name


match any


� The not keyword inverts the condition.

� One class map can use another class map for classification.

� Nested class maps allow generic template class maps to be used in other class maps.

� The any keyword can be used to match all packets.

Use the match commands to specify various criteria for classifying packets. Packets are checked to determine whether they match the criteria specified in the match commands. If a packet matches the specified criteria, that packet is considered a member of the class and is forwarded according to the QoS specifications set in the traffic policy.

Packets that fail to meet any of the matching criteria are classified as members of the default traffic class called “class-default.” The MQC does not necessarily require that users associate a single traffic class to one traffic policy. Multiple types of traffic can be associated with a single traffic class by using the match any command.

The match not command inverts the condition specified. This command specifies a match criterion value that prevents packets from being classified as members of a specified traffic class. All other values of that particular match criterion belong to the class.

The MQC allows multiple traffic classes (nested traffic classes, which are also called nested class maps) to be configured as a single traffic class. To achieve this nesting, use the match class-map command. The only method of combining match-any and match-all characteristics within a single traffic class is with the match class-map command.

www.CareerCert.info




Configuring Classification UsingSpecial Options (Cont.)

match access-groupmatch cos | precedence | dscp | qos-group | flowmatch source address mac | input-interfacematch destination addressmatch vlanmatch fr-de | fr-dlcimatch ipmatch protocolmatch packet (length)match mpls


Use the match commands with several other condition types:

The match access-group command specifies a numbered or named access control list (ACL) whose contents are used as the match criteria against which packets are checked to determine if they belong to the class specified by the class map.

To match a packet based on a Layer 2 class of service (CoS) or Inter-Switch Link (ISL) marking, use the match cos command in class-map configuration mode. You can also configure a QoS policy to include IP precedence marking for packets entering the network with match precedence. To identify one or more differentiated services code point (DSCP), Assured Forwarding (AF), and Certificate Server (CS) values as a match criterion, use the match dscp command in class-map configuration mode.

To use the source MAC address as a match criterion, use the match source-address mac command in QoS class-map configuration mode. To configure a class map to use the specified input interface as a match criterion, use the match input-interface.

To configure a class map to use the destination IP address as a match criterion, use the match destination-address.

You can also use the VLAN as a classification criterion using the match VLAN command.

Several conditions can be set for Frame Relay traffic, such as traffic having the discard eligibility bit set to 1, with match fr-de, or the data-link connection identifier (DLCI) on which the traffic is received or sent, with match fr-dlci.

The match ip condition accepts three possible criteria:

• Real-Time Transport Protocol (RTP) (match ip rtp <port number> to identify a voice traffic type by its port, such as H.323 or Media Gateway Control Protocol [MGCP])

www.CareerCert.info



• match ip precedence, replaced with match precedence

• match ip dscp, replaced by match dscp.

When you know the protocol, you can use the condition match protocol <protocol_name>.

You can also use the packet size to identify a type of traffic, with match packet length min | max (value in bytes).

In Multiprotocol Label Switching (MPLS) networks, you can match traffic matching a specific label value, with match mpls experimental topmost <value>.

Example: Using the match Command The following example shows a traffic class configured with the class-map match-any command:

Router(config)# class-map match-any cisco2 Router(config-cmap)# match ip dscp 46 Router(config-cmap)# match ip precedence 5 Router(config-cmap)# match qos-group 4 Router(config-cmap)# match access-group 101

In traffic class called cisco2, the match criteria are evaluated consecutively until a successful match is located. The packet is first evaluated to determine whether ip dscp 46 can be used. If ip dscp 46 is not a successful match, then ip precedence 5 is evaluated. If ip precedence 5 is not a successful match, then qos-group 4 is evaluated. If qos-group 4 is not a successful match, then access-group 101 is evaluated. Each line is evaluated in order to see if the packet matches that criterion. When a successful match occurs, the packet is classified as a member of traffic class cisco2. If the packet matches none of the specified criteria, the packet is classified as a member of the “class-default” traffic class.

Example: Nested Traffic Class to Combine match-any and match-all Characteristics in One Traffic Class

The only method of including both match-any and match-all characteristics in a single traffic class is to use the match class-map command. To combine match-any and match-all characteristics into a single class, a traffic class created with the match-any instruction must use a class configured with the match-all instruction as a match criterion (through the match class-map command), or vice versa. The following example shows how to combine the characteristics of two traffic classes, one with match-any and one with match-all characteristics, into one traffic class with the match class-map command. The result of traffic class class4 requires a packet to match one of the following three match criteria to be considered a member of traffic class class4:

(IP protocol and qos-group 4), or destination MAC address 1.1.1, or access-group 2.

In this example, only the traffic class called class4 is used with the traffic policy called policy1:

Router(config)# class-map match-all class3 Router(config-cmap)# match protocol ip Router(config-cmap)# match qos-group 4 Router(config-cmap)# exit Router(config)# class-map match-any class4 Router(config-cmap)# match class-map class3 Router(config-cmap)# match destination-address mac a.b.c.d.e.f Router(config-cmap)# match access-group 2

www.CareerCert.info



Router(config-cmap)# exit Router(config)# policy-map policy1 Router(config-pmap)# class class4 Router(config-pmap-c)# police 8100 1500 2504 conform-action transmit exceed-action set-cos-transmit 4 Router(config-pmap-c)# exit

www.CareerCert.info




Monitoring Class Maps

� Displays all class maps and their matching criteria.

show class-map [class-name]

router>

router>show class-map

Class Map class-3 Match access-group 103

Class Map class-2 Match ip dscp 46

Class Map class-1 Match input-interface Ethernet1/0

The show class-map command lists all class maps with their match statements.

The show class-map command with a name of a class map displays the configuration of the selected class map.

The example of show class-map in the illustration shows three class maps:

� class-3 will match any packet to access-group 103.

� class-2 matches IP packets with ip dscp 46.

� class-1 matches any input from interface Ethernet 1/0.

www.CareerCert.info



Policy Maps This topic describes how a policy map is used to assign a QoS policy to a class of traffic.


Policy Maps

� “What will be done to this traffic?”

� Defines a traffic policy, which configures the QoS features associated with a traffic class previously identified using a class map.

� A traffic policy contains three major elements:

– A case-sensitive name

– A traffic class

– The QoS policy associated with that traffic class

� Up to 256 traffic classes can be associated with a single traffic policy.

� Multiple policy maps can be nested to influence the sequence of QoS actions.

Use the policy-map command to create a traffic policy. The purpose of a traffic policy is to configure the QoS features that should be associated with the traffic that has been classified in a user-specified traffic class or classes. A traffic policy contains three elements: a case-sensitive name, a traffic class (specified with the class command), and the QoS policies.

The name of a traffic policy is specified in the policy-map CLI command (for example, issuing the policy-map class1 command would create a traffic policy named class1). After the policy-map command is issued, the user is placed into policy-map configuration mode. You can enter the name of a traffic class, and then you can configure the QoS features to apply to the traffic that matches this class.

The MQC does not necessarily require that users associate only one traffic class to a single traffic policy. When packets match to more than one match criterion, multiple traffic classes can be associated with a single traffic policy.

Note A packet can match only one traffic class within a traffic policy. If a packet matches more

than one traffic class in the traffic policy, the first traffic class defined in the policy will be used.

www.CareerCert.info



Configure and Monitor Policy Maps This topic describes the Cisco IOS MQC commands required to configure and monitor a policy map.


Configuring Policy Maps

policy-map policy-map-name

router(config)#

class {class-name | class-default}

router(config-pmap)#

class class-map-name condition


� Enter policy-map configuration mode.

� Policy maps are identified by a case-sensitive name.

� Enter the per-class policy configuration mode by using the name of a previously configured class map.

� Use the name “class-default” to configure the policy for the default class.

Configure service policies by using the policy-map command. You can use up to 256 classes within one policy map by using the class command with the name of a preconfigured class map.

The table shows starting and resulting configuration modes for the class-map, policy-map, and class commands:

Configuration Modes

Starting Configuration Mode Command Configuration Mode

Router(config)# class-map Router(config-cmap)#

Router(config)# policy-map Router(config-pmap)#

Router(config-pmap)# class Router(config-pmap-c)#

All traffic that is not classified by any of the class maps that are used within the policy map is part of the default class “class-default.” This class has no QoS guarantees, by default. The default class, when used in the output direction, can use one FIFO queue or flow-based weighted fair queuing (WFQ), or class-based weighted fair queuing (CBWFQ). The default class is part of every policy map, even if not configured.

www.CareerCert.info




Configuring Policy Maps (Cont.)

description description


<Per Hop Behaviour mechanism>

router(config-pmap-c)#

� You should use descriptions in large and complex configurations.

� The description has no operational meaning.

� Per-class service policies are configured within the per-class policy-map configuration mode.

� Some of the QoS mechanisms supported by MQC include:

– Class-based weighted fair queuing (CBWFQ)

– Low latency queuing

– Class-based policing

– Class-based shaping

– Class-based marking

Policy maps, like class maps, should use descriptions in large QoS implementations where a large number of different policy maps are used.

Renaming a policy map would normally require the renaming of all the references to the policy map. Using the rename command simplifies the renaming process by automatically renaming all references.

Example: Policy Map Example The example shows the configuration of a policy map using three classes. The first two classes were separately configured using the class-map command. The third class was configured by specifying the match condition after the name of the class:

class-map match-all Test1 match protocol http match access-group 100 class-map match-any Test2 match protocol http match access-group 101 ! policy-map Test class Test1 bandwidth 100 class Test2 bandwidth 200 ! access-list 100 permit tcp any host 10.1.1.1 access-list 101 permit tcp any host 10.1.1.2

Class Test1 has one match conditions evaluated in the default match-all strategy. Classes Test2 has two conditions, evaluated independently with the match-any strategy.

www.CareerCert.info



Example: Policy Map In the example diagram, a policy is created to prioritize voice traffic.


Policy Map Example

Example policy:� Gives strict priority (LLQ) to voice traffic.

� Voice signaling gets 10% of the bandwidth.

� Other traffic has weighted fair queue.

class-map VOICEmatch protocol rtp audio!class-map match-any SIGNALINGmatch protocol sccpmatch protocol h323match protocol sip!class-map AllTrafficmatch any!

policy-map QueueAllclass VOICEset ip dscp efpriority percent 45

!class SIGNALINGset ip dscp CS3bandwidth percent 10

!class AllTrafficfair-queue

You can recognize voice traffic by its protocol, RTP audio. You can recognize signaling by its type, which can be Skinny Client Control Protocol (SCCP), H.323, or Session Initiation Protocol (SIP). Voice affects a strict priority over the other traffic, with 45 percent of the interface available bandwidth and will be marked with a DSCP of Expedited Forwarding (EF). Signaling receives 10 percent of the interface available bandwidth, but without strict priority, and will be marked with a DSCP of CS3. The rest of the traffic will be matched to the AllTraffic class and will use flow-based weighted fair queuing. In this example, no traffic will be matched to the class-default traffic class.

When you use the bandwidth keyword, you can assign a percentage of the available bandwidth on the interface (bandwidth percent 45). You can also assign a bandwidth in kilobits (bandwidth 80) or based on the remaining available bandwidth (bandwidth remaining percent 45). In this last case, you allocate bandwidth based on what is available, not on the total interface bandwidth.

Note This example does not take into account a possible already existing tag on the frames.

Traffic is recognized by its type, not by an already existing tag. This means that the router on which this policy is configured is the trust boundary. Tags on traffic entering this router are not trusted. The router classifies and tags frames whether an already existing tag can be

seen or not.

www.CareerCert.info




Monitoring Policy Maps

� Displays the configuration of all classes for a specified service policy map or of all classes for all existing policy maps.

show policy-map [policy-map]

router>

router>show policy-mapPolicy Map Test

Class Test1Strict Priority

Bandwidth 100 (kbps) Max Threshold 64 (packets)Class Test2

Weighted Fair QueueingBandwidth 200 (kbps) Max Threshold 64 (packets)

Class Test3Weighted Fair Queueing

Bandwidth 300 (kbps) Max Threshold 64 (packets)

You can use the show policy-map command to verify the configuration of a policy map.


Monitoring Policy Maps (Cont.)

show policy-map interface interface-name [input | output]

router>

router>show policy-map interface FastEthernet0/0 outputFastEthernet0/0

Service-policy output: Test (1101)

Class-map: Test1 (match-any) (1103/3)0 packets, 0 bytes5 minute offered rate 0 bps, drop rate 0 bpsMatch: access-group 101 (1107)Match: access-group 102 (1111)Match: protocol http (1115)Weighted Fair Queueing

Output Queue: Conversation 265Bandwidth 100 (kbps) Max Threshold 64 (packets)(pkts matched/bytes matched) 0/0(depth/total drops/no-buffer drops) 0/0/0

...Class-map: class-default (match-any) (1143/0)

25 packets, 19310 bytes5 minute offered rate 1000 bps, drop rate 0 bpsMatch: any (1147)

The show policy-map command also displays live information if the interface keyword is used. The sample output shows the parameters and statistics of the policy map that is attached to outbound traffic on interface FastEthernet0/0.

www.CareerCert.info



Service Policy This topic describes how a service policy is assigned to an interface.


Service Policy

� “Where will this policy be implemented?”

� Attaches a traffic policy configured with a policy map to an interface.

� Service policies can be applied to an interface for inbound or outbound packets.

The last configuration step when configuring QoS mechanisms using the MQC is to attach a policy map to the inbound or outbound packets using the service-policy command.

Using the service-policy command, you can assign a single policy map to multiple interfaces or assign multiple policy maps to a single interface (a maximum of one in each direction, inbound and outbound). You can apply a service policy for inbound or outbound packets.

www.CareerCert.info



Attach Service Policies to Interfaces This topic describes the MQC commands used to attach a service policy to an interface.


Attaching Service Policies to Interfaces

� Attaches the specified service policy map to the input or output interface.

service-policy {input | output} policy-map-name

router(config-if)#

class-map HTTPmatch protocol http!policy-map PMclass HTTPbandwidth 2000class class-defaultbandwidth 6000

!

interface fastethernet0/1service-policy output PM!

Use the service-policy interface configuration command to attach a traffic policy to an interface and to specify the direction in which the policy should be applied (either on packets coming into the interface or packets leaving the interface).

The router immediately verifies the correctness of parameters that are used in the policy map. If there is a mistake in the policy map configuration, the router will display a message explaining what is wrong with the policy map.

The sample configuration shows how to use a policy map to separate HTTP from other traffic. HTTP is guaranteed 2 Mb/s. All other traffic belongs to the default class and is guaranteed 6 Mb/s.

www.CareerCert.info




MQC Configuration Example

class-map class1match access-group 101!class-map class2match access-group 102!policy-map policy1class class1bandwidth 3000queue-limit 30class class2bandwidth 2000

!interface e1/1service-policy output policy1!interface fa1/0/0service-policy output policy1

Class Definition

Policy Definition

Policy Applied to Interfaces

Example: Complete MQC Configuration Traffic Classes Defined

In the following example, two traffic classes are created and their match criteria are defined. For the first traffic class, called class1, access control list (ACL) 101 is used as the match criterion. For the second traffic class, called class2, ACL 102 is used as the match criterion. Packets are checked against the contents of these ACLs to determine if they belong to the class:

Router(config)# class-map class1 Router(config-cmap)# match access-group 101 Router(config-cmap)# exit

Router(config)# class-map class2 Router(config-cmap)# match access-group 102 Router(config-cmap)# exit

Traffic Policy Created

In the following example, a traffic policy called policy1 is defined to contain policy specifications for the two classes—class1 and class2. The match criteria for these classes were defined in the traffic classes.

For class1, the policy includes a bandwidth allocation request and a maximum packet count limit for the queue reserved for the class. For class2, the policy specifies only a bandwidth allocation request:

Router(config)# policy-map policy1 Router(config-pmap)# class class1 Router(config-pmap-c)# bandwidth 3000 Router(config-pmap-c)# queue-limit 30 Router(config-pmap-c)# exit Router(config-pmap)# class class2 Router(config-pmap-c)# bandwidth 2000 Router(config-pmap-c)# exit

www.CareerCert.info



Traffic Policy Attached to an Interface

The following example shows how to attach an existing traffic policy (which was created in the preceding section) to an interface. After you define a traffic policy with the policy-map command, you can attach the traffic policy to one or more interfaces to specify the traffic policy for those interfaces by using the service-policy command in interface configuration mode. Although you can assign the same traffic policy to multiple interfaces, each interface can have only one traffic policy attached at the input and a single traffic policy attached at the output:

Router(config)# interface e1/1 Router(config-if)# service-policy output policy1 Router(config-if)# exit Router(config)# interface fa1/0/0 Router(config-if)# service-policy output policy1 Router(config-if)# exit

www.CareerCert.info



Policy Map Examples This topic provides several simple examples of policy maps.


Policy Map Examples

interface f0/0

fair-queue

!

policy-map policy2

class class2bandwidth percent 30class class3bandwidth percent 40

class class-default

bandwidth percent 30!

policy-map policy3

class class4

priority percent 30

class class5

bandwidth percent 40

Class-Based Weighted Fair Queue

Weighted Fair Queue

Low Latency Queue

Weighted Fair Queue:

Weighted Fair Queue prioritizes “quiet” traffic over “heavy” traffic. In other words, in a queue where a few small Telnet packets compete with many large FTP packets for bandwidth, the WFQ system will prioritize Telnet packets. For serial interfaces at E1 (2.048 Mb/s) and below, WFQ is used by default. For other interfaces, simply apply WFQ to the interface by entering:

interface f0/0 fair-queue

Class Based Weighted fair Queue:

CBWFQ is a type of fair queue in which the administrator can create classes and apply the WFQ on these classes; thus allocating a minimum bandwidth to guarantee chosen traffic. The following bandwidth command applies CBWFQ queuing to several classes:

policy-map policy2 class class2 bandwidth percent 30 class class3 bandwidth percent 40 class class-default bandwidth percent 30

www.CareerCert.info



Class 2 receives a guaranteed minimum bandwidth equal to 30 percent of the interface available bandwidth, class 3 receives a guaranteed minimum bandwidth equal to 40 percent of the interface available bandwidth, and the rest of the traffic receives a guaranteed minimum bandwidth equal to the remaining 30 percent of the interface available bandwidth.

Practically, the router determines time intervals on the interface and sends the relevant proportion of each type of traffic during these time intervals. For example, suppose that all packets are of the same size, and that the router uses ten time intervals per second. Each time interval allows sending 100 packets. The router will send 30 packets of class 2 traffic, 40 packets of class 3 packets, and then 5 packets of other packets before moving to the next interval and sending 30 packets of class 2 traffic, and so on.

By default, it is recommended that you should reserve 25 percent of the interface bandwidth for high-priority system traffic (such as routing protocol traffic) and for the class-default class. The LLQ/CBWFQ traffic classes should not reserve this 25 percent reserved bandwidth. Therefore, by default, the total available bandwidth on an interface is equal to 75 percent of interface bandwidth. You can use the max-reserved-bandwidth interface command to change the default reserved bandwidth on an interface.

Low Latency Queuing

Low Latency Queuing is a mix of priority queue and class-based weighted fair queue. One type of traffic has absolute priority over the others. After the prioritized traffic queue has emptied, the other types of traffic are served in a CBWFQ manner. As soon as some prioritized traffic arrives in the buffer, it is sent immediately. For example:

Policy-map policy3 class class4 priority percent 30 class class5 bandwidth percent 40 class class-default bandwidth percent 30

In this example, class 4 is given absolute priority. When the class 4 traffic buffer is empty, class 5 and all the rest are served in a CBWFQ manner. Class 4 traffic gets a maximum guarantee rate, which is 30 percent of the interface available bandwidth. Class 5 traffic gets a minimum guarantee rate, which is 40 percent of the interface available bandwidth. Class class-default traffic gets a minimum guarantee rate, which is 5 percent of the interface available bandwidth.

www.CareerCert.info




Policy Map Examples (Cont.)policy-map policy1

class class1

random-detect

!

policy-map policy2

class class2

police 80000

!

policy-map policy3

class class3

shape average 80000

!

interface serial 1/0

ip rtp header-compression

!

…/…

ppp multilink interleave

ppp multilink fragment-delay 20

Congestion Avoidance

Policing

Shaping

Compression

Fragmentation and Interleaving

Congestion avoidance:

Congestion avoidance acts by dropping packets before congestion occurs. It is adapted to TCP traffic. To configure congestion avoidance, use the following random-detect commands:

Policy-map policy1 class class1 random-detect

When class1 type of traffic fills the buffer, packets will start to drop randomly to activate the TCP window sliding mechanism and slow the traffic incoming throughput. This command has several variations, such as random-detect dscp based, to drop less important traffic first. Refer to the Cisco IOS MQC documentation for all the random-detect command options.

Policing

Policing allows a certain rate to a type of traffic. Traffic exceeding the rate can be dropped, re-marked, or transmitted. To configure policing, use the following police commands:

Policy-map policy2 class class2 police 80000

Class2 traffic exceeding 80 kb/s at the interface level is dropped. Refer to the Cisco IOS MQC documentation for all the police command options.

Shaping:

Shaping allows a certain rate to a type of traffic. Traffic exceeding the rate is buffered. If the network usage for this type of traffic reduces, the buffered traffic is sent. To configure shaping, use the following shape commands:

Policy-map policy3 class class3 shape average 80000

www.CareerCert.info



Class3 traffic exceeding 80 kb/s at the interface level is buffered if possible. Refer to the Cisco IOS MQC documentation for all the shape command options.

Link efficiency mechanisms

Link efficiency mechanisms include compression and link fragmentation and interleaving. Link efficiency mechanisms are usually reserved for links with 768 kb/s or less bandwidth. There are several types of compression (header compression, payload compression). The following class-based RTP header compression example compresses the RTP header for the voice traffic class. This is useful for voice traffic over a point-to-point link to reduce the header overhead:

Router(config)# policy-map policy1 Router(config-pmap)# class voice Router(config-pmap-c)# compression header ip rtp

All voice traffic class traffic will have its RTP header compressed.

Link fragmentation and interleaving (LFI) defines a maximum frame size on a link and fragments any frame larger than the threshold. The following example enables this feature on a point-to-point multilink PPP link:

ppp multilink interleave ppp multilink fragment-delay 20

The interleave command enables the LFI feature. It is possible to refine the command by defining the fragment threshold. This threshold is based on the time taken to send a frame on the interface. The router knows the frame size and speed of the link. It determines the time required to send any incoming frame on the interface. In this example, any frame that would take more than 20 milliseconds to be sent has to be fragmented.

Note All the preceding examples aim at helping you recognize the different types of QoS features

available on the Cisco IOS CLI. Each of these types has several variations not covered in this course.

www.CareerCert.info



QoS on a Switch This topic describes basic QoS policies on a switch.


QoS on Switches: Voice VLANs

Switches can be Layer 2 or Layer 3 and can have a limited set of features or extended possibilities. They can be configured as trust boundaries and perform full traffic identification, classification, and marking. When connecting to controllers and APs, switches receive already tagged packets and frames. Wired IP phones also send tagged frames. By default, the IP phone sends IEEE 802.1p-tagged packets with the CoS set to a value of 5 for voice bearer traffic and 3 for voice signaling traffic.

Because most PCs do not have an 802.1Q-capable network interface card (NIC), PCs send the packets untagged. This means that the frames do not have an 802.1p field. Also, unless the applications running on the PC send packets with a specific CoS value, this CoS field is not present.

Even if the PC is sending tagged frames with a specific CoS value, Cisco IP phones can zero out this value before sending the frames to the switch. This is the default behavior. Voice frames coming from the IP phone have a CoS of 5 and data frames coming from the PC do not have a CoS value. When the switch receives these frames, the switch can take into account these values for further processing based on its capabilities. For example, the switch can be configured to trust the CoS setting from the IP phone and maps the trusted CoS to the corresponding DSCP value for output queuing.

www.CareerCert.info




QoS on Switches: Trust Policies

� Port can be configured to trust CoS, DSCP, or Cisco IP phone (default = untrusted).

� Has default CoS-to-DSCP and DSCP-to-CoS maps.

� Can set the default CoS by port.

� Can use class-based marking to set DSCP.

Cisco Catalyst switches offer superior and highly granular QoS based on Layer 2 through Layer 4 information to ensure that network traffic is classified and prioritized and that congestion is avoided in the best possible manner.

You can configure Cisco switches to trust the incoming CoS, the incoming DSCP, or recognize traffic coming from a Cisco IP phone. After you define the trust policy, the switch can perform a default mapping between type of service (ToS) and CoS, or CoS and DSCP. In other words, if you decide to trust the CoS value on a given port, you can also configure the switch to determine and fix the DSCP value read in the Layer 3 section of the frame. This allows you to determine a Layer 2 to Layer 3 QoS map and apply it to all packets received on the switch.

Cisco switches can also classify, reclassify, police (determine if the packet is in or out of predetermined profiles and affect actions on the packet), and mark or drop incoming packets before the packets are placed in the shared buffer. Packet classification allows the network elements to discriminate between various traffic flows and enforce policies based on Layer 2 and Layer 3 QoS fields.

Use the Layer 2 frame information to carry out classification:

� Prioritization values in Layer 2 frames:

— Layer 2 802.1Q frame headers are used in trunks, except for native VLAN frames.

— Other non-802.1Q frame types cannot carry Layer 2 CoS values.

Use the Layer 3 packet information to carry out classification:

� Prioritization bits in Layer 3 packets:

— Layer 3 IP packets with DSCP.

www.CareerCert.info




Mapping Tables

� During QoS processing, the switch represents the priority of all traffic (including non-IP traffic) with an internal DSCP value.

� During classification, QoS uses configurable mapping tables to derive the internal DSCP (a 6-bit value) from received CoS value.

� Before the traffic reaches the scheduling stage, QoS uses the configurable DSCP-to-CoS map to derive a CoS value from the internal DSCP value.

During QoS processing, the switch represents the priority of all traffic (including non-IP traffic) with an internal DSCP value. During classification, if you configure the switch to trust the incoming CoS value, the switch will use the configurable CoS-to-DSCP mapping tables to derive the internal DSCP (a 6-bit value) from the received CoS value.

On an ingress interface configured to trust the incoming DSCP value (instead of trusting the incoming CoS), if the DSCP values are different between the QoS domains, you can apply the configurable DSCP-to-DSCP-mutation map to the interface that is on the boundary between the two QoS domains.

The CoS-to-DSCP, DSCP-to-CoS, and the IP-precedence-to-DSCP maps have default values that might or might not be appropriate for the network.

Actions at the egress interface include queuing and scheduling:

� Queuing evaluates the internal DSCP and determines which of the egress queues should be used for placing the packet. The DSCP value is mapped to a CoS value, which selects one of the egress queues.

� Scheduling services the egress queues based on their configured weighted round-robin (WRR) weights and thresholds. One of the queues can be the expedite queue, which is serviced until empty before the other queues are serviced. Congestion avoidance techniques include tail drop and weighted random early detection (WRED) on Gigabit-capable Ethernet ports and tail drop (with only one threshold) on 10/100 Ethernet ports.

The maps like CoS-to-DSCP, DSCP-to-CoS, and so on, have default values that might or might not be appropriate for the network. If these values are not appropriate for the network, you can modify the maps by using the switch CLI.

www.CareerCert.info



Note Capabilities vary depending on the Cisco switch hardware platform and software images.

Refer to the Cisco Catalyst Switch QoS configuration guide. Refer to the Cisco Catalyst Switch QoS configuration guide relevant to your specific platform. For example, for the Cisco Catalyst 3750 model, refer to http://www.cisco.com/en/US/docs/switches/lan/catalyst3750/software/release/12.2_44_se/configuration/guide/swqos.html.

www.CareerCert.info



Configure QoS on a Switch This topic describes how to configure QoS policies on a switch.


Configuring Classification and Marking on a Switch

mls qos trust [cos [pass-through dscp] | device cisco-phone | dscp]

Switch(config-if)#

mls qos cos {default-cos | override}

Switch(config-if)#

� Configures the port to trust state on an interface.

� When a port is configured with trust DSCP and the incoming packet is a tagged non-IP packet, the CoS value for the packet is set to 0, and the DSCP-to-CoS map is not applied.

� If DSCP is trusted, the DSCP field of the IP packet is not modified, but it is still possible that the CoS value of the packet is modified according to the DSCP-to-CoS map.

� Defines the default CoS value of a port or assigns the default CoS to all incoming packets on the port.

Example Default Standard 2960 QoS Configuration

QoS is disabled. There is no concept of trusted or untrusted ports because the packets are not modified (the CoS, DSCP, and IP precedence values in the packet are not changed). Traffic is switched in pass-through mode (packets are switched without any rewrites and are classified as best effort without any policing).

This figure shows two of the QoS configuration commands that are available on the Catalyst switches. The defaults for its interfaces are as follows:

mls qos trust [cos [pass-through dscp] | device cisco-phone | dscp]

On the Catalyst 2960 Series Switches, QoS is enabled as soon as you enter the mls qos trust command. On some other switches, you may need to enable QoS support first by entering, in global configuration mode:

mls qos

www.CareerCert.info



Syntax Description

Parameter Description

cos (Optional) Specifies that the CoS bits in incoming frames are trusted and derives the internal DSCP value from the CoS bits.

pass-through dscp (Optional) Configures the interface to classify ingress packets by trusting the CoS value and to send packets without modifying the DSCP value (pass-through mode).

device cisco-phone (Optional) Classifies ingress packets by trusting the value sent from the Cisco IP phone (trusted boundary).

dscp (Optional) Classifies ingress packets with packet DSCP values (most significant 6 bits of the 8-bit service-type field). For non-IP packets, the packet CoS value is set to 0. This keyword is available only if your switch is running the enhanced image (EI) software.

To define the default CoS value for an interface, use the mls qos cos interface configuration command. Use the no form of this command to remove a prior entry. QoS assigns the CoS value specified with this command to untagged frames received on trusted and untrusted ports. The default CoS value is 0.

mls qos cos cos-value

Syntax Description


cos-value Default CoS value for the interface; valid values are from 0 to 7.

Note Capabilities vary depending on the Cisco switch hardware platform and software images.

Refer to the Cisco Catalyst Switch QoS configuration guide. Refer to the Cisco Catalyst Switch QoS configuration guide relevant to your specific platform. For example, for the Cisco Catalyst 3750 model, refer to http://www.cisco.com/en/US/docs/switches/lan/catalyst3750/software/release/12.2_44_se/configuration/guide/swqos.html

www.CareerCert.info




Configuring Classification and Marking on a Switch (Cont.)

mls qos map cos-dscp dscp1...dscp8

Switch(config)#

mls qos map dscp-cos dscp-list to cos

Switch(config)#

� Defines the CoS-to-DSCP mapping.

� For dscp1...dscp8, enter eight DSCP values that correspond to CoS values 0 to 7. Separate each DSCP value with a space.

� Defines the DSCP-to-CoS mapping.

� For dscp-list, enter DSCP values separated by spaces. Then, enter the to keyword.

� For cos, enter the CoS value to which the DSCP values correspond. The CoS range is 0 to 7.

The commands listed in this figure show how to change the default CoS-to-DSCP and DSCP-to-CoS mappings.

CoS-to-DSCP Default Mapping

Marker Value

CoS Values 0 1 2 3 4 5 6 7

DSCP Values

0 8 16 24 32 40 48 56

To define the ingress CoS-to-DSCP mapping for trusted interfaces, use the mls qos map cos-dscp command. Use the CoS-to-DSCP map to map the CoS of packets arriving on trusted interfaces (or flows) to a DSCP where the trust type is trust-cos. This map is a table of eight CoS values (0 through 7) and their corresponding DSCP values. Use the no form of this command to remove a prior entry.

mls qos map cos-dscp values

Syntax Description


values

Eight DSCP values, separated by spaces, corresponding to the CoS values; valid values are from 0 to 63.

www.CareerCert.info



DSCP-to-CoS Default Mapping

Marker Value

DSCP Values

0 8, 10 16, 18 24, 26 32, 34 40, 42 48 56

CoS Values 0 1 2 3 4 5 6 7

To define an egress DSCP-to-CoS mapping, use the mls qos map dscp-cos command. The DSCP-to-CoS map is used to map DSCP values in incoming packets to a CoS value, which is used to select one of the four egress queues. The CoS mapped value is written into the ISL header or 802.1Q tag of the transmitted frame on trunk interfaces. You can enter up to eight DSCP values separated by a space and up to eight CoS values separated by a space. Use the no form of this command to remove a prior entry.

mls qos map dscp-cos dscp-values to cos-values

Syntax Description


dscp-values DSCP values; valid values are from 0 to 63.

to Defines mapping.

cos-values CoS values; valid values are from 0 to 63.

www.CareerCert.info




Configuring Classification and Marking on a Switch (Cont.)

This figure shows a configuration example on a Catalyst switch where the CoS-to-DSCP map has been changed from the default.

The default map:

Marker Value

CoS Values 0 1 2 3 4 5 6 7

DSCP Values

0 8 16 24 32 40 48 56

The map after configuration:

Marker Value

CoS Values 0 1 2 3 4 5 6 7

DSCP Values

0 10 18 26 34 46 48 56

The interface has been set to trust the CoS value using the mls qos trust command, using both the cos and device cisco-phone options. The result of the configuration is that the switch interface trusts CoS only when a Cisco IP phone is attached. The switch uses Cisco Discovery Protocol to detect if a Cisco IP phone is attached and to pass the voice VLAN ID information to the Cisco Ip phone.

The last command in the configuration is the switchport priority extend cos 0 command. Use the switchport priority extend cos 0 interface configuration command to enable the IP phone to override the CoS marking from the PC attached to the IP phone with a CoS value of 0:

switchport priority extend {cos value | trust}

www.CareerCert.info



Syntax Description


cos value

Sets the IP phone port to override the priority received from the PC or the attached device.

The CoS value is a number from 0 to 7. The highest priority is 7. The default is 0.

trust Sets the IP phone port to trust the priority received from PC or the attached device.

www.CareerCert.info




Implementing QoS with AutoQoS

� AutoQoS is available for LAN and WAN routers and switches.

� One command enables Cisco QoS for VoIP on a given port, interface, or PVC.

Using Cisco AutoQoS, network administrators can implement the QoS features that are required for VoIP traffic without in-depth knowledge of the following underlying technologies:

� PPP

� Frame Relay

� ATM

� Service policies

� Link efficiency mechanisms, such as link fragmentation and interleaving (LFI)

The AutoQoS VoIP feature simplifies QoS implementation and speeds up the provisioning of QoS technology over a Cisco network. AutoQoS VoIP also reduces human error and lowers training costs. With the AutoQoS VoIP feature, one command (the auto qos command) enables QoS for VoIP traffic across every Cisco router and switch.

Network administrators can also use existing Cisco IOS commands to modify the configurations that are automatically generated by the AutoQoS VoIP feature in case the default AutoQoS configuration is not sufficient.

www.CareerCert.info



Monitor QoS on a Switch This topic describes commands used to monitor QoS on a switch.


Monitoring QoS on a Switch

show mls qos interface [interface-id] [policers]

Switch>

Switch> show mls qos interface fastethernet0/1

FastEthernet0/1 trust state:trust cos trust mode:trust cos COS override:dis default COS:0 pass-through:none trust device:cisco-phone

� Displays QoS information at the interface level.

After configuring QoS on a Catalyst switch, the network administrator should verify proper operation of QoS and verify that the policies have been configured. In the example, the trust state has been set for CoS, and the default value of CoS is 0.

Use the show mls qos interface user EXEC command to display QoS information at the interface level.

show mls qos interface [interface-id] [policers]

Syntax Description


interface-id (Optional) Displays QoS information for the specified interface.

policers (Optional) Displays all the policers configured on the interface, their settings, and the number of policers unassigned (available only when the switch is running the Extended Feature Set software).

www.CareerCert.info




Monitoring QoS on a Switch (Cont.)

show mls qos maps [cos-dscp | dscp-cos]

Switch>

Switch> show mls qos maps

Dscp-cos map: dscp: 0 8 10 16 18 24 26 32 34 40 46 48 56 -----------------------------------------------cos: 0 1 1 2 2 3 7 4 4 5 5 7 7

Cos-dscp map: cos: 0 1 2 3 4 5 6 7 --------------------------------dscp: 0 8 16 24 32 40 48 56

� Displays QoS mapping information.

Another important monitoring command is the show mls qos maps command, which displays the CoS-to-DSCP and DSCP-to-CoS mappings. Use the maps to generate an internal DSCP value, which represents the priority of the traffic.

show mls qos maps [cos-dscp | dscp-cos]

Syntax Description


cos-dscp (Optional) Displays CoS-to-DSCP map.

dscp-cos (Optional) Displays DSCP-to-CoS map.

www.CareerCert.info





Summary

� MQC is a modular approach to designing and implementing an overall QoS policy.

� Applying an overall QoS policy involves three steps: defining class maps to identify classes of traffic, defining QoS policy maps, and assigning the policy maps to interfaces.

� Each class of traffic is defined in a class map module.

� The class-map global configuration command is used to create a class map and enter the class-map configuration mode. The show class-map command lists all class maps with their match statements.

� A policy map module defines a traffic policy, which configures the QoS features associated with a traffic class previously identified using a class map.


Summary (Cont.)

� The service-policy command assigns a single policy map to multiple interfaces or assigns multiple policy maps to a single interface (a maximum of one in each direction, inbound and outbound).

� The policy-map statement is where the queuing mechanism mainly is defined.

� QoS classification and marking on workgroup switches are based on DiffServ and CoS. There must be mapping between Layer 2 and Layer 3.

� Several types of classification and marking are available on Cisco Catalyst Switches 6500, 4000, 3750, 3500, and 2950.

� Use the show mls qos interface command to display general QoS information.

www.CareerCert.info



www.CareerCert.info


Lesson 5

Understanding Current Best-Practice Guidelines

Overview The previous lessons gave you most of the elements you need to configure quality of service (QoS) in a voice over wireless LAN (VoWLAN) deployment. Some questions still remain that you may be asked when deploying a Control and Provisioning of Wireless Access Points (CAPWAP)-based solution for VoWLAN. Many clients do not understand the benefit of controllers and see the traffic between access points (APs) and controllers as redundant and a source of congestion. This last lesson will give you some best-practice recommendations, show you the CAPWAP traffic pattern, and help you evaluate its effect on the LAN throughput.

Objectives Upon completing this lesson, you will be able to understand the current best practices for QoS over wireless implementations. This ability includes being able to meet these objectives:

� Describe how frame sizes affect throughput on the wireless side

� Configure QoS on a switch linking APs to the controller

� Describe CAPWAP traffic classification

� Evaluate CAPWAP traffic volume and its effect on your network

� Manipulate CAPWAP marking

www.CareerCert.info



Throughput This topic describes how frame sizes affect throughput on the wireless side.


Throughput ConsiderationsDay 1 Day 2 Day 5

300 600 900 1200 1500 Frame Size (Bytes)

11g–54 Mb/s 11.4 19.2 24.6 28.4 31.4 Throughput (Mb/s)

11b–11 Mb/s 2.2 3.6 4.7 5.4 6 Throughput (Mb/s)

An important consideration when deploying IEEE 802.11 QoS is to understand the offered traffic, not only in terms of bit rate, but also in terms of frame size, because 802.11 throughput is sensitive to the frame size of the offered traffic. The figure shows the affect that frame size has on throughput: As packet size decreases, so does throughput. For example, if an application offering traffic at a rate of 3 Mb/s is deployed on an 11-Mb/s 802.11b network but uses an average frame size of 300 bytes, no QoS setting on the AP allows the application to achieve its throughput requirements. This is because 802.11b cannot support the required throughput for that throughput and frame size combination. The same amount of offered traffic, having a frame size of 1500 bytes, does not have this issue.

This is one of the reasons why you should control the number of voice devices in the cell. Voice devices send small packets at a steady pace. In an 802.11a or 802.11g cell, traffic is slower on the wireless side (54-Mb/s max half duplex) than on a typical Ethernet link (100 Mb/s or 1000-Mb/s full duplex). APs commonly have two radios, so approximately double the throughput read in this table to take into account wireless flows that are coming from both 802.11g and 802.11a and that are being transferred onto the Ethernet cable. The total is still far less than 100 Mb/s. QoS is, therefore, important for traffic going to the AP from the switch and being sent into the wireless space.

With the new generations of 802.11n APs offering wireless throughput of up to 300 Mb/s, 100 Mb/s might not be enough to relay wireless traffic to the switch without congestion. IEEE 802.11n AP Ethernet ports are 1000 Mb/s for that reason.

The links to the controller are usually 1000-Mb/s ports, and several ports are available on larger model controllers to aggregate ports. When a controller performs at full capacity, these links might be heavily loaded, and QoS might need to be applied.

www.CareerCert.info



Switch QoS Configuration This topic describes the QoS configuration on switches connecting to APs and controllers.


AP and Cisco WLC Switch Example QoS Configuration

interface GigabitEthernet1/0/1

switchport access vlan 100

switchport mode access

mls qos trust dscp

spanning-tree portfast

interface GigabitEthernet1/0/13

switchport trunk encapsulation dot1q

switchport trunk allowed vlan 11-13,60,61

switchport mode trunk

mls qos trust cos

AP Switch Configuration

The QoS configuration of the AP switch is relatively trivial because the switch must trust the differentiated services code point (DSCP) of the CAPWAP packets that the AP passes to it. There is no class of service (CoS) marking on the CAPWAP frames coming from the AP because it is not a trunk switch port. The following is an example of this configuration:

interface GigabitEthernet1/0/1 switchport access vlan 100 switchport mode access mls qos trust dscp spanning-tree portfast end

Cisco WLC Switch Configuration

The following example chooses to trust the CoS of settings of the Cisco Wireless LAN Controller (Cisco WLC), because trusting CoS on the Cisco WLC port allows a central location for the management of WLAN QoS, rather than having to manage the WLC configuration and an additional policy at the WLC switch connection. Other customers that want a more precise degree of control can implement QoS classification policies on the WLAN-client VLANs.

interface GigabitEthernet1/0/13 switchport trunk encapsulation dot1q switchport trunk allowed vlan 11-13,60,61 switchport mode trunk mls qos trust cos end

www.CareerCert.info




AP and Cisco WLC Switch QoS Considerations

AP traffic:� Trusting DSCP on the AP port is trusting the WLC QoS configuration.

� Additional QoS policies for queuing might need to be added onto the AP port.

Cisco WLC traffic:� Traffic leaving the WLC can be either upstream (to the WLC or network)

or downstream (the AP and WLAN client).

� CoS and DSCP values for downstream traffic are determined by theWLC QoS policy and can be trusted.

� CoS value for upstream traffic is determined by the WLC QoS policy and is trusted.

� DSCP value for upstream traffic is determined by the WLAN client DSCP request and is not trusted.

� Additional QoS policies for queuing might need to be added onto the Cisco WLC port.

The switch configuration for ports going to the APs trusts DSCP. In trusting the AP DSCP values, the access switch simply trusts the policy set for that AP port destination by the Cisco WLC. The maximum DSCP value assigned to client traffic is based on the QoS policy applied to the WLANs on that AP. Note that the example configuration shown previously addresses only the classification and that queuing commands can be added, depending on local QoS policy. Trusting the AP port DSCP also means that the port is safe, in the sense that you can expect nothing other than the AP to be connected to that port.

The QoS classification decision at the WLC-connected switch is a bit more complicated than at the AP-connected switch, because the choice can be to trust either the DSCP or the CoS of traffic coming from the Cisco WLC. In this decision, there are a number of points to consider:

� Traffic leaving the Cisco WLC can be either upstream (to the core network) or downstream (to the AP and WLAN client). The downstream traffic is CAPWAP-encapsulated, and the upstream traffic originally came from the WLAN clients via the AP. Therefore, packets from the WLC port are either CAPWAP-encapsulated and destined for the AP or decapsulated WLAN client traffic leaving the Cisco WLC port and destined for the final reciprocating host.

� The QoS policies on the Cisco WLC control the DSCP values of outer CAPWAP headers. DSCP values on the encapsulated WLAN client traffic are not altered and reflect those set by the WLAN client.

� When a controller receives CAPWAP-encapsulated traffic, it discards the outer header. The outer header QoS values are needed only between the AP and the controller. Only those DSCP values originally requested by the client remain. You can trust these values only if you trust all the clients. In most cases, the choice is not to trust them and, therefore, not to trust DSCP values in controller upstream traffic.

� CoS values of frames leaving the Cisco WLC are set by the Cisco WLC QoS policies, regardless of whether they are upstream, downstream, encapsulated, or decapsulated.

www.CareerCert.info



CAPWAP Traffic Classification This topic describes how CAPWAP traffic can be classified on LAN and WAN links.


CAPWAP Traffic

Wired CAPWAP control traffic: UDP 5246� Initialization traffic: image traffic marked best effort, ACKs marked CS6

� Background traffic: marked CS6

Wired CAPWAP data traffic: UDP 5247

802.11 management frames (CAPWAP encapsulated): marked CS6� Probes

� Association requests and responses

802.11 data frames (802.1X)� From clients: marked depending on WLAN policy

� From controller: marked CS4

You can separate CAPWAP AP packets generally into the following two types:

� CAPWAP control traffic: Identified by User Datagram Protocol (UDP) port 5246

� CAPWAP data traffic: Identified by UDP port 52471

CAPWAP control traffic generally into the following two additional types:

� Initialization traffic: Generated when a CAPWAP AP is booted and joins a CAPWAP system; for example, the traffic generated by controller discovery, AP configuration, and AP firmware updates.

Note CAPWAP image packets from the controller are marked best effort, but their

acknowledgement is marked CS6. Note that there is no windowing of the protocol, and each additional packet is sent only after an acknowledgement. This type of handshaking minimizes the impact of downloading files over a WAN.

� Background traffic: Generated by a CAPWAP AP when it is an operating member of a WLAN network using established communication with a WLC; for example, CAPWAP heartbeat, Radio Resource Management (RRM), and rogue AP measurements. Background CAPWAP control traffic is marked CS6.

1 CAPWAP control traffic uses UDP 5246 and CAPWAP data traffic uses UDP 5247.

www.CareerCert.info



You can divide CAPWAP data traffic generally into the following two additional types:

� 802.11 management frames: IEEE 802.11 management frames such as probe requests, association requests, and responses are classified automatically with a DSCP of CS6.

� 802.11 data frames: Client data and 802.1X data from the client is classified according to the WLAN QoS settings, but packets containing 802.1X frames from the WLC are marked CS4. IEEE 802.11 data traffic classification depends on the QoS policies applied in the WLAN configuration and is not automatic. The default classification for WLAN data traffic is best effort.

www.CareerCert.info



CAPWAP Traffic Volumes This topic describes network volume generated by CAPWAP traffic.


CAPWAP Traffic Volume

CAPWAP control:� Discovery: 97 bytes request, 106 bytes response

� Join: 3000 bytes typical exchange

� Config: initial (6000 bytes) and config changes (360 bytes)

� Initial RRM: initial (1400 bytes) and config changes (375 bytes)

� Heartbeats: 4 packets/30s, 97 bytes/106 bytes

� RRM exchanges: 396 bytes/60s, 2660 bytes/180s

CAPWAP data encapsulation:� 14 bytes added to the 802.11 header + 46 bytes of additional

Layer 4, Layer 3, and Layer 2 Ethernet headers (60 bytes total)

The traffic volume generated by CAPWAP control messages is as follows:

� CAPWAP discovery messages: Use the CAPWAP discovery requests sent by the AP to determine which WLCs are present in the network. The AP also uses these messages to update information about backup controllers. A discovery request packet is 97 bytes, which includes the 4-B frame check sequence (FCS). A discovery response packet is 106 bytes, which includes the 4-B FCS.

� CAPWAP join messages: The AP uses a CAPWAP join request packet to inform the WLC that it wants to service clients through the controller. The join request phase is also used to discover the maximum transmission unit (MTU) supported by the transport. If a join response is received for the initial request, the AP forwards frames without any fragmentation. The join response also initiates the heartbeat timer (a 30-second value), which, when it expires, deletes the WLC-AP session. The timer is refreshed upon the receipt of the echo request or acknowledgements. If the initial join request does not yield any response, the AP sends out another join request with the test element, which brings the total payload to 1500 bytes. If the second join request does not yield a response either, the AP continues to cycle between the large and small packets and eventually times out to start over from the discovery phase. The initial join request sent by the AP is always padded with a test element of 1596 bytes. If a join response is received for the initial request, the AP forwards frames without any fragmentation. If the initial join request does not yield any response, the AP sends out another join request with the test element, which brings the total payload to 1500 bytes. Packet sizes for the join request and response messages vary based on the description, but

www.CareerCert.info



the typical packet exchange between the AP and the WLC (ap-manager interface) is 3000 bytes.

� CAPWAP config messages: The CAPWAP config requests and responses are exchanged between the APs and the controllers to create, change (update), or delete the services offered by an AP.

In general, an AP sends a configure request message to send its current configuration to its WLC.

The configure request can be sent in two scenarios:

The first scenario comprises the initial phase when the AP joins a controller and needs to be provisioned with all 802.11 settings that are configured on the controller.

In the case of on-demand administrative changes, such as a change to a WLAN parameter the following occurs:

� The WLC sends the CAPWAP config response message type to the AP to acknowledge the receipt of the CAPWAP config request from the AP. This provides an opportunity for the WLC to override the AP's requested configuration. There are no special message elements contained by such a frame. The initial exchange between the AP and the WLC (ap-manager interface) is approximately 6000 bytes and a one-time configuration change averages 360 bytes and involves two packets each from the AP and the ap-manager interface of the WLC.

� Initial CAPWAP RRM messages: An RRM-related information exchange takes place after the AP is provisioned. In the event of an RRM-related configuration change, another packet exchange occurs between the AP and the ap-manager interface of the WLC. A typical exchange between the AP and the WLC (ap-manager interface) is approximately 1400 bytes. In the event of an RRM-related configuration change, there is a four-packet exchange between the AP and the ap-manager interface of the WLC. This exchange averages 375 bytes.

� Heartbeats: The CAPWAP architecture provides for a heartbeat timer that is accomplished by a series of echo requests and echo responses. An AP periodically sends echo requests to determine the state of the connection between the AP and the WLC. In response, the WLC sends the echo response to acknowledge the receipt of the echo request. The AP, then, resets the heartbeat timer to the EchoInterval. The CAPWAP protocol specification draft contains a detailed description of these timers. The system heartbeat, coupled with fallback mechanism, is four packets every 30 seconds.

� RRM Exchanges: There are two ongoing RRM exchanges. The first one is the load and signal measurement. The second sequence of packets is the noise measurement that includes a statistics information request and response sequence. The first one, at every 60-second interval, consists of four packets. This exchange always adds up to 396 bytes. The second is done every 180 seconds. This short exchange of packets averages approximately 2660 bytes and typically lasts 0.01 seconds.

The traffic generated by CAPWAP data encapsulation depends on the client traffic itself. The CAPWAP data frame header adds 14 bytes to the existing 802.11 packets. This header is added before the encapsulated 802.11 frame and includes this:

Light Weight Access Point Protocol [0-40] Flags: %00000000 [42-48] 00.. .... Version: 0 ..00 0... Radio ID: 0 .... .0.. C Bit - Data message [0-29]

www.CareerCert.info



.... ..0. F Bit - Fragmented packet [0-34] .... ...0 L Bit - Last fragment [0-30] Fragment ID: 0x00 [43-55] Length: 74 [44-52] Rec Sig Strngth Indic:183 dBm [46-77] Signal to Noise Ratio:25 dB [47-76]

Because CAPWAP frames can be fragmented, a Fragment ID field is included in the CAPWAP information section. The total packet size can be determined if you add the original frame and the IP fragment.

To this CAPWAP encapsulation, you must add the outer Layer 3 and Layer 2 headers. In the case of a typical voice packet using G.711, the voice payload is 160 bytes long, to which 40 bytes of Layer 3 and Layer 4 overhead are added, and 28 bytes of 802.11 Layer 2 overhead. On the wireless side, 228 bytes are used. When going through the AP, the packet is encapsulated into CAPWAP, adding 14 bytes, and new outer Layer 3, Layer 4, and Layer 2 headers are created, adding 60 bytes to the frame (18-B Ethernet header and FCS, 20-B IP header, 8-B UDP header, and 14-B CAPWAP segment). The frame is 288 bytes long when leaving the AP. Although the frame is 30 percent larger, it leaves a medium for which throughput is at best 30 Mb/s. Typically, throughput is 100 Mb/s. This overhead is usually negligible on a campus infrastructure network.

www.CareerCert.info




CAPWAP Traffic Volume (Cont.)

Example of a quiet network:

� CAPWAP control is the main traffic.

� 20 minutes exchange of initial traffic shows 0.001% network utilization.

� 20 minutes exchange of ongoing traffic shows 0.35 kb/s per AP onaverage.

Cisco testing has found that the average background traffic per AP is approximately 305 b/s. These tests were conducted on 802.11a/b/g access points.2

Calculating the average initial traffic per AP is more difficult, because the average time taken for an AP to go from rebooted to operational is a function of the WAN speed, as well as that of the WLC and AP. In reality, the difference is minimal. While on a lab test network, the best of initial traffic might average 2614 b/s over 18 seconds. With a WAN link with 100-ms RTT, the average is 2318 b/s over 20.3 seconds.

An hour-long sample presents the break-up of protocols shown here. This sample was taken on a very quiet network so that data traffic would not affect too severely the CAPWAP traffic. This break-up shows the relative percentage of CAPWAP traffic without active data traffic.

The operation of CAPWAP does not introduce heavy bandwidth requirements on the infrastructure, and in most typical deployments, there is no need to add extra capacity to the infrastructure to accommodate Cisco Unified Wireless Network architecture. Keep these facts about the operation of CAPWAP in mind:

� Although latency is an important consideration, the AP-to-WLC link must not exceed 100-ms round-trip latency.

� There are two separate ports for the operation of CAPWAP:

— CAPWAP data

— CAPWAP control traffic

� CAPWAP operation is broken down into two broad categories:

— One-time exchanges

— On-going exchanges

2 The addition of 802.11n capabilities might change this traffic measurement.

www.CareerCert.info



A 20-minute sample that includes initial exchanges results in an average utilization statistic of 0.001 percent.

A 20-minute sample of on-going exchanges results in a maximum utilization statistic of 0.35 kb/s.

www.CareerCert.info



CAPWAP Marking Manipulation This topic describes how CAPWAP marking can be changed.


CAPWAP Traffic Reclassification

� CS6 is suitable in most cases

� CS6 is also the default for routing protocols such as BGP, OSPF,EIGRP, and so on.

� Some networks might have a QoS policy contradicting with CAPWAP = CS6 on the WAN.

� CAPWAP traffic can be reclassified, if necessary, based on its ports 5246 and 5247.

The DSCP classification of CAPWAP control traffic is CS6, which is an IP routing class, and is intended for IP routing protocols such as Border Gateway Protocol (BGP), Open Shortest Path First (OSPF), Enhanced Interior Gateway Routing Protocol (EIGRP), and so on.

The current CAPWAP DSCP classification represents a classification that, although optimal for the WLAN system, may not align with the QoS policies and needs of each customer.

In particular, customers might want to minimize the amount of CS6-classified traffic generated by the WLAN network. They might want to stop CS6 traffic that is generated by client activity such as probe requests. The simplest mechanism to do this would be to reclassify the CAPWAP data CS6 traffic to a different DSCP. The fact that the CAPWAP UDP port used for control is different from that used by CAPWAP data and the default DSCP marking allow for remarking this traffic without resorting to deep packet inspection.

In addition, customers might want to ensure that CAPWAP initialization traffic does not affect routing traffic. The simplest mechanism for ensuring this is to mark CAPWAP control traffic that is in excess of the background rate with a lower priority.

www.CareerCert.info




CAPWAP Traffic Reclassification (Cont.)

class-map match-all CAPWAPDATACS6match access-group 110

match dscp cs6 !

policy-map CAPWAPDATACS6

class CAPWAPDATACS6set dscp cs3

!interface FastEthernet0

ip address 192.168.203.1 255.255.255.252service-policy input CAPWAPDATACS6

!access-list 110 permit udp 192.168.101.0 0.0.0.255 host 192.168.60.11 eq 5247access-list 110 permit udp 192.168.101.0 0.0.0.255 host 192.168.60.11 eq 5246

� Example: Re-marking client CAPWAP control and data traffic to CS3:

Re-Marking Client-Generated CS6 Packets

This illustration shows a sample configuration for remarking CAPWAP data packets marked as CS6 to a more appropriate value of CS3. This moves the traffic to a more suitable classification, at the level of call control, rather than at the level of network control. Traffic coming from APs in the 192.168.101.0/24 subnet and destined for the controller at 192.168.60.11 is filtered. CAPWAP control traffic, sent to port 5247, is re-marked from CS6 to CS3. Client traffic, sent to port 5246, is re-marked to CS3 only if its original tag was CS6, identifying management information.

www.CareerCert.info




CAPWAP Traffic Rate Limiting

class-map match-all CAPWAPCTRLCS6

match access-group 110match dscp cs6

!policy-map CAPWAPCTRLCS6

class CAPWAPCTRLCS6police 8000 conform-action transmit exceed-action set-dscp-transmit 26

!interface FastEthernet0

ip address 192.168.203.1 255.255.255.252service-policy output CAPWAPCTRLCS6

!

access-list 110 permit udp 192.168.101.0 0.0.0.255 host 192.168.60.11 eq 5246

� Example: Controlling CAPWAP control traffic data rate:

Changing the DSCP of CAPWAP Control Traffic Above a Predefined Rate

Here is an example of rate limiting the CAPWAP control traffic from the WAN site to minimize the effect of the CS6-marked control traffic on routing traffic. Note that the rate limit configuration does not drop nonconforming traffic but simply reclassifies that traffic.

Note Note that this is an example, and not a recommendation. Under normal circumstances, and following the design guidelines for deploying APs over a WAN connection, it is unlikely that CAPWAP control traffic would affect the WAN routing protocol connection.

www.CareerCert.info





Summary

� The size of the average 802.11 frame is directly related to the throughput of clients in the cell.

� Switch link to the controller is typically set to trunk mode, trusting CoS, while links to APs are typically set to access mode, trusting DSCP.

� There are several forms of CAPWAP traffic, some related to client encapsulation; some others to AP management and RF control.

� CAPWAP represents a very light overhead on network resources, usually very acceptable on campus types of networks.

� Although CAPWAP control traffic is tagged CS6 by default, it is possible to re-mark it in specific scenarios.

www.CareerCert.info



Module Summary This topic summarizes the key points that were discussed in this module.


Module Summary

� Several QoS mechanisms allow for optimizing of queuing and congestion management on the wired side, while on the wireless side, the 802.11e and its partial certification, WMM, improves prioritization in the cell.

� Because all wireless traffic is CAPWAP-encapsulated, QoS levels requested by clients are differentiated from and capped by QoS levels that are configured for the WLAN.

� On the controller, QoS parameters are configured globally, whilesome others affect only determined WLANs; most of them can be applied as templates from the Cisco WCS.

� The QoS policy defined on the wireless side is extended to the wired side by configuring queuing and CoS-to-ToS mapping.

� The LWAPP encapsulation optimizes traffic management; the CAPWAP overhead itself does not add heavily to the LAN global traffic.

Think of quality of service (QoS) as an end-to-end problem. When a wireless phone initiates a call, each packet sent and received must travel from one end to the other at as steady a pace as possible, within a delay below 150 ms if possible. To achieve this objective, you must manage wireless and wired resources to prioritize voice traffic in internetworking device buffers to ensure that it receives proper bandwidth guarantees all along the path.

This QoS configuration starts in the wireless space, where Wi-Fi Multimedia (WMM) allows wireless devices themselves to prioritize voice over the other types of traffic. It continues on the controller, where QoS configuration allows linking wireless QoS to wired QoS. The same QoS global strategy extends to the wired network, where you must configure routers and switches to apply the same logic and ensure that traffic coming from the controller is recognized, classified, and prioritized properly.

Good QoS is the second key element to proper voice over wireless LAN (VoWLAN) implementation, the first one being appropriate WLAN and access point (AP) design. When both QoS and WLANs are configured to accommodate VoWLAN, voice in wireless environments can fulfill the same expectations as wired VoIP deployments.

www.CareerCert.info


© 2009 Cisco Systems, Inc. Implementation of QOS for Wireless Applications 1-153

References For additional information, refer to these resources:

� Cisco Enterprise Mobility 4.1 Design Guide: http://www.cisco.com/en/US/docs/solutions/Enterprise/Mobility/emob41dg/emob41dg-wrapper.html

� IEEE 802.11e protocol: http://standards.ieee.org/getieee802/

� WMM: http://www.wi-fi.org/white_papers/whitepaper-090104-wmm

� Cisco WLC configuration guide: http://www.cisco.com/en/US/products/ps6366/products_installation_and_configuration_guides_list.html

� Cisco WCS configuration guide: http://www.cisco.com/en/US/products/ps6305/products_installation_and_configuration_guides_list.html

www.CareerCert.info



Module Self-Check Use the questions here to review what you learned in this module. The correct answers and solutions are found in the Module Self-Check Answer Key.

Q1) Which of the following is true? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) ToS and CoS both use a 1-B field. B) IP precedence is a form of CoS. C) DSCP is a form of IP precedence. D) DSCP is a form of ToS.

Q2) How long is the field in which CoS information is marked? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) 4 bits B) 1 byte C) 4 bytes D) 8 bytes

Q3) Which of the following is true? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) AF 32 offers a higher drop probability level than AF 31. B) EF offers a higher drop probability level than DSCP 56. C) The AF class offers a higher drop probability level than the ECN class. D) AF 31 offers a higher drop probability level than AF 32.

Q4) Where should trust boundary ideally be set? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) as close as possible to the source B) as close as possible to the destination C) all along the path D) at the edge of the LAN

Q5) What is the aCWmax with 802.11e? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) depends on the WMM/802.11e queues B) (aCWmin x2 )-1 C) 16 D) 1023

Q6) How many access categories (ACs) does 802.11e define? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) 1 B) 2 C) 4 D) 8

www.CareerCert.info



Q7) What is the length of an AIFS? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) the same as SIFS B) the same as DIFS C) at least one DIFS D) two SIFS

Q8) What is the purpose of TSpec? (Source: Identifying General Considerations for Wired and Wireless QoS)

A) for a client to inform the AP about the requested TC B) for WMM clients to use HCCA to reserve the medium before sending critical

traffic C) for the AP to specify which type of traffic is allowed in the cell D) for the wireless infrastructure to translate wireless QoS into wired QoS

Q9) How does an incoming frame marked with DSCP exit a controller? (Source: Describing Wireless QoS Deployment Schemes)

A) with only the DSCP marking B) with DSCP marking, and 802.1p if the controller is on a trunk port C) with only 802.1p marking D) with 802.1p marking and DSCP if the controller is on a trunk port

Q10) In the case of a WLAN limited to 802.1p value 3 and a wireless client requesting Platinum level, how does the AP tag the frame sent to the controller? (Source: Describing Wireless QoS Deployment Schemes)

A) the equivalent of 3 both in the inner and outer headers B) the equivalent of 3 in the inner header and of 5 in the outer header C) the equivalent of 5 in the inner header and of 3 in the outer header D) the equivalent of 5 both in the inner and the outer header

Q11) In the case of a client requesting DSCP 56 (Platinum) in a WLAN limited to 802.1p value 3, how is the frame tagged when leaving the controller toward the wired network? (Source: Describing Wireless QoS Deployment Schemes)

A) 802.1p 3 and DSCP 56 B) 802.1p 3 and DSCP CS3 C) 802.1p 5 and DSCP 56 D) untagged, because the controller cannot decide for the wired infrastructure

Q12) Which of the following is true? (Source: Describing Wireless QoS Deployment Schemes)

A) H-REAP cannot tag locally switched traffic at Layer 2 because H-REAPs are on access ports.

B) For locally switched traffic, H-REAPs tag frames with 802.1p, limited to the WLAN QoS level.

C) For locally switched traffic, H-REAP does not tag at Layer 2, but tags with the DSCP value.

D) For the H-REAP to tag locally switched traffic, the local switch must have a specific CoS to DSCP map.

www.CareerCert.info



Q13) Which of the following is true? (Source: Configuring the Controller and Cisco WCS for QoS)

A) In 802.11e, the equivalent to Platinum for 802.1p is 5, and Cisco recommends 6, following the IETF recommendations.

B) In 802.11e, the equivalent to Platinum for 802.1p is 6, and Cisco recommends 5, following the IETF recommendations.

C) In WMM, the equivalent to Platinum for 802.1p is 5, and Cisco recommends 6, following the IEEE 802.11e recommendations.

D) In WMM, the equivalent to Platinum for 802.1p is 6, and Cisco recommends 5, following the IEEE 802.11e recommendations.

Q14) What is the default 802.1p tagging for WLANs? (Source: Configuring the Controller and Cisco WCS for QoS)

A) no tagging B) 1 C) 3 D) 6

Q15) What is the purpose of the QoS Role? (Source: Configuring the Controller and Cisco WCS for QoS)

A) to allow specific WLANs to override the global controller WMM to 802.1p mapping

B) to allow specific bandwidth contracts for temporary workers C) to allow QoS profiles based on the authentication type D) to allow specific mapping between inner and outer DSCP values

Q16) What happens if both “WMM allowed” and “7920 AP CAC” are selected? (Source: Configuring the Controller and Cisco WCS for QoS)

A) Cisco Wireless IP 7921 phones can associate but Wireless IP 7920 phones cannot associate because of conflicting information from the WMM allowed parameter.

B) Wireless IP 7920 phones can associate but Wireless IP 7921 phones cannot associate because WMM is not set to mandatory.

C) Both Wireless IP 7921 and 7920 phones with recent firmware can associate. D) Wireless IP 7921 phones cannot associate because of incorrect WMM settings

and Wireless IP 7920 phones cannot associate because of incorrect WMM and Wireless IP 7920 CAC combinations.

Q17) What is the purpose of TSM templates on the Cisco WCS? (Source: Configuring the Controller and Cisco WCS for QoS)

A) to determine the size of the TSM logs both on the controller and the Cisco WCS

B) to determine which client traffic TSM should log C) to determine which radios TSM should be enabled for D) to determine which threshold should flag which color alerts on Cisco WCS for

TSM events

www.CareerCert.info



Q18) Which of the following is a correct class map statement? (Source: Configuring the Wired Infrastructure for QoS)

A) class-map test1 match policy map test2

B) class-map test1 set ip dscp 56

C) class-map test1 match protocol telnet

D) interface s0/0 class-map test1

Q19) Which of the following is a correct policy-map statement? (Source: Configuring the Wired Infrastructure for QoS)

A) policy-map test2 class test1

B) policy-map test2 match protocol telnet

C) interface s0/0 policy-map output test2

D) policy-map test2 match class-map test1

Q20) Which of the following is a correct low latency queuing example on a router? (Source: Configuring the Wired Infrastructure for QoS)

A) policy-map test2 priority 40

B) policy-map test2 bandwidth 40 percent

C) pervice-policy test2 priority 1 bandwidth 40

D) plass-map test2 priority queue 1 bandwidth 40

Q21) Which of the following is a correct Layer 2 switch QoS configuration? (Source: Configuring the Wired Infrastructure for QoS)

E) interface f0/1 mls qos trust cos

F) interface f0/1 service-policy trust dscp

G) interface f0/1 qos map 0 1 2 3 4 5 6 7 to 0 0 3 3 4 5 5 7

H) interface f0/1 switchport trust tos

Q22) If two clients are connected to the same access point at 54 Mb/s, which of the following is true? (Source: Understanding Current Best-Practice Guidelines)

A) Client sending smaller frames will get an overall higher throughput. B) Client sending larger frames will get an overall higher throughput. C) Because both clients connect at the same speed, their throughput will be the

same. D) Clients with smaller frames will get higher throughput only if the cell

experiences congestion.

www.CareerCert.info



Q23) Which of the following is a correct common configuration for a switch port connecting to a CAPWAP AP in local mode? (Source: Understanding Current Best-Practice Guidelines)

A) switchport mode trunk mls qos trust cos

B) switchport mode trunk mls qos trust dscp

C) switchport mode access mls qos trust cos

D) switchport mode access mls qos trust dscp

Q24) Which of the following is the port used by Layer 3 control CAPWAP traffic? (Source: Understanding Current Best-Practice Guidelines)

A) UDP 15666 B) TCP 15666 C) UDP 5247 D) TCP 5246

Q25) How is CAPWAP control traffic tagged by default? (Source: Understanding Current Best-Practice Guidelines)

A) CAPWAP control traffic is not tagged by default. B) CAPWAP control traffic tag depends on the WLAN QoS configuration. C) CAPWAP control traffic is tagged 802.1p 7. D) CAPWAP control traffic is tagged DSCP CS6.

Q26) How much overhead does the CAPWAP encapsulation add to the 802.11 frame? (Source: Understanding Current Best-Practice Guidelines)

A) 1 byte B) 4 bytes C) 6 bytes D) 14 bytes

www.CareerCert.info



Module Self-Check Answer Key Q1) D

Q2) C

Q3) A

Q4) A

Q5) A

Q6) C

Q7) C

Q8) A

Q9) B

Q10) C

Q11) A

Q12) B

Q13) B

Q14) A

Q15) B

Q16) C

Q17) D

Q18) C

Q19) A

Q20) A

Q21) A

Q22) B

Q23) D

Q24) C

Q25) D

Q26) D

www.CareerCert.info



www.CareerCert.info


Module 2

Voice over Wireless Architecture

Overview Voice-ready wireless is an end-to-end approach that addresses the convergence of VoIP and wireless networks and allows administrators to extend the mobility benefits of wireless networks to their voice communications. Critical to this solution are client devices with fully integrated advanced wireless LAN capabilities to ensure optimal performance. A voice-ready wireless network is founded on Cisco Unified Wireless Network.

Designed to complement a variety of voice clients, the Cisco Unified Wireless Network is optimized for voice. It builds upon a highly-scalable architecture with a low total cost of ownership, which is designed to support pervasive deployments that are typical of customers with mobile voice applications. In addition, innovative features of the solution, such as end-to-end quality of service and fast secure roaming, backed by a portfolio of access points with enhanced radios, make the Cisco Unified Wireless Network "voice-ready."

In this module, you will examine design considerations and deployment guidelines for an implementation of voice over wireless LAN (VoWLAN) technology on the Cisco Unified Wireless Network infrastructure. You will learn how a wireless network can be tailored for voice and why bad designs provide poor results. Last, but not least, this module will give you the tools to estimate if an already deployed network is “voice-ready.”

Module Objectives Upon completing this module, you will be able to describe the voice over wireless architecture. This ability includes being able to meet these objectives:

� Describe the evolution of voice architecture

� Describe VoWLAN call flow

� Design wireless for voice

� Verify voice readiness

www.CareerCert.info



www.CareerCert.info


Lesson 1

Describing the Evolution of Voice Architecture

Overview Voice over IP communications may seem very far from the traditional phone system approach. Nevertheless, many common principles apply. Both Voice over IP and traditional telephony aim at bringing the sound of a person’s voice to the ear of a distant hearer, and vice versa. If you understand traditional phone communication principles, you will be able to easily comprehend the evolution towards Voice over IP communication. This lesson describes the traditional voice system, and how it evolved to a more complex model, populated with complex devices such as PBXs or Central Switches.

Objectives Upon completing this lesson, you will be able to describe traditional voice architecture. This ability includes being able to meet these objectives:

� Describe traditional voice networks

� Describe how voice is integrated into the data flow with VoIP

www.CareerCert.info



Traditional Voice Network This topic describes how calls are conducted in a traditional voice network.


Components of a Traditional Phone Network

� Voice calls are two cable links

� Connections are source jack to destination jack, directly or viaa trunk to another operator

In the early years of telephony, the first exchange systems consisted of cable pairs running from end user telephones to local operator offices. Each operator sat in front of a vertical panel containing banks of jacks, each of which was the local termination of a telephone line. To initiate a call, the caller would lift the telephone handset and turn a wheel. Turning this wheel would generate a current that would trigger a light corresponding to the jack for the caller on the operator’s panel.

The operator would plug a headset to the jack and be in communication with the caller over this simple 2-cable circuit. The calling party would then explain which number (or which place) was to be reached, and the operator would unplug the headset, then plug a cord from the caller jack to the recipient jack. If the call was intended for a distant location, the cord would be plugged to a link (called a trunk) to another operator office, where a second operator, closer to the destination, would repeat the operation until caller and recipient line were finally connected.

At the end of World War I, in the USA, an intercity call typically took 15 minutes to fully connect, when all the trunks along the way were available.

www.CareerCert.info


© 2009 Cisco Systems, Inc. Voice over Wireless Architecture 2-5


Components of a Traditional Phone Network (Cont.)

� CO switches replace manual operators

� Protocols are developed for interswitch communication and negotiation

During the twentieth century, automation and increased line density greatly improved this system. The operator’s panel slowly became a manual switch, for which destinations were entered as codes instead of physical jacks, and then was slowly automated. Today, when you are about to dial a number, the action of lifting the handset or pressing the line button generates a signal to your local automatic central office (CO) switch. The CO returns a dial tone indicating that it is ready to receive digits. The number you dial is sent to the CO.

Originally, this number used to be sent in the form of pulses on a rotary dial phone. These digits must be produced at a specific rate and within a certain level of tolerance. Each pulse consists of a “break” and a “make,” which are achieved by opening and closing the local loop circuit. More recent systems use another coding technique, called dual tone multifrequency (DTMF), for which each number is represented by a combination of two sounds that are sent at the same time and created by touch buttons. This technique is more robust: in the pulse system, the loss of one pulse would mean the wrong number.

The CO receives the dialed number, then redirects the call to the recipient circuit or forwards the information over a trunk to another CO. The phone call process is very close to the initial system, in the sense that two end users would be connected over a dedicated 64-kb/s circuit. It is still far more efficient, as the connection between lines is automated and commonly takes a fraction of a second.

A hidden aspect of this process is the communication between switches. A system must be in place to replace the intelligence of human operators. COs need to exchange information, about the availability and load on trunk lines and about the ongoing call, ranging from “recipient number busy” to “one side ending the call.” A common protocol used for this purpose in traditional telephony is Signaling System 7, or SS7.

www.CareerCert.info



When mechanical switches became electronic switches, in 1960s and 1970s, the voice carried over the trunk lines started to be sent as binary digits. In other words, when you speak, the local switch captures the received sound. The sound is then sampled at a rate of 8000 samples per second. Each sample is 8 kb/s which yields a connection rate of 64 Kb/s. Each sample is sent on the link as a binary code. This sampling solution is known as the Nyquist theorem and is used in pulse code modulation (PCM).

These samples, sent as digital values, can be manipulated just like any binary value (compressed, grouped with others, simplified, and so on). To increase the capacity of the telephone network, it is necessary to be able to send multiple calls over the same circuit. A method known as time-division multiplexing (TDM) was developed to perform this task. Each timeslot is 64 kb/s. which is the same size as the output from PCM, and the trunk itself has a larger bandwidth to accommodate several concurrent calls. To further reduce the size of the conversations transmitted across the network, a series of coder-decoders (codecs) were developed to code and decode the original PCM signal. Some of the more popular codec types are G.711, G.729, and G.723. G.711 is used for PCM and is considered to have the highest voice quality. The other two codecs use far less bandwidth, at the expense of the voice quality.

www.CareerCert.info




Components of a Traditional Phone Network (Cont.)

� PBXs act as enhanced mini CO switches

When the CO connects a whole enterprise instead of an individual user, new requirements appear. An example would be a phone number dialed from the enterprise office. The destination point can be a phone on the other side of the world, or on the next desk. As the quantity of phones increases with CO connections, so does the price. This cost is even more compounded by employees calling other employees within the same building. A connection outside the CO is created, just to turn back around and connect to the destination inside the same building. It soon became obvious that large, medium, or even small companies might benefit from having their own simplified CO switches to save cost on internal calls. This simplified in-building CO switch is called a private branch exchange (PBX). When an internal user lifts the handset, the PBX sends the dial tone. Then when the user dials a number, the PBX examines the destination number. If it is an internal number, it connects the two endpoints directly without leaving the PBX. If it is an external number, it relays the call to the external CO switch. The PBX acts as a bridge between the external world and the internal phone network. As only some of the users will be calling an outside number at any one time, the connection from the PBX to the CO switch can have a capacity of only a few phone calls. The PBX can assign internal numbers to the phones inside a building or campus, so that internal calls are processed completely independently. It takes care of call establishment, voice sampling and encoding, call maintenance, and termination.

Over the years, new PBXs have offered more integrated services such as voice-mail storage space. When an internal number is busy, voice-mail storage is available to leave a message. PBXs can also report phone utilization to an internal billing system. Some other examples of “intelligent functions” are redirecting a call to another phone number if the destination phone is busy or if its user presses a “busy” button, linking more than two phones together to build conference call systems, automatically redialing a number that was busy at configurable intervals, or temporarily storing incoming calls while the recipient of the call dials another number (call parking, associated with a music on hold system). A PBX today is therefore a hybrid device, which acts as a bridge and a simplified CO switch, but also brings call management services to the internal voice network.

www.CareerCert.info



VoIP Network This topic describes how calls are conducted in a VoIP network.


VoIP Business Case

� Cost savings – long-distance toll bypass

– Flexibility

– Network convergence

� Advanced features and applications

– Advanced call routing

– Unified messaging

– Telephony application services

As voice packets over trunks were already digitalized, it soon made sense to integrate the voice stream into existing IP networks, merging voice and data into the same binary infrastructure and creating Voice over IP (VoIP).

Originally, VoIP return on investment (ROI) calculations centered on toll-bypass and converged network savings. Although these savings are still relevant today, advances in voice technologies allow organizations and service providers to differentiate their product offerings by providing:

� Cost savings: Traditional voice flow used in the public switched telephone network (PSTN) environment dedicates 64 kb/s of bandwidth per voice channel. This approach results in unused bandwidth when there is no voice traffic. VoIP shares bandwidth across multiple logical connections, which makes more efficient use of the bandwidth and thereby reduces bandwidth requirements. This consolidation results in substantial savings on capital equipment and operating costs.

� Flexibility: The sophisticated functionality of IP networks allows organizations to be flexible in the types of applications and services that they provide to their customers and users. Service providers can easily segment customers. This segmentation helps them to provide different applications, custom services, and rates depending on the traffic volume needs and other customer-specific factors.

www.CareerCert.info



� Advanced features: VoIP applications can provide many types of new services, which act as differentiators pushing toward the migration from PSTN to VoIP, such as:

— Advanced call routing: When multiple paths exist to connect a call to its destination, some of these paths may be preferred over others based on cost, distance, quality, partner handoffs, traffic load, or various other considerations. Least-cost routing and time-of-day routing are two examples of advanced call routing that can be implemented to determine the best possible route for each call.

— Unified messaging: Unified messaging improves communications and productivity. It provides a single user interface for messages that have been delivered over a variety of media. For example, users can read their email, listen to their voice mail, and view fax messages by accessing a single inbox.

— Voice security: Some mechanisms in the IP network allow the administrator to ensure that IP conversations are secure. Encryption of sensitive signaling header fields and message bodies protects the packets in case of unauthorized packet interception.

— Telephony application services: XML services on Cisco Unified IP phones give users another way to access more business applications. Some examples of XML-based services on IP phones are user stock quotes, inventory checks, direct-dial directories, announcements, and advertisements. Cisco Unified IP phones are equipped with pixel-based displays that can show full graphics instead of just text on the window. The pixel-based display capabilities allow you to use sophisticated graphical presentations for applications on Cisco IP phones and make them available at any desktop, counter, or other location.

www.CareerCert.info




VoIP Networks

In traditional telephony, voice streams are digital. Voice leaves the user endpoint as an analog wave and is encoded into a binary flow in the CO switch. Most of the intelligence is in the CO switch. All along the path, CO switches use complex protocols, such as SS7, to negotiate call parameters. The binary flow is converted back to an analog voice before reaching the destination point, which passively replays the received sound.

In a VoIP network, the sound is digitalized directly in the VoIP phone. The voice stream is digital all the way between endpoints. A VoIP phone connects to the network and benefits from the ability of lower layers to partly manage the medium. For example, a VoIP phone connected to an Ethernet segment can use the Ethernet Carrier Sense Multiple Access with Collision Detection (CSMA/CD) or Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) protocols to detect collisions or sense the carrier availability.

www.CareerCert.info




VoIP Networks (Cont.)

The phone itself has an IP address and integrates into the IP network. A voice packet leaving the phone is treated just like any other packet: it uses IP for addressing and is routed over the network. Major routing protocols like Open Shortest Path First (OSPF) or Border Gateway Protocol (BGP) can be used to route voice packets as they already route data packets. This allows greater flexibility in the network organization. The intelligence does not need to reside anymore in an interface between the endpoint and the rest of the world, like with CO switches and PBXs. The intelligence can be distributed throughout the network among different components.

Phones still need to have numbers and perform voice-specific functions, such as call handling and management. The call intelligence is moved to a “call manager.” In Cisco networks, this call manager is called Cisco Unified Communications Manager. Redundancy can be built just like for any other IP component: as call management is a feature, it can be installed on one or several dedicated platforms. Call management functions can also be implemented directly in routers. It is common to see scenarios in which a dedicated Cisco Unified Communications Manager system performs the call management and a router has partial voice ability. In case of network or Cisco Unified Communications Manager failure, the router is able to ensure partial call management. The voice network can survive the loss of connectivity to the Cisco Unified Communications Manager system.1 A VoIP network is more distributed and usually more robust than a traditional voice network.

1 This feature is called Survivable Remote Site Telephony (SRST).

www.CareerCert.info




Voice and Data Integration

One of the many advantages of VoIP is the integration of voice with data, which allows many possibilities that voice alone cannot achieve, such as:

� Video telephony: traditional voice circuits were built and sized to carry one conversation stream per circuit. In an IP network, if bandwidth permits, images can be sent along with the voice stream.

� Web conferencing: the integration of voice in web application creates new possibilities, with many features such as whiteboard or application sharing, and one-to-many or many-to-many conversations.

� Applications bridges: as voice traffic is digitalized, it can be managed like any other binary value and converted or translated from one application to another. For example, automated systems can retrieve the content of a voice-mail message and send it as an audio attachment to a preconfigured email address.

� Customer contact center types of features: when an organization is deployed worldwide and has the VoIP ability to bypass PSTN costs, greater flexibility becomes possible for customer contact centers. For example, a user calling from a given country can be redirected to an agent speaking the right language, even if the agent is located in another country. When a user calls a help center, an automated audio menu can be presented to that user and the user responses used to redirect the call to the right service, which might be located anywhere in the company IP network.

The vocabulary used in a VoIP network is different from the one used in traditional voice environments. You will use terms such as:

� IP phone: An IP endpoint for voice communication. The main differences between an IP phone and a traditional phone are that an IP phone has the ability to encode voice directly into a digital stream2 and also has IP capabilities. An IP phone can be a DHCP client and use protocols such as User Datagram Protocol (UDP) or TCP.

2 In a traditional voice system, encoding is done at the level of the PBX or the CO switch.

www.CareerCert.info



� Gatekeeper or call agent: provides Call Admission Control (CAC), bandwidth control, and management. It determines if the call can be placed, or “admitted” (if there is enough bandwidth to allow it), and converts the dialed number into a destination IP address. In a Cisco voice network, this role is handled by Cisco Unified Communications Manager.

� Gateway: provides translation between VoIP and non-VoIP networks such as the traditional telephone network (PSTN). Gateways also provide physical access for local analog and digital voice devices such as telephones, fax machines, key sets, and PBXs. In other words, a gateway is a VoIP-capable networking device, such as a router with VoIP functionalities, which has one or several cards on which a PSTN, a fax, or an analog phone can be connected. Most voice-enabled company edge routers act as gateways. They can route a call over the IP network if the WAN link is available, or redirect it to a PSTN backup network if the WAN is not available or congested.

� Multipoint control unit: provides real-time connectivity for participants in multiple locations to attend the same videoconference or meeting. It performs the same functions as the PBX conference call feature, usually with more services, such as whiteboard or application sharing.

� Application server: provides services such as voice mail or unified messaging, IP contact center, and Presence.

� Videoconference station: provides access for end-user participation in videoconferencing. The videoconference station usually contains a video capture device for video input and a microphone for audio input. The user can view video streams and hear the audio that originates from a remote user station.

www.CareerCert.info




Call Control

The Communications Manager handles phone numbers and helps them establish and manage calls.

Just like traditional phones, IP phones need specialized devices to help them control calls. Controlling a call means, at minimum:

� Determining if the call can be placed in good conditions based of information such as bandwidth availability, user rights, reachability of the dialed string, and so on.

� Routing the call, that is, helping the calling phone to reach the called destination.

� During the call, monitoring the quality of the call and taking proactive actions based on this quality, such as bandwidth reservation or voice stream redirection through another path if the original link becomes congested.

� When one ends terminates the call, releasing the line, so that another call may be placed.

Some organizational models are so centralized that the phone literally receives all its instructions from the Cisco Unified Communications Manager. In these models, the phone has so few functions that when a user presses a key, the phone has to inform the Communications Manager to know how to react. The Communications Manager would return instructions such as “play a beep”. Some organizational models are more distributed and put more intelligence on local devices, which can be smaller specialized units connected to the main Communications Manager system. A more intelligent phone supposes more complex local software. Each organizational model, more or less centralized, relies on a different protocol that manages the communication parameters between the network devices and defines the roles.

www.CareerCert.info




Communications Manager Locations

In a centralized model, all phones in all locations rely on a central communications manager.

In a distributed model, local sites

have local communications

managers.

In most cases, VoIP phones perform some call management functions, and communication managers perform some others. Most organizational VoIP models fall into one of two categories, associated with two different families of protocols, based on if call management functions are centralized or distributed:

� Centralized model: All of the features are grouped in a central location. The IP phones or the branch systems are just used to connect to the central system. This model is easy to configure, as everything is done in a single place. If the link to the central location fails, Survivable Remote Site Telephony allows the branch to continue to function with little impact on the end users.

� Distributed model: Each branch has its own communications manager unit. This model is more expensive, as call agents and applications need to be duplicated and consistently maintained at several locations. This model is typically deployed where there are high concentrations of IP phones (large branches or distributed headquarters).

Some of the VoIP protocols associated with each model are only signaling protocols, but most of them also define how voice sampling can occur. In other words, they define how voice is encoded into a digital symbol and decoded back to a sound (thus determining what is called a codec), what bandwidth must be available for the call to succeed and how to manage the call during the conversation.

www.CareerCert.info




VoIP Signaling Protocols

Protocol Description

H.323 ITU standard protocol for interactive conferencing; evolved from H.320 ISDN standard; flexible, complex

MGCP IETF standard for PSTN gateway control; thin device control

SIP IETF protocol for interactive and noninteractive conferencing; simpler, but less mature, than H.323

SCCP or “Skinny” Cisco proprietary protocol used between Cisco Unified Communications Manager and Cisco VoIP phones

The level of centralization desired dictates the protocols that phones will have to use to communicate with the communications manager. The VoIP protocols that you may be using in voice over wireless deployment include:

� H.323: H.323 is a standard that specifies the components, protocols, and procedures that provide multimedia communication services—real-time audio, video, and data communications—over packet networks, including IP networks. H.323 is part of a family of ITU-T recommendations called H.32x. This H32x family of protocol describes multimedia communication services over a variety of networks. It is actually an umbrella of standards that define all aspects of synchronized voice, video, and data transmission. It also defines end-to-end call signaling. It has long been the most important standard for VoIP networks.

� MGCP: Media Gateway Control Protocol (MGCP) is a method for PSTN or VoIP gateway control or thin device control. With MGCP, a gateway is the entry point to the voice network. It can even be a phone. Specified in RFC 2705, MGCP defines a protocol that controls VoIP gateways that are connected to external call-control devices, referred to as call agents. MGCP provides the signaling capability for thin edge devices that may not have a full voice-signaling protocol such as H.323 implemented. For example, any time an event occurs at the voice port of a gateway, such as a phone being taken off the hook, the voice port reports that event to the call agent. The call agent then signals for the device to provide a service, such as dial tone signaling.

� SIP: Session Initiation Protocol (SIP) is a detailed protocol that specifies the commands and responses to set up and tear down calls. SIP also details features such as security, proxy, and TCP or UDP services. SIP and its partner protocols, Session Announcement Protocol (SAP) and Session Description Protocol (SDP), provide announcements and information about multicast sessions to users on a network. SIP defines end-to-end call signaling between devices. SIP is a text-based protocol that borrows many elements of HTTP, using the same transaction request and response model and similar header and response codes. It also adopts a modified form of the URL addressing scheme that is used within email that is based on Simple Mail Transfer Protocol (SMTP). SIP is becoming

www.CareerCert.info



more and more common for many providers of VoIP services over the Internet. Since it is an open standard, many new features can be added to it. This makes it very flexible, but renders interoperability sometimes difficult.

� SCCP: Skinny Client Control Protocol (SCCP) is a Cisco proprietary protocol used between Cisco Unified Communications Manager and Cisco VoIP phones. The end stations (telephones) that use SCCP are called Skinny clients. The client communicates with the Cisco Unified Communications Manager system using connection-oriented (TCP/IP-based) communication to establish a call with another H.323-compliant end station.

www.CareerCert.info




H.323 Suite

� Approved in 1996 by the ITU-T

� Umbrella of protocols designed for audio-video communications

� Subdivided into subprotocols such as H.225 call signaling, H.225Registration, Admission and Status, H.245 control signaling

� Widely used with gateways, gatekeepers, or third-party H.323 clients, especially video terminals in Cisco Unified Communications

H.323 is probably the most widely used protocol for audio and video communications on packet networks. It was designed with a thorough examination of the requirements for multimedia communication over IP networks, including audio, video, and data conferencing. It defines an entire, unified system for performing these functions, and was recommended by the ITU in 1996. It is often described as an “umbrella of protocols” more than a protocol, because it is subdivided in several subprotocols, each performing a specific action:

� H.225 call signaling: this is used to establish a connection between two H.323 devices. This connection is achieved by exchanging H.225 protocol messages on a call-signaling channel.

� H.225 Registration, Admission, and Status: Registration, Admission, and Status (RAS) is the protocol between endpoints (terminals and gateways) and gatekeepers. The RAS is used to perform registration, admission control, bandwidth changes, and status and disengage procedures between endpoints and gatekeepers. A RAS channel is used to exchange RAS messages. This signaling channel is opened between an endpoint and a gatekeeper prior to the establishment of any other channels (such as the voice-bearing channels).

� H.245 control signaling: H.245 control signaling is used to exchange end-to-end control messages governing the operation of the H.323 endpoint. These control messages carry information such as capabilities exchange, opening and closing of logical channels used to carry media streams, flow-control messages, and other general commands and indications.

In IP communications environments, H.323 is widely used with gateways, gatekeepers, and third-party H.323 clients, especially video terminals. It is a peer-to-peer protocol, in the sense that H.323-enabled devices communicate are not in a client/server relationship. Each device is fully capable of negotiating all communication parameters. The consequence is that bringing all H.323 functionalities to an IP phone implies creating a rather complex endpoint. It is common to see H.323 at the gateway level, the IP equivalent to the PBX mentioned before, and a simpler protocol between the gateway and the end devices inside the network.

www.CareerCert.info




Media Gateway Control Protocol

� Defined in RFC 3435, which renders RFC 2705 obsolete

– Architecture and requirements defined in RFC 2805

– Centralized device control with simple endpoints for basic and enhanced telephony services

– Master-Slave protocol

� Media gateway is controlled by the Media Gateway Controller (MGC)

� MG contacts its MGC when initialized

� MGC tells MG

– Media streams to establish

– Tones to play and events to monitor

– Digit maps against which to map received digits

The philosophy of MGCP is often described as the exact opposite of the philosophy of H.323, although this comparison is not entirely technically correct.

MGCP as a centralized control architecture has the advantage of centralized gateway administration and provides for largely scalable IP telephony solutions. The system is built around call agents, or Cisco Media Gateway Controllers (Cisco MGCs), and media gateways.

A Cisco MGC controls a number of dumb terminals, the media gateways. The Cisco MGC receives signaling information (like dialed digits) from the media gateway and can instruct it to alert the called party, to send and receive voice data, and so on. A MGCP Cisco MGC works as a software switch for a VoIP network; it really does nothing more than simply direct the media gateways and signaling gateways that perform all the work. All the dial plan information resides on a separate call agent. The call agent, which controls the ports on the gateway, performs call control. The gateway does media translation between the PSTN and the VoIP networks for external calls.

www.CareerCert.info




Session Initiation Protocol

� Is compliant with IETF RFC 2543 (1999), RFC 3261 (2002), and RFC 3665 (2003)

� Is based on the logic of the World Wide Web

� Is widely used with gateways and proxy servers within service provider networks

� Is a peer-to-peer protocol for which end devices (user agents) initiate sessions

� Is based on ASCII text for easy implementation and debugging

� Causes interoperability difficulties between vendors

SIP is a protocol developed by the Internet Engineering Task Force (IETF) Multiparty Multimedia Session Control (MMUSIC) working group as an alternative to H.323. SIP features are compliant with IETF RFC 2543, published in March 1999; RFC 3261, published in June 2002; and RFC 3665, published in December 2003. Because it is a common standard based on the logic of the World Wide Web and very simple to implement, SIP is widely used with gateways and proxy servers within service provider networks for internal and end-customer signaling.

SIP is a peer-to-peer protocol for which user agents (UAs) initiate sessions, like H.323. But unlike H.323, SIP uses ASCII text-based messages to communicate. Therefore, it simplifies troubleshooting and incoming signaling traffic content analysis.

SIP was designed to set up a "session" between two points and to be a modular, flexible component of the Internet architecture. SIP has many possible features, and each vendor is free to integrate some or all of them. As a result, and as paradoxical as it seems, SIP is now a 12-year old protocol with a vast number of interoperability issues. While SIP has been successfully deployed in some environments, those are generally "closed" environments. Interoperability is often built on strictly limited common sets of features.

www.CareerCert.info




Protocols Comparison

� MGCP:

– comes from the Telco world

– Very centralized

– Ideal for large and scalable deployments with few central points of control

� H.323:

– More distributed than MGCP

– Supports voice, but also any multimedia communication

� SIP:

– Built to be very distributed

– Based on Internet protocols (DNS, HTTP, MIME, URLs)

– So open that interoperability is sometimes difficult

It is often difficult to understand the differences between these various protocols unless you are a voice professional. This difficulty mainly comes from the fact that these protocols were developed by different groups, at different times, for different purposes, but using some common components.

MGCP is the most centralized of these protocols. MGCP comes from the Telco engineering world. With MGCP, the communications manager knows and controls the state of each individual connection to the PSTN, legacy PBX, voice-mail systems, plain old telephone service (POTS) phones, and so on. This is implemented with the use of a series of plain-text commands sent over UDP port 2427 between the communications manager and the slave devices.

H 323 is more distributed. H.323 was developed in the Enterprise LAN community as a video-conferencing technique. H.323 is extended with non-standard features in such a way as to avoid conflicts between vendors. Where MGCP focuses on voice communications, the H 323 family of protocols covers all aspects of multimedia communications. Given its multi-subprotocols nature, H.323 can be implemented in various platforms. Endpoints may have some intelligence and participate (in a limited way) to call management.

SIP was developed by the IETF, reusing many familiar Internet elements: SMTP, HTTP, URLs, MIME, and DNS. SIP was designed to set up a "session" between two points and to be a modular, flexible component of the Internet architecture. It is very distributed. Each endpoint can manage and control its call parameters. It has a loose concept of a call (that being a "session" with media streams), has no support for multimedia conferencing, and largely leaves the integration of sometimes disparate standards up to each vendor.

MGCP is an ideal protocol to use for building large, scalable systems, with one or a few central points of control. With MGCP, adding endpoints and gateways is just a matter of linking them to the central point of control.

www.CareerCert.info



H 323 is more of a hybrid, and can be deployed in most environment sizes. Depending on the environment size and types of applications to support, several subprotocols can be implemented and points of control can be distributed or more concentrated.

SIP is built to be flexible, but its openness is precisely what makes interoperability difficult. SIP is ideal for integrating voice into an Internet architecture without adding heavy central points of control.

www.CareerCert.info




Skinny Client Control Protocol

� Cisco proprietary terminal control protocol

� Stimulus protocol: For every event, the end device sends a message to the Cisco Unified Communications Manager system

� Cisco Unified Communications Manager translates into larger protocols (H.323, MGCP, SIP, and so on)

� Can be used to control gateways

� Proprietary nature allows quick additions and changes

In a Cisco network, the communications manager function is allocated to specialized software: Cisco Unified Communication Manager. This program is flexible, able to operate a heavy control on thin voice devices, but also able to delegate part of the tasks to more intelligent phones. It can also communicate with simplified versions of the software, running on specific routers with Cisco IOS embedding voice functions. This voice feature is called Cisco Unified Communications Manager Express. It can. by itself, manage phones in a more or less centralized manner, or act as a relay for branch offices to a larger Cisco Unified Communication Manager system located at the corporate headquarters. In case of loss of connectivity to the Cisco Unified Communications Manager system, Cisco Unified Communications Manager Express can still offer basic connectivity and functions to the local phones.

Skinny Client Control Protocol (SCCP) is a Cisco proprietary protocol that is used for communications between Cisco Unified Communications Manager, the Cisco implementation of Voice Control Server, and terminal endpoints. SCCP is a stimulus protocol, meaning any event (such as the phone being put on the hook or taken off the hook or the pressing of a button) causes a message to be sent to Cisco Unified Communications Manager. Cisco Unified Communications Manager then sends specific instructions back to the device to tell it what to do about the event. Therefore, each time a user presses a phone button, data traffic is sent between Cisco Unified Communications Manager and the terminal endpoint. SCCP is widely used with Cisco IP phones. The major advantage of SCCP within Cisco Unified Communications Manager networks is its proprietary nature, which allows you to make quick changes to the protocol and add features and functionality.

SCCP is a simplified protocol used in VoIP networks. Cisco Unified Communications Manager acts as a gateway that handles the major protocols that can be used on the WAN, such as H.323, MGCP, or SIP, and translates into SCCP only the information and instructions that are relevant to the phone. The communication between the phone and the gateway is therefore light, even “skinny.” As enterprises usually have voice gateways between their internal and the external networks, the combination of Cisco Unified Communications Manager and SCCP allows them to simplify exchanges between the phones and the gateway.

www.CareerCert.info





Summary

� In traditional voice networks, calls are switched along dedicated circuits running from callers’ phones, through devices such as the local PBX and CO switches, until they reach their final endpoints.

� In a VoIP network, voice traffic joins the data traffic, and voice endpoints are both voice- and IP-enabled.

www.CareerCert.info


Lesson 2

Describing VoWLAN Call Flow

Overview With the introduction of wireless communication, wireless IP phones, such as the Cisco Unified Wireless IP Phone 7921G, can provide voice communication within the corporate wireless local area network. The wireless phones depend upon and interact with wireless access points and Cisco IP telephony components, including Cisco Unified Communications Manager, to provide wireless voice communications. To properly design such a network, you need to understand how a voice call is placed over a wireless network. This lesson will guide you through the hardware components and protocols involved in such a communication process.

Objectives Upon completing this lesson, you will be able to describe voice traffic in a wireless network environment. This ability includes being able to meet these objectives:

� Describe how a wireless VoIP call is set up and how data flows between endpoints

� Describe wireless VoIP protocols and standards

� Describe the main hardware and software components involved in wireless VoIP communication

� Describe some wireless VoIP phones

www.CareerCert.info



Call Setup and Data Flow This topic describes how a wireless phone associates with control devices and places a call.


Call Setup and Data Flow

WLAN Controller

A wireless IP phone is a combination of a wireless networking device and a VoIP phone. To be able to place and receive calls, the phone has to do the following:

1. Associate with the wireless infrastructure. Just like any other wireless device, the phone must go through the authentication and association phase. It then registers with the access point (AP) or the controller and behaves like a normal wireless client. Because voice traffic is time-sensitive, the wireless infrastructure might add extra control procedures to ensure that the network provides sufficient connection quality for voice communications.

2. Register with a call agent. Recall that the call agent in a Cisco voice infrastructure is called Cisco Unified Communications Manager. The simplified version of Cisco Unified Communications Manager is Cisco Unified Communications Manager Express. Both provide a phone with voice-related items such as these:

� Latest firmware: When new firmware is released, it can be placed on Cisco Unified Communications Manager or Cisco Unified Communications Manager Express and be pushed automatically to the phone.

� Configuration file: Cisco Unified Communications Manager and Cisco Unified Communications Manager Express store a configuration file associated with each phone. When the phone registers with Cisco Unified Communications Manager or Cisco Unified Communications Manager Express, it retrieves this configuration file, which contains information such as which codec to use, what the extension number of the phone is, what to display on its screen, which ring tone to use, and so on.

www.CareerCert.info




Call Setup and Data Flow (Cont.)

WLAN Controller

Once associated with the wireless infrastructure and registered with Cisco Unified Communications Manager, the phone is ready to communicate. When a user wants to place a call, the phone sends a first packet to the AP. If there is not enough bandwidth to allow the call, and if wireless Call Admission Control (CAC) is configured, the wireless infrastructure refuses this packet and the call cannot be initiated. If there is sufficient bandwidth, the wireless infrastructure accepts the packet, and may reserve bandwidth for this call.

1. The first packets are typically a call initialization request followed by the dialed number, which are sent to the Cisco Unified Communications Manager system with which the phone has registered. The server examines the request and decides to allow or deny the call, based on criteria such as bandwidth, security configuration, phone profile, and so on. For example, international calls can be refused when initiated from certain phones, while allowed when coming from other phones. This configuration is done in the Cisco Unified Communications Manager system and is based on corporate or network policies.

2. If the call is allowed, the Cisco Unified Communications Manager system tries to reach the destination endpoint. This can imply going over an IP network to another Cisco Unified Communications Manager system that will make the receiving phone ring. If the IP network is not available, the call request might flow through the public switched telephone network (PSTN).

3. As soon as the receiving phone answers the call, a quick negotiation is done about some call parameters such as the common voice encoding system (codec) to use. Then the voice traffic flows directly from one phone to the other without being relayed through Cisco Unified Communications Manager anymore, until one of the endpoint signals that the communication is terminated. Cisco Unified Communications Manager then releases the line and records the call duration.

During the conversation, any endpoint can request Cisco Unified Communications Manager assistance, to park or redirect a call for example.

www.CareerCert.info



VoIP over Wireless Protocols This topic describes wireless protocols and standards that wireless IP phones and infrastructure need to initiate and establish a VoIP over Wireless LAN (VoWLAN) call.


Wireless Network Access

Client first step is to associate to the wireless infrastructure. Like any other client, phone needs to be configured for:� Authentication:

– Open

– Pre-shared key

– 802.1X-based

� Encryption:

– No encryption

– WEP

– TKIP/MIC

– AES

WLAN Controller

Before placing a call, a wireless phone must connect to the wireless network. This process involves several protocols. A wireless IP phone is not different from a data wireless device, and goes through the authentication and association sequences. Two elements need to be determined:

� Authentication: Most enterprise types of phones, including Cisco Unified Wireless IP Phone 7921G, Nokia E-series phone, SpectraLink phone, or Vocera phone, support both pre-shared key and IEEE 802.1X types of authentication. The subtypes supported in the 802.1X family depend on the phone model. They are carried vie the Extensible Authentication Protocol (EAP), and can be, for example, EAP-FAST (Extensible Authentication Protocol-Flexible Authentication via Secure Tunneling), EAP-TLS (Extensible Authentication Protocol-Transport Layer Security), EAP-PEAP (Extensible Authentication Protocol-Protected Extensible Authentication Protocol), or LEAP (Lightweight Extensible Authentication Protocol).

� Encryption: Most phones support no encryption, WEP (Wired Equivalent Privacy), WPA (Wi-Fi Protected Access), and WPA2 (Wi-Fi Protected Access 2). However, WPA2 is more CPU-intensive than WEP or WPA, and most phone CPUs are less powerful than laptops CPUs. The WPA2 encryption and decryption process is normally run at the wireless card level, and should not impact the voice communication performances. When WPA2 was released, some vendors decided to transfer the encryption and decryption tasks to the central CPU, to ensure compatibility. This might impact phone performance. Always verify that your phone is fully WPA2-compatible if you want to use this encryption. Encryption still adds overhead to the frame, but wireless infrastructures designed for voice communication ensure that bandwidth is not a limiting factor for call establishment.

www.CareerCert.info



On a Wireless IP Phone 7921G, the following wireless security features are supported:

� Authentication:

— Cisco Wireless Security Suite IEEE 802.1X Cisco LEAP authentication: Optional password prompt at power up.

— EAP-FAST

— EAP-TLS

— EAP-PEAP

— WPA versions 1 and 2

— WPA pre-shared key (WPA-PSK) versions 1 and 2

� Encryption:

— 40- and 128-bit static WEP

— Temporal Key Integrity Protocol (TKIP) and message integrity check (MIC)

— Advanced Encryption Standard (AES)

www.CareerCert.info




Wireless VoIP Phone Requirements Overview

� Cell design avoids configurations in which frames are delayed: smaller cells with higher speeds, call admission control

� Design ensures that clients associate to the best cells and phones get prioritized

WLAN Controller

Wireless IP phones have specific requirements regarding medium access timing. If the normal Carrier Sense Multiple Access with Collision Avoidance (CSMA/CA) process occurs in the cell, the voice packets may be queued behind large packets sent by data devices, and the conversation quality degraded. To properly design a wireless network to allow voice, network designers must make sure that:

� Cells have the right sizes and specifications: cell design avoids, as much as possible, configurations in which frames are delayed. Cells are smaller, to avoid the possibility of large packets getting sent from distant, low-speed outer perimeter (1 Mb/s for 802.11b and 6 Mb/s for 802.11g and 802.11a). Reducing transmission time for all packets is critical if large packets precede smaller VoIP packets that are waiting for RF time. Cells are also smaller to reduce the risk of issues such as multipath.

� Wireless phones associate to the best cells: some APs can communicate their load values to wireless phones. They include in their beacons an information element called the Quality of Service Basic Service Set (QBSS) load element, which informs the clients about their load levels. When a client is in range of two APs, it can choose to associate not only to the AP offering the best signal, but also to the one with the lowest load level.

� The wireless phone gets prioritized over other devices: if data and voice traffic are sent from the same cell, voice traffic must be prioritized. Wireless phones and APs are usually compatible with Wi-Fi Multimedia (WMM) to ensure traffic prioritization. When a wireless phone needs to associate to the wireless network, it sends information about the fact that it will be sending voice traffic: it specifies the type of traffic it wishes to send. This process is called Traffic Specification (TSpec). Based on this information, the AP allows the phone to associate or forbids it, with wireless CAC, then prioritizes the voice flow with WMM.1

1 The QBSS load element and TSpec are both part of the 802.11e protocol and WMM certification. Phones using TSpec usually do not need to use QBSS at the same time.

www.CareerCert.info




Cisco Unified Communications ManagerRegistration

� Wireless registration: Phone needs to obtain an IP address (static or DHCP)

� Voice registration: Phone registers with Cisco Unified Communications Manager or Cisco Unified Communications Manager Express to:

– Get the latest firmware

– Get the configuration file

– Signal itself to the voice infrastructure

� Configuration file: also describes how to encode voice streams

Once the wireless phone is associated to the wireless infrastructure, the second step is for the phone to use the wireless infrastructure to place and receive calls. To achieve this, the phone must make the voice network aware that it is now connected and available. It also has to obtain an extension number associated with configuration parameters. Unless the phone connects to a voice ISP over the Internet, it usually retrieves this information from a local server or router, running one of the control protocols mentioned in the lesson “Describing the Evolution of Voice Architecture.” In a Cisco voice infrastructure, this voice control service is ensured by Cisco Unified Communications Manager or Cisco Unified Communications Manager Express software. Theses platforms can support several protocols, such as Skinny Client Control Protocol (SCCP), H323, Media Gateway Control Protocol (MGCP), or Session Initiation Protocol (SIP).

The phone registers with Cisco Unified Communications Manager and receives its configuration. The phone is then ready to send and receive voice traffic. Whenever the wireless IP phone needs to place a call, Cisco Unified Communications Manager interacts with it just like with any other wired IP phone. Even if the wireless infrastructure allows the call, Cisco Unified Communications Manager may forbid it for any of various reasons, such as if the bandwidth on the WAN is insufficient. CAC for voice occurs after CAC for wireless has allowed the IP phone access to the wireless infrastructure.

www.CareerCert.info




Codecs

� G.711 packets are eight times the size of G.729 packets, but it does not consume eight times the bandwidth.

� G.711 has a higher MOS.

G.711 G.729

Sample 20 ms 20 ms

Packet size (bytes) 160 20

Wireless overhead (bytes)

28 28

Layer 3 / 4 overhead 40 40

Packets per second 50 50

Total bandwidth consumption (kbps) 91.2 (per stream) 35.2 (per stream)

MOS 4.4 3.7

A VoIP phone digitalizes the voice stream into a binary code. This process is called encoding. At the other end, the binary code is decoded and retransformed into an audio voice stream. One or several possible coder-decoder (codec) methods used on the stream are contained in the configuration file received by the phone from Cisco Unified Communications Manager.

The encoding operation consists of taking samples of the sound received by the microphone and converting each of them to a binary sequence. In a normal situation, 8000 samples are taken per second, then gathered into groups of samples representing 20 ms of voice traffic each.2

Each 20 ms is converted into a longer or shorter sequence of bits. The longer the sequence, the more information it contains and the closer the reconstruction at the other end will be to the original sound. On the other hand, a longer sequence consumes more bandwidth. The choice of a codec is a compromise between bandwidth consumption and acceptable quality. The resulting voice stream quality can be evaluated using several methods. A common method is called mean opinion score (MOS) testing. With MOS testing, a pre-defined set of sentences is spoken using various codecs and transmission conditions, and a pool of testers grade each sample on a scale from 0 to 5, where 5 would be hypothetical perfect similitude to live conditions. Each codec can then be compared to the other when transmitted in similar conditions. In voice over wireless, two main codecs are used:

G.711: this codec, also called Pulse Code Modulation (PCM), is the reference codec for the whole voice over IP industry. It is said to be the best, because its sampling system offers voice restitution very close to the original sound. It has the highest MOS of all voice codecs, up to 4.4. Each 20-ms sample is 160 B long. As each sample is 20 ms, using G.711 implies that 50 packets must be sent each second for the voice stream to be seamless.3 For the wireless

2 By default, Cisco devices send sound extracts of 20 ms, regardless of the codec used. Each sample contains 10 ms of effective voice sound. Each packet of 20 ms of audio contains two 10-ms samples. 3 1 second is 1000 ms; 1000 / 20 = 50 packets of 20 ms each per second.

www.CareerCert.info



infrastructure, this means that a wireless phone must be able to access the media 50 times per second.

As each packet is 160 B, 1280 b, 50 packets per second represents 64 kb/s of bandwidth consumption for the payload. This is why G.711 is said to use 64kb/s. Each packet has Layer 3 and Layer 4 overhead of about 40 B, as well as wireless overhead from the header and frame check sequence (FCS). Without encryption and with a default 802.11 header, the header and FCS use 28 B (224 b). A G.711 call, therefore, sends 50 packets of 1824 b/s, which uses 91.2 kb/s of bandwidth in the cell.

G.729: this codec is often used in VoIP networks, because it uses a lot less bandwidth than G.711 without much loss to its MOS. Each 20 ms sample is 20 B long (160 B), which represents 8 kb/s for the voice flow itself. After adding the Layer 4, Layer 3, and 802.11 header, as well as the FCS overhead, bandwidth consumption reaches 35.2 kb/s, a lot less than G.711. The MOS for G.729 is 3.7, which is, of course, less than the MOS for G.71, but still acceptable in most situations. This is why G.729 is often considered a good alternative to G.711 when bandwidth is an issue.

Nevertheless, the default of G.711 is often recommended in LAN and WLAN deployments. Using another codec can save some bandwidth, but some issues may occur when a compressed codec is converted to another compressed codec. The translation results in some losses, and the sound quality degrades dramatically.

Translation losses can occur with wireless phones in a corporate network. Suppose that your wireless phone uses a compressed codec such as G.729. Calling a desk phone from a wireless phone gives very good result; calling a Global System for Mobile Communications (GSM) phone from a desk phone gives a good result; calling the same GSM phone from the wireless phone gives a poor sound. The translation from G.729 to the GSM codec degrades voice quality, because the coding techniques are different.

www.CareerCert.info




“Call Legs” or Voice Streams� To give a full-duplex impression, 50 voice packets must be sent and 50 voice

packets must be received every second.

� Voice streams are sometimes named “call legs,” even though the term has a different meaning in voice networks.

WLAN Controller

WLAN Controller

In most voice over wireless deployments, 91.2 kb/s per call can easily be afforded in a well-designed cell, so G.711 is usually the default and preferred codec. However, the calculation on the previous page does not take into account what is sometimes called “wireless call leg,” or voice stream. In a voice conversation, both users have the impression of a full-duplex situation: each user can speak and hear the other one at the same time. This seems easy on a full-duplex Ethernet LAN segment, but wireless networks are CSMA/CA, which is a half-duplex technology. This means that when a voice conversation is measured on a wireless segment, both streams, up and down, have to be taken into consideration to evaluate the bandwidth requirement.

These streams are sometimes named “call legs4,” and it is said that a voice device needs two legs (two streams). This is true for each phone on each side, making a total of four voice streams for a wireless-phone-to-wireless-phone conversation. Both phones are not necessarily wireless, and are not necessarily in the same cell.

This implies that each phone actually needs 91.2 kb/s per voice stream, which makes a little less than 183 kb/s needed per active phone call. When adding the management frames (acknowledgements and probes) and the control frame overhead, a G.711 phone needs around 200 kb/s. Here again, in a 54 Mb/s cell, 200 kb/s can easily be afforded. Even a 1 Mb/s cell could allow several theoretical phone conversations. A G.729 phone needs about 80 kb/s total per active phone call.

4 In voice networks, the term “call leg” does not refer to exactly the same concept. A call leg is a logical connection between a voice enabled router and either a telephony endpoint over a bearer channel or another endpoint using a session protocol. This is why using the term “voice stream” is preferable and causes less confusion when speaking with voice professionals, even if several vendors in the wireless industry use the term “call leg.”

www.CareerCert.info




RTP vs. RTCP

� Each call is partly voice flow (RTP) and partly information about the voice flow (RTCP)

� RTP flow is urgent, RTCP is less critical

Voice traffic is often described as a real-time, UDP-based, steady stream of information. Voice traffic is actually composed of two streams, built on two different protocols:

� The voice flow itself, made up of the data packets carrying the digitalized voice samples: this traffic relies on UDP, mainly because many features of TCP are not relevant for voice traffic. For example, what would be the use of resending a lost packet? If a sample is missing, it will not be possible to replay it later. Still, UDP itself does not offer enough functionality to ensure smooth voice flow. Another protocol is associated to UDP for voice traffic: Real-Time Transport Protocol (RTP). RTP adds a header, positioned just after the UDP section in the frame, where extra information is added, including:

— payload type: This indicates the kind of content that is being sent, the codec.

— Sequence number: Each packet has a sequence number. When a packet is lost, the receiving phone can detect the loss and take action. Some of them play “silence,” some replay the previous packet, and some play the average value between the previous and the next packet. As samples are only 20-ms long, it is almost impossible, when playing a normal voice conversation, to detect when a single packet is missing. Notice that the sequence number is only used for loss detection, not to resend a lost packet.

— Timestamp: This shows when the sample was taken. It is used to replay the packets at the right speed and in the right order.

— Delivery monitoring: RTP can be associated with another protocol to monitor the quality of the call.

� The control traffic, which carries information about the quality of the call: this traffic relies on TCP or UDP, depending on the implementation, and is sent in parallel with that traffic. The main protocol used for this purpose is Real-Time Transport Control Protocol (RTCP). RTCP gathers statistics on the media connection and information such as the number of bytes sent, packets sent, and lost packets, as well as the jitter and round trip delay. An application may use this information to increase the quality of service, perhaps

www.CareerCert.info



by limiting flow or using a different codec. For example, when delay increases, the Cisco Unified Wireless IP Phone 7921G can dynamically increase the size of its buffer.

RTCP information is not sent for each voice packet, and it is considered acceptable if some RTCP packets are lost or delayed. If too many voice packets are lost, the conversation is affected. If RTCP packets are lost, the receiving end simply asks the emitter to resend. Despite its name, RTCP relies on UDP in most Cisco implementations. Some other vendors use TCP, but RTCP is independent of the Layer 4 underlying protocol used.

www.CareerCert.info




Delay

� Depending on cell conditions, the sending of a voice packet can be delayed.

� End-to-end delay should be less than 150 ms.

� If delay is beyond 30 ms within the cell, the packet is usually dropped.

In an 802.11 cell using CSMA/CA, before a wireless client sends a frame, it calculates a backoff timer value. This timer indicates the random number of slots (units of times) the wireless client will wait before sending the frame. At each step of the countdown, the client listens to the cell. If another device starts sending at that time, the first client adds the frame duration value in the 802.11 header for the other device to its backoff timer and begins counting down again from the new value. If many devices are in the cell, this process can add considerably to the total countdown time. When the backoff timer finally reaches zero, and if no other device is sending, the client sends its frame.

As radio networks are half-duplex, the client does not know while sending if a collision has occurred or if the destination point is receiving the frame. For that reason, the destination point needs to send an acknowledgement to the sender. If this acknowledgement is not received, the client assumes that the frame was lost, picks another random number, and begins counting down to send the frame a second time.

These two parameters, possible delay due to the variable countdown time and possible delay due to resends, mean that a voice packet can reach the AP less than a millisecond after entering the phone wireless buffer, while the packet that follows can take several tens of milliseconds.

The ITU defines acceptable delay for voice packet transmission in a voice network, end-to-end, as 150 ms or less. This is true whether both ends are in the same building or on different sides of the globe.5 For this reason, and also because voice traffic is real-time, it is necessary to reduce, as much as possible, the delay taken by a wireless frame sent by a wireless client to reach the AP. Delay can be reduced by applying wireless quality of service (QoS) to the Service Set Identifier (SSID), but also by designing cells so that the number of concurrent users is limited and controlled, and so that each user achieves the minimum defined access values (RSSI, max SNR, minimum throughput, and so on), to ensure distributed timely access to the wireless medium when needed.

5 Most wireless phones will drop a frame if it hasn’t reached the access point 30 milliseconds after having entered the network buffer.

www.CareerCert.info



Another issue is related to the cell itself. Because the end-to-end maximum delay is 150 ms, the time spent in the cell must be far less. For that reason, most wireless IP phones consider that if a frame takes more than 30 ms to reach the AP, it has few chances of reaching the destination within an overall 150 ms. In many phones, if a frame has not reached the AP 30 ms after having entered the phone buffer, the frame is simply discarded and the next one sent instead. It is better to maintain the flow than to keep sending the same frame forever. This creates another issue: When too many collisions occur and packets are re-sent, VoWLAN quality degrades due to variations in delay and to the dropped frames.

www.CareerCert.info




Jitter

Jitter is defined as variation in the delay of received packets. At the sending side, packets are queued in the buffer in a continuous stream, with the packets spaced evenly apart. Due to medium congestion and resends, this steady stream can become lumpy and the delay between each packet can vary instead of remaining constant. Sending buffers partly compensate for this issue: a phone does not have to send one packet exactly every 20 ms. A packet can be slightly delayed and sent just next to another one.6 When a voice-enabled device receives an RTP audio stream for VoIP, it must compensate for the jitter that is encountered. The mechanism that handles this function is the playout delay buffer. The playout delay buffer must buffer these packets and then play them out in a steady stream. The playout delay buffer is also sometimes referred to as the dejitter buffer. The Cisco Unified Wireless IP Phone 7921G has a dynamic jitter buffer, which can increase its size if the RTCP information shows that the jitter and delay are worsening.

If the jitter is so large that it causes packets to be received out of the range of the buffer, the out-of-range packets are discarded and dropouts are heard in the audio. It is important to understand that losses can occur due to lost packets, that is packets that do not reach the destination endpoint, or due to packets that reach the destination endpoint, but too late to be played while respecting the voice flow. These late packets are the most common type of lost packets. The receiving end prefers to play what should be heard at that time rather than stop the voice flow.

6 At 54 Mb/s, sending a voice packet takes about 0.034 ms. Several packets can be sent within a few milliseconds to empty the sending buffer.

www.CareerCert.info



For losses as small as one packet, the voice reconstructing algorithm interpolates what it thinks the audio should be and no problem is audible. Typically, the expected value is the intermediate between the previous and the next value. Some systems replay the previous value. When jitter exceeds the system’s ability to make up for the missing packets, audio problems are noticeable.

Audio problems are first noticeable due to lower MOS, in the sense that the voice heard seems different than the original speaker’s voice. As the voice quality decreases, its MOS decreases. It is often qualified as “metallic.” If losses increase, reconstruction becomes impossible and the voice becomes choppy, with clearly audible holes in the stream flow. Degradation of the MOS is higher. A VoIP network tries to limit its maximum loss rate due to lost or late packets, in order to maintain a certain MOS level.

www.CareerCert.info




VoWLAN Call Quality

� MOS is related to codec and packet loss rate

� Jitter and collisions increase the packet loss rate

� For G.711, the target is 1% packet loss or less at the –67dBm boundary to keep the MOS at an acceptable levels

The number of calls on a Wi-Fi channel is limited by a number of factors. First, the media used by the AP and VoWLAN clients is in the RF spectrum, which cannot be shielded from electromagnetic interference like shielded twisted-pair Category 5 cable. The closest Wi-Fi comes to segmentation is channel separation. This open shared media of 802.11 creates the possibility for high packet loss. Most of this packet loss is addressed through retransmission of 802.11 frames, which in turn causes jitter.

In 802.11a as well as 802.11g, the highest coverage range is achieved by the lowest data rate, which is 6 Mb/s. The lowest packet error rate is also at 6 Mb/s, for the same given power level. This seems to indicate that the lower speed should always offer the best MOS, and that wireless voice devices should use the lowest speed for best efficiency.

This common opinion does not take into account the parameters examined before, which are that the lower the speed, the longer it takes to send a single packet. As a consequence, fewer devices can send in a given timeframe. To increase the number of users in the cell, you want each of them to send as fast as possible. Also, lower speeds mean longer distances, and therefore higher risk of interferences and multipath.

Each cell has an “optimal size” that is a compromise between range and speed. An acceptable coverage area for voice is an area that maintains a MOS at an acceptable level. Looking at the MOS for G.711 shows that the packet error rate should be 5 percent or less for voice communication to be acceptable. The MOS scores are ranked as follows:

� 4.4: Top G.711 MOS score

� 4.3–4.0: “Very satisfied” to “satisfied”

� 4.0–3.6: “Some users satisfied”

www.CareerCert.info



A packet error rate of 5 percent reduces the MOS to a level of “some users satisfied” quality of speech. The coverage area edge for a phone is where the coverage area drops the MOS to the “very satisfied” category. This coverage area edge is referred to as a cell edge in this course. A cell edge with a 1-percent packet error rate is needed for voice, because of the likelihood of multiple phone clients, multiple data clients, co-channel interference, and other unaccounted-for interference. This means that the edge of the cell is set at 1 percent loss rate, but the client can go beyond this edge and still be associated to the AP. As the MOS degrades beyond this point, a design will make this point as a roaming point towards another AP.

www.CareerCert.info



Infrastructure Hardware and Software Components

This topic describes the hardware and software components commonly used in the Cisco Unified Wireless Network solution for voice over wireless networks.


Cisco Unified Wireless Network Components for Voice Over Wireless

Network Services

Network Management

Network Unification

Access Points

Client Devices

In a traditional LAN, phones and computers use cables to transmit messages and data packets over a wire conductor. Wireless LANs use radio waves to carry the messages and data packets. WLANs require AP devices that receive and transmit radio signals. Theoretically any AP can support a wireless IP phone by offering support for voice on a WLAN and prioritizing the voice over any other wireless traffic.

The AP can be autonomous, providing direct communication between the wireless space and the wired network, or be controller-based, using the Lightweight Access Point Protocol (LWAPP) and Control and Provisioning of Wireless Access Points (CAPWAP) protocols to communicate with a controller. In the first case, the AP is directly configured for wireless parameters such as SSID, wireless CAC, supported speeds, and so on. In the second case, the configuration is done on a controller and downloaded to the AP.

In a controller-based solution, the AP is a relay between wireless IP phones and the controller. As a user move from one location to another within the corporate WLAN environment, the wireless device roams out of range of one AP and into the range of another. The autonomous AP or controller uses the wired network to transmit credentials to the next autonomous AP or controller, in order for roaming to occur without disconnection. Autonomous APs allow only Layer 2 roaming, whereas controllers allow Layer 3 roaming. Each AP and each controller has a hard-wired connection to an Ethernet switch that is configured on the LAN. The switch provides access to gateways and the Cisco Unified Communications Manager server. At the management layer, devices such as Cisco Wireless Control System (Cisco WCS) provide single interfaces to control and monitor the wireless network.

www.CareerCert.info




Cisco Unified Wireless LAN Access Points

Cisco Unified Wireless LAN APs can be:� Autonomous or lightweight

(CAPWAP or LWAPP) based

� Single- or dual-band:

– 802.11b/g

– 802.11a/b/g

– 802.11a/b/g/n

Within the Cisco Unified wireless architecture, there are two categories of APs: standalone and CAPWAP- or LWAPP-based. Cisco standalone APs consist of the original Cisco Aironet product line. Most models are available in or are capable of being field-upgraded to a lightweight (CAPWAP or LWAPP) mode of operation. This feature permits an enterprise to standardize on a common AP platform that can be deployed in mixed wireless topologies.

Some standalone or lightweight APs are as follows:

� Cisco Aironet 1130 AG Series Access Point: a dual-band (a/b/g) AP with integrated antennas. It is designed to be ceiling-mounted and makes use of an integrated dual-band antenna. The Aironet 1130 AG Series Access Point is available in a lightweight (CAPWAP) version for implementation in centralized deployments based on wireless LAN controllers (WLCs). The standalone version can be upgraded for lightweight operation. The part number for the CAPWAP AP is AIR-LAP1131AG-x-K9, where x is the regional code.

� Cisco Aironet 1240 AG Series Access Point: a dual-band 802.11 a/b/g AP designed for deployments in challenging RF environments such as retail and warehousing. The Aironet 1240 AG Access Point possesses external connections for antennas in both bands. It is the most feature-rich AP in the standalone category and is also available in a lightweight (CAPWAP) version. For greatest flexibility, the standalone version can be upgraded later to a lightweight mode of operation. Other notable features include pre-installed certificates for CAPWAP operation mode and the ability to support Hybrid Remote-Edge Access Point (H-REAP). The part number for the CAPWAP AP is AIR-LAP1242AG-x-K9, where x is the regional code.

� Cisco Aironet 1300 Series Access Point: a single band 802.11b and g AP/bridge designed for outdoor deployments. It comes with an integrated antenna or can be ordered with RP-TNC connectors to support external antenna applications. The CAPWAP AP part number is AIR-LAP1310G-x-K9 where x is the regional code.

www.CareerCert.info



� Cisco Aironet 1252 Access Point: a new, third-generation AP, the Aironet 1252 Access Point is a business-class AP that supports draft two of the emerging 802.11n standard. 802.11n offers combined data rates up to 600 Mb/s using multiple-input multiple-output (MIMO) technology. The Cisco 1252 is available in a dual-band a/b/g or a single-band b/g radio configuration and can be deployed as a standalone AP (standalone) or as part of a unified (controller) wireless deployment. In order to offer maximum deployment flexibility, the Aironet 1252 Access Point is equipped with reverse-polarity threaded Neill-Concelman (RP-TNC) connectors for use with a variety of external 2.4- and 5-GHz antennas. In order to support the greater throughput rates offered by 802.11n, the Cisco 1252 incorporates a gigabit 10/100/1000 interface. The Cisco 1252 is designed to be deployed in challenging RF environments where high bandwidths are needed. Part numbers for the standalone version include: AIR-AP1252AG-x-K9 (dual-band) and AIR-AP1252G-x-K9 (single-band). Part numbers for the Cisco Unified Wireless versions include: AIR-LAP1252AG-x-K9 (dual-band) and AIR-LAP1252G-x-K9 (single-band), where x is the regional code.

� Cisco Aironet 1140 Series Access Point: a new, third-generation AP, the Aironet 1140 Series Access Point is a business-class AP that supports draft two of the emerging 802.11n standard. 802.11n offers combined data rates up to 600 Mb/s using MIMO technology. It is designed to be ceiling-mounted and makes use of an integrated dual-band antenna. The Aironet 1140 AG Series Access Point is available in a lightweight (CAPWAPP) version for implementation in centralized deployments based on Cisco WCS.

� Cisco 500 Series Wireless Express Access Point: single-band 802.11g AP for small or medium-sized businesses (SMBs) with carpeted offices or similar indoor environments. It is designed to be set to autonomous mode or to work in conjunction with Cisco 500 Wireless Express Mobility Controllers, with fewer features than the Cisco Unified Wireless Network Enterprise Solution.

Cisco CAPWAP APs The following model can be used only in controller-based topologies:

� Cisco Aironet 1500 Series Access Point: A dual-band AP specifically designed for outdoor, point-to-point, and multipoint MESH deployments. The 802.11a band is used for backhaul while the b/g band is used for wireless client access. The Aironet 1500 Series Access Point uses (patent-pending) Adaptive Wireless Path Protocol (AWPP) for optimal routing through MESH topologies. Although voice works in MESH deployment, the mesh design and the outdoor environments are very different from the VoWLAN specifications. For that reason, Cisco does not recommend running voice over MESH networks and does not support it.

www.CareerCert.info




Wireless LAN Controllers

The following summarizes the various Cisco WLCs and their features:

� Cisco 2100 Series WLC: a standalone WLC that supports up to 6 (2106 model), 12 (2112 model), or 25 (2125 model) APs, with 8 Fast Ethernet interfaces. Two of the Fast Ethernet interfaces can be used to power (802.3af) directly-connected APs. The interface can be configured as dot1q trunks to provide connection into the wired network. The Cisco 2100 Series WLC is ideal for small-to-medium-sized offices, where an H-REAP would otherwise be unsuitable because of the number of users, WAN requirements, or client roaming requirements.

� Cisco 4402 WLC: a standalone WLC that supports 12, 25, or 50 APs. It comes with two Gigabit Ethernet ports based on small form-factor pluggable (SFP) that can be configured as dot1q trunks to provide connection into the wired network, or the Gigabit ports can be link-aggregated to provide an EtherChannel connection to the switched network. This is ideal for medium-sized offices or buildings.

� Cisco 4404 WLC: a standalone WLC that supports 100 APs. It comes with four SFP-based Gigabit Ethernet ports that can be configured as dot1q trunks to provide connection into the wired network. The Gigabit ports can be link-aggregated to provide an EtherChannel connection to the switched network. This is ideal for large offices and buildings, and even small campuses.

� Cisco Wireless LAN Controller Module (WLCM): a WLC module specifically designed for the Cisco integrated service router (ISR) series. It is currently available in 6-, 8-, and 12-AP versions. The Cisco WLCM appears as an interface on the ISR router that can be configured as a dot1q trunk to provide routed connectivity to the wired network. This is ideal for small-to-medium-sized offices requiring integrated solutions.

www.CareerCert.info



� WS-C3750G: a WLC that supports either 25 or 50 APs and comes integrated with Cisco Catalyst 3750 Series Switches. The WLC’s backplane connections appear as two Gigabit Ethernet ports, which can be configured separately as dot1q trunks to provide a connection to the Catalyst 3750 Series Switch. Alternatively, the Gigabit ports can be link-aggregated to provide a single EtherChannel connection to the Catalyst 3750 Series Switch. Because the WLC is integrated directly, it has access to all of the advanced routing and switching features available in the stackable Catalyst 3750 Series Switch. It is ideal for medium-sized offices or buildings. The “’50-AP” version can scale up to 200 APs when four Catalyst 3750 Series Switches are stacked together as a virtual switch.

� Cisco Catalyst 6500 Series Wireless Services Module (WiSM): a WLC module specifically designed for the Cisco Catalyst 6500 Series. It is currently available in a 300-AP version. The Cisco WiSM can be associated to other Cisco Catalyst 6500 Series service modules and is ideal for networks requiring a centralized management solution.

www.CareerCert.info




Cisco Secure ACS

Cisco Secure Access Control Server (ACS) provides a centralized identity-networking solution and user-management experience across all Cisco devices and security-management applications. Cisco Secure ACS ensures enforcement of assigned policies by allowing network administrators to control the following:

� Who can log into the network or access to the network

� The privileges each user has in the network

� The accounting information recorded, in terms of security audits or account billing

� The access and command controls enabled for each configuration administrator

When a wireless IP phone tries to associate to the wireless infrastructure, the controller can delegate the authentication tasks to the secure ACS server, which acts as an authentication, authorization, and accounting (AAA) RADIUS server, allowing the association based on a successful authentication, and returning a specific profile the allowed user. This is very important in mixed environments, where data devices access the same SSIDs as voice devices. You can return a specific profile for voice devices, allowing a higher QoS value or specific filtering.

www.CareerCert.info




Cisco Unified Communication Manager

� Call processing

� Signaling and device control

� Dial plan administration

� Phone feature administration

� Directory services

� Programming Interface to external applications (API)

Cisco Unified Communications Manager (formerly Cisco Unified CallManager) is the core call-processing component of the Cisco Unified Communications solution. It is a program that can be installed on Windows or Linux and provides voice, video, mobility, and presence services for businesses with up to 30,000 users. Cisco Unified Communications Manager not only registers phones and provides telephony services, but also includes a suite of integrated voice applications and utilities, allowing analysis and reporting of call detail records, real-time application monitoring, and conference calls.

Cisco Unified Communications Manager has open telephony application programming interfaces (APIs) that support integration of additional communication services such as:

� Unified messaging: Enables users to access voice, fax, and text messages through a single email or telephone account

� Multimedia conferencing: Supports audio and video conferencing, typically from remote locations

� Collaborative contact centers: Integrates multiple communication tools, such as telephone, email, and the web

� Interactive multimedia response systems: Capable of detecting and responding with prerecorded or dynamically generated audio or video, typically to manage large volumes of interactions.

www.CareerCert.info




Cisco Unified Communications Manager Express

� Call processing based on Cisco IOS

� Integrated voice and data platform

� Cisco Unified Communications Manager integration

� Remote maintenance and troubleshooting

� Dual-line ephone-dns

� PBX capabilities

Cisco Unified Communications Manager Express is a call-processing application based on the Cisco IOS software, mainly for enterprise branch offices or small businesses. It offers fewer abilities, but also less complexity than the Cisco Unified Communications Manager program itself. It is an ideal solution for small or medium-sized organizations.

Cisco Unified Communications Manager Express allows Cisco ISRs to provide call processing for locally attached IP and analog phones. All the necessary files and configurations for IP phones are stored internally on the router, providing a single platform solution. In addition, the solution offers PSTN interfaces, WAN interfaces, integrated voice mail and automated attendant, and a full phone portfolio. Cisco IOS software supports H.323, SCCP, and SIP signaling, advanced QoS, and inter-working with an H.323 gatekeeper or SIP Proxy Server, all available for use with Cisco Unified Communications Manager Express deployments.

Cisco Unified Communications Manager Express also has the following features:

� Interoperability with Cisco Unified Communications Manager: administrators can deploy Cisco Unified Communications Manager at larger sites and deploy Cisco Unified Communications Manager Express at branch office locations where local call processing is required. Using H.323 or SIP trunking, calls can be routed over the WAN with the name and number information for the calling party, plus compressed voice for better WAN bandwidth utilization.

� Remote maintenance and troubleshooting: as Cisco Unified Communications Manager Express is a specific form of Cisco IOS for ISR routers, administrators can configure remote access to the router, and use the Cisco IOS command-line interface (CLI) to configure it. It is also possible to use a specific GUI to configure and administer the Cisco Unified Communications Manager Express.

www.CareerCert.info



Wireless IP Phones This topic describes some wireless IP phones commonly used in VoWLAN deployments.


Cisco Wireless IP Phone 7921G

� Second-generation wireless IP phone that supports a dual-band 802.11a/b/g radio

� Voice communications in conjunction with Cisco Unified Communications Manager and Cisco Unified Communications Manager Express

� Supports

– 802.11a, 802.11b, 802.11g

– G.711a, G.711u, G.729a, and G729ab

– SCCP, SRST, RTP/RTCP

– TSpec and U-APSD, PS-Poll

– CCX, CDP, XML

– 802.1X, Cisco LEAP, EAP-FAST, EAP-PEAP, WPA/WPA2, WEP. CKIP

The Cisco Unified Wireless IP Phone 7921G provides wireless voice communication over a network. Like traditional analog telephones, you can place and receive phone calls and access features such as hold, transfer, and speed dial. In addition, because the phone connects to your wireless LAN, you can place and receive phone calls from anywhere in your wireless environment.

The Wireless IP Phone 7921G is a second-generation wireless IP phone that supports 802.11b/g and 802.11a. It does not support 802.11n. This phone supports several subfamilies of the main codecs used in voice over wireless environments, namely G.711a, G.711 μ, G.729a, and G729ab. G.711a and G.711 μ are two slightly different versions of the same G.711; the “μ-law” is used primarily in North America, whereas the “a-law” is in use in most countries outside North America. G.729a and G729ab are two slightly different ways of obtaining a G.729 result, the main difference being that G.729ab supports some additional features such as voice activity detection (VAD) and comfort noise generation (CNG).

The following protocols are supported:

� SCCP

� Cisco Unified Communications Manager versions 4.3, 5.1, 6.0 and later

� Cisco Unified Communications Manager Express versions 4.3, 5.1, 6.0 and later

� Cisco Unified Survivable Remote Site Telephony (SRST) version 4.1 or later

� Cisco Compatible Extensions version 4.0 with WMM, traffic specification (TSpec), and unscheduled automatic power save delivery (U-APSD)

� Power-save poll (PS-Poll) with non-WMM-capable APs

www.CareerCert.info



� XML

� RTP and RTCP

� Cisco Discovery Protocol

The following wireless security features are supported:

� Authentication:

— Cisco Wireless Security Suite IEEE 802.1X Cisco LEAP authentication: Optional password prompt at power up

— EAP-FAST

— EAP-TLS

— EAP-PEAP

— WPA versions 1 and 2

— WPA-PSK versions 1 and 2

� Encryption:

— 40- and 128-bit static WEP

— TKIP and MIC

— AES

www.CareerCert.info




Cisco Wireless IP Phone 7921G (Cont.)

� Traditional telephony functionality

� Wireless access to corporate directory and phone services using speed-dial buttons

� Multi-line appearances

� Caller ID

� Call-handling features

� Profile choices

– Four network

– Personal

� Access to web-based services

� Support for five languages

The Wireless IP Phone 7921G provides traditional telephony functionality, such as call forwarding and transferring, call pickup, redialing, speed dialing, conference calling, and voice messaging system access. In addition, the Wireless IP Phone 7921G provides the following features:

� Wireless access to your corporate directory and phone services

� Multi-line appearances

� Speed-dial buttons using numeric keys

� Caller ID for incoming calls

� Call-handling features such as forwarding, transferring, holding, call parking, conference calling, call pickup, and group pickup

� Choice of four network profiles

� Choice of personal profiles for different environments, such as outdoors, or meetings

� Access to web-based services such as weather, stock reports, and phone directories

� Language support: English, French, Germany, Norwegian, and Japanese

www.CareerCert.info




Cisco Wireless IP Phone 7925G

� Similar features to the Wireless IP Phone 7921G (802.11a/b/g)

� More rugged

– Resistant to shocks

– Weatherproof

� Internal antenna

� Standard USB connector

� Bluetooth 2.0

The Cisco Unified Wireless IP Phone 7925G is designed for challenging environments, in which dust or humidity might damage other models or make them non-operational. The Wireless IP Phone 7925G complies with the following specifications:

� IP54: the phone is dust-protected (dust deposits are permitted, but their volume must not affect the function of the unit). The phone is also splash-protected. Water can be sprayed on the unit from any direction without affecting the functionality of the phone.

� Bluetooth 2.0: the phone has Bluetooth, and Bluetooth activity can occur during WLAN activity.

� Battery Life: the phone’s battery life depends on the battery used with it.

— Standard Battery: up to 150 hrs standby or up to 10 hrs talk time

— Extended Battery: up to 200 hrs standby or up to 13 hrs talk time

� Durability: the phone is built to resist up to a 5 ft (1.5 m) drop onto concrete without its carry case

� Restriction of Hazardous Substances Directive (RoHS): RoHS-compliant (and lead-free)

The Wireless IP Phone 7925G also supports the standard features described for the Wireless IP Phone 7921G.

www.CareerCert.info




SpectraLink� NetLink e340/i640 Wireless Telephone:

– 802.11b-only, Wired Equivalent Privacy (WEP), 40-bit and 128-bit, WPA-PSK

– Basic WMM

– SpectraLink SVP Server required: can also run a SCCP version (integrates with Cisco Unified Communications Manager and Cisco Unified Communications Manager Express)

� 8000 series:

– 802.11a/b/g, supports WEP, WPA/WPA2, PSK and enterprise

– WMM

– SpectraLink SVP Server required (no SCCP support)

Some other wireless IP phones can integrate into the Cisco Unified Wireless Network infrastructure: for example, some Polycom wireless IP phones in the SpectraLink series.

The SpectraLink NetLink e340/i640 Wireless Telephone operates over an 802.11b. It supports SCCP. The handset can register to Cisco Unified Communications Manager or Cisco Unified Communications Manager Express. A Cisco Unified Communications Manager template for the handset must be configured for the specific features and lines to be accessed by the handset. The handset can be automatically added to the network and registered with Cisco Unified Communications Manager, or each of these steps can be configured manually. After the handset is registered, it receives its configuration information from Cisco Unified Communications Manager.

More recent series, such as the SpectraLink 8000 series, offer a wider range of features and protocols supported, such as 802.11a//b/g radios and full WMM. These phones do not have SCCP support, and a SpectraLink SVP server is required. Cisco Unified Communications Manager or Cisco Unified Communications Manager Express are not required in that case.

The NetLink SVP Server is an Ethernet LAN device that provides call management and control for the SpectraLink phones. SpectraLink phones do not support QBSS or Cisco Compatible Extensions, and therefore rely on the SVP server for CAC.

www.CareerCert.info




Vocera Badges

� Push-to-talk badge-to-badge communication system

� Relies on a Vocera Server

� Phone-to-badge communication is possible with Cisco Unified Communications Manager

The Vocera Communications System enables wireless voice communication that users control with naturally spoken commands. The system is primarily targeted at hospitals, hotels, retail stores, and other in-building environments where mobile workers must stay in contact to perform their jobs. The Vocera Communications System consists of two key components:

� Vocera System Software: Controls and manages call activity

� Vocera Communications Badge: A lightweight, wearable, voice-controlled communication device that operates over a WLAN (IEEE 802.11b/g)

The Vocera Communications Badge is a small wireless device that provides a voice-controlled user interface for the Vocera Communications System. The Vocera Communications Badge enables instant, hands-free conversations among people throughout the workplace. It contains a speaker, microphone, wireless radio, and a backlit LCD that shows caller ID, text messages, and alerts.

The badges are centrally maintained by the Vocera Server from a single configuration file for all badges. Badges do not have keyboards, so this single configuration file is uploaded to all badges. Badges do not maintain static directory numbers or IDs (as a typical phone would have). Instead, each badge defines its identity when a user logs in. As a necessary consequence of this centralized management, all badge properties, including the SSID and security settings that allow it to connect to the network, must be the same. In turn, all APs to which the Vocera badge can connect must also share the same SSID and security settings. The Vocera server can link to Cisco Unified Communications Manager, allowing a network IP phone to dial an internal extension that relays to one, several, or all Vocera badges in the network.

www.CareerCert.info




Vocera Badges (Cont.)

802.11b/g, multicast and UDP unicast packets� Authentication:

– Open

– Wi-Fi Protected Access pre-shared key (WPA-PSK)

– WPA-PEAP, LEAP

� Encryption:

– 64/128 bit Wired Equivalent Privacy (WEP)

– Temporal Key Integrity Protocol (TKIP) / Message Integrity Check (MIC)

– Cisco Temporal Key Integrity Protocol

The Vocera badge supports the following wireless networks, and authentication and encryption capabilities:

� Wireless Network Support: IEEE 802.11b/g wireless network with multicast and UDP unicast packet delivery

� Authentication: Types supported include the following:

— Open

— WPA-PSK

— PEAP

— LEAP

� Encryption: Types supported include the following:

— 64/128 bit WEP

— TKIP

— MIC

— Cisco Key Integrity Protocol (shown as “CKIP” on many configuration pages)

www.CareerCert.info




Nokia Dual Mode Solutions for Enterprise

� Nokia has developed an SCCP client on Nokia dual-mode E-Series devices

� The Nokia E-Series natively supports SIP; SCCP needs to be added

� In VoWLAN, operates as an IP Phone with Cisco Unified Communications Manager

� In public GSM network, operates as a GSM phone

� Phones run on Symbian OS

� 802.11 b/g

Nokia E65 Nokia E51 Nokia E71

Nokia developed a “Dual Phone” line, called “Eseries”. The Nokia Eseries consists of several models, such as the Nokia E60, Nokia E61, Nokia E70 (first generation), Nokia E51, Nokia E66 and Nokia E71 (second generation). The Nokia Eseries phones are referred to as dual-mode, because they can operate in both cellular and Wi-Fi mode. Nokia designed the phones to operate with both radios on at the same time. This allows each user to choose a preferred method of calling as the default, but always have the option to choose either cellular or VoIP at the time of dialing. All 802.11b and 802.11g speeds are supported, which allows for use in most currently deployed wireless networks.

All Eseries phones run on Symbian OS, from version 6.0 for the first generation to version 9.2 for the latest. They include a variety of GSM frequencies and 3G (Wideband CDMA [WCDMA]) cellular network support for seamless roaming across different countries, as well as Wi-Fi, Bluetooth, and infrared plus USB 2.0 compatibility.

By default, an Eseries phone ships with a SIP client. If used as it is with Cisco Unified Communications Manager, it is seen as a third-party phone, with only standard RFC-based features (call, hold, blind transfer, consult transfer, conference). By installing a SCCP client on the Nokia phone, it is possible to integrate it into the Cisco Unified Communications infrastructure and to use all the SCCP features available through the phone softkeys.

www.CareerCert.info




Cellular and Wireless Mobility

A “single-mode phone” is only used on the 802.11 wireless networks. Nokia Eseries phones also utilizes the same wireless network, but are considered a dual-mode devices. A dual-mode phone has the capability to run on both a mobile cellular network and the WLAN network. Dual-mode phones can offer additional value to customers due to their flexible usage, increased productivity, and potential for return on investment (ROI).

When outside the corporate network, the phone connects to the cellular operator network. It can also connect to Cisco Unified Communications Manager using a VPN connection via the General Packet Radio Service (GPRS) network. Once inside the corporate network, the phone registers through the internal wireless network to Cisco Unified Communications Manager and behaves like any other internal VoIP phone.

www.CareerCert.info




Transferring Between VoWLAN and GSM

Nokia supports roaming from Wi-Fi to cellular only with the use of Cisco Unified Mobility Manager� Bridges the gap between public and enterprise calling

� Requires the use of a Cisco Unified phone that supports XML

� Both GSM and 802.11 radios are active

� Single number reach pinned to network

An integral concept for wireless networks is, of course, device roaming. It is important to understand what roaming is, how and when it occurs, what types of roaming there are, and how the types differ.

For 802.11 data networks, roaming also refers to physical movement, but it is often associated with data connectivity while physically moving. This type of roaming does not include moving from the 802.11 environment to the cellular network. The present Nokia Eseries generation does not support automatic roaming from Wi-Fi to cellular.

If you are on an active VoWLAN call and roam out of coverage area, your call will be disconnected. When this occurs both your phone conversation and registration to Cisco Unified Communications Manager will most likely be lost. To regain connectivity, simply roam back into coverage area and, assuming you are not using manual registration, the phone should eventually re-register with Cisco Unified Communications Manager. The time it takes to re-register can range from several seconds to several minutes. To actively roam from the 802.11 network to the GSM network, it is possible to transfer the voice conversation from the 802.11 internal network to the GSM network by pressing a series of button on your phone to switch the phone from WLAN to GSM. You can then move safely outside the 802.11 cell. Provided you have the agreement of the mobile GSM provider, it is also possible to switch from GSM to the internal WLAN.

The next generation of Nokia Eseries will integrate automated roaming features between WLAN and GSM networks.

www.CareerCert.info





Summary

� A wireless VoIP phone needs to associate to the wireless infrastructure, just like any other wireless device, before registering to a Voice Control Server, such as Cisco Unified Communications Manager.

� Several processes and protocols work in the wireless and wired infrastructure to make sure that phone packets get prioritized and reach their destinations as fast as possible.

� The wireless infrastructure comprises access points and controllers; the voice infrastructure includes the Cisco Unified Communications Manager system or Cisco Unified Communications Manager Express system

� Wireless phones include the Cisco Unified Wireless IP Phone 7921G, as well as other brands such as SpectraLink phones, Vocera badges, and Nokia Eseries phones.

www.CareerCert.info



www.CareerCert.info


Lesson 3

Designing Wireless for Voice

Overview The packet loss and jitter requirements of VoIP and the increased mobility and increased expectations of wireless phone users place demands on connection quality and coverage that are beyond the performance capacity of typical wireless data deployment. RF planning, design, and implementation are key to a successful voice over wireless deployment. Correctly designing, planning, implementing, operating, and maintaining the WLAN RF environment is critical. This lesson provides you with the fundamental considerations to address when designing a wireless LAN for voice applications.

Objectives Upon completing this lesson, you will be able to outline the design considerations for different wireless network implementations. This ability includes being able to meet these objectives:

Describe the general site survey guidelines for VoWLAN

Describe RF design guidelines for VoWLANs

Describe combined WLAN services

Describe VoWLAN security

Describe VoWLAN roaming

Describe the benefits of Cisco Compatible Extensions for VoWLAN roaming

Describe controller position and configuration parameters for VoWLAN optimal roaming

Describe campus design requirements for supporting VoWLANs

Describe voice support in mesh networks

www.CareerCert.info



General Site Survey Guidelines This topic describes the general principles involved in designing a voice over wireless LAN (VoWLAN) network.


Methodology

VoWLANs require careful RF planning. A thorough voice site survey is often required in order to determine the proper levels of wireless coverage and to identify sources of interference. Access point (AP) placement and antenna selection choices can be made much more easily with the help of the results of a valid voice site survey.

Undoubtedly, the most important consideration is the transmit power of each wireless phone. Ideally, each phone will learn the transmit power of its AP and adjust its transmit power to that of the AP.

Although the majority of the wireless networks today are deployed after extensive RF site surveys, those surveys are generally performed with data service in mind, not voice. VoWLAN phones are likely to have different roaming characteristics and different coverage requirements than those of a typical WLAN adapter for a mobile data client such as a laptop. Therefore, an additional site survey for voice is often recommended in order to prepare for the performance requirements of multiple VoWLAN clients. This additional survey also provides an opportunity to tune the APs to ensure that the VoWLAN phones have enough RF coverage and bandwidth to provide proper voice quality. Voice over wireless deployment should be organized in four phases:

1. Assess the network requirements: Which types of devices need to be supported wirelessly? Will the VoWLAN only be used for wireless phones? Will data be supported? What about data collection terminals? Will users be stationary or on the move while using the VoWLAN? These questions are important answer when determining what the backbone needs are. If users are mobile, good hand-offs during client roaming are a must. If wireless phones are being used, then the wireless network needs to be built with better receive signal strength for quality of service (QoS) and good hand-offs between APs. How robust you

www.CareerCert.info



need the wireless network to be depends on what types of VoWLAN devices it has to support.

2. Determine AP count and placement: The number of users and devices supported and the bandwidth desired determines the number of APs that are needed. Many tools are available today to determine AP placement. For example, the Cisco Wireless Control System (WCS) has a planning mode feature under the Maps menu that allows Cisco WCS to estimate how many APs are needed for a given floor in a building.1 The planning mode feature takes into account the following when predicting range and coverage:

Protocol IEEE 802.11b/g or IEEE 802.11a

Coverage or capacity

Throughput

Square feet

3. Deploy APs: Proper installation of APs and antennas is very important. APs should be installed as surveyed. If an AP is using external antennas, then the antennas should also be installed as surveyed. Antenna installation is crucial to recreate the RF propagation as surveyed.

4. Verify the deployment: If using a Cisco Unified Wireless Network solution with controllers, you should let the WLAN controllers automatically manage the APs for power and channel upon completion of installation. As a final step, it is crucial that you conduct verification tests to ensure that everything works as desired.

1 The WCS planning tool does not replace a site survey. It can provide an estimation of the number of access points needed, but cannot take into account the specifics of the building, such as sources of interference, reflection patterns, and so on. A site survey is a vital part of the access point count and placement determination process.

www.CareerCert.info




Customer RequirementsProtocol requirements

802.11b/g802.11a

Client devicesPhonePDAWireless NIC

Client densityTypicalPeak

Security level802.1X EAP authentication typeEncryption methodLegacy devices (WEP)

Which protocols and spectrum does the customer require?

Which protocols are required? Do you have the choice of the spectrum to use, or does the customer already have existing devices, both for voice and data?

802.11b/g

802.11a

What client devices are present in the network?

Phone

PDA

Wireless network interface card (NIC)

What is the expected client density?

Typical

Peak

What security requirements does the customer have?

802.1X Extensible Authentication Protocol (EAP) authentication type

Encryption method

Are there legacy devices (Wired Equivalent Privacy [WEP])

www.CareerCert.info




General Signal Level for VoWLANSignal of –67 dBmor higherPacket Error Rate (PER) no higher than 1% Minimum SNR of 25 dB (–92 dBmnoise level)

Adding signal does not always increase SNR.

The general principle in VoWLAN design is that the signal level should be high. The lower the signal level, the greater the chance that the signal will have to be resent, causing delay, jitter, and poor voice quality (according to a mean opinion score [MOS]). A general recommendation for Cisco Wireless IP phones is to set the edge of the cell at a signal level of –67 dBm. In other words, you should design the cells so that when a client gets to the area where the Received Signal Strength Indicator (RSSI) is –67 dBm, it is leaving the cell area as you designed it.2 Prior to this point, this client should already have heard the signal from another AP, in order to roam smoothly.

The RSSI is, of course, only one of the criteria to take into consideration. Another important one is the noise level. The –67 dBm recommendation is valid in a normal office environment where the noise floor is at about –94 dBm. A client always tries to use a given speed depending on RSSI and signal-to-noise ratio (SNR) thresholds. If the signal coming from the AP is loud, the RSSI will be high. If the noise level is also high, the client will revert to a lower speed to avoid losses. It is therefore very important to take care both of the RSSI level and of the noise level at any given point of the wireless network.

The size of the cell can be modified by changing the power level on the AP. Increasing the power level does not always increase the SNR. It does increase the signal strength at a given point, but it also increases multipath issues. When the power level is increased on several APs, the co-channel interference also increases, thus the global impact on the SNR value is not positive.

2 Keep in mind that the RSSI is relative to vendor implementations. The values given have been calculated using the Cisco Unified Wireless IP Phone 7921G as a reference.

www.CareerCert.info




Data Rate(Mb/s)

Data Cell

+15 dB recom

mended from

cutoff values

VoWLAN CellMinimum

SignalStrength(dBm)

MinimumSNR(dB)

MinimumSignal

Strength(dBm)

Minimum SNR(dB)

54 -71 25 -56 40

36 -73 18 -58 33

24 -77 12 -62 27

18 -79 11 -65 26

11 or 12 -82 10 -67 25

9 -85 9 -71 24

5.5 or 6 -89 8 -74 23

2 -91 6 -76 21

1 -94 4 -79 19

Recommended SNR Values

The RSSI is intended to be used as a ‘relative value’ within the client card chipset. It is a 1-byte value, so that it can have any value from 0 to 255, but vendors prefer to use arbitrary scales from 0 to a vendor-specific maximum value (for instance, Cisco uses 101, some other vendors 60 or 37). This number is not associated with any particular power scale (such as milliwatts) and is not required to be of any particular accuracy or precision. The RSSI value is used internally by the microcode in the adapter and this is why vendors are not forced to use a compatible standard.

This vendor-related understanding of the RSSI value implies that the each wireless device has a different sensitivity and different RSSI and SNR requirements. When surveying a wireless network for voice, you should always use the device intended for the deployment, along with the vendor specifications for which RSSI and SNR values are required to achieve which speeds. The values given in the table above are relevant to Cisco clients, such as the Cisco Unified Wireless IP Phone 7921G or the Cisco Aironet IEEE 802.11a/b/g Wireless LAN Client Adapter.3

Each client needs a minimum signal strength and SNR to perform at a given speed. A Cisco card needs at least a 10-dB SNR to achieve an 11 Mb/s data rate or more, provided that the signal strength is –82 dBm or better. To achieve a higher data rate, both the signal strength and SNR must increase. In a VoWLAN deployment, a general principle is that the required SNR should be 15 dB higher than the minimum SNR required for data transmission. This requirement ensures that the signal quality will be optimum at any given speed.

3 Although the antennas of the Cisco Aironet IEEE 802.11a/b/g Wireless LAN Client Adapters are different from the antennas of the Wireless IP Phone 7921G, both devices have the same requirements for RSSI and SNR.

www.CareerCert.info




Cell OverlapCell overlap is designed so that when a VoWLAN device gets to the –67 dBm area, it is already in good range of another access point.20-percent overlap between cells is recommended

One of the requirements of VoIP deployments is for high availability. The same high-availability strategies used in wired networks can be applied to the wired components of the VoWLAN solution. One area unique to the VoWLAN availability is RF coverage high availability, providing RF coverage that is not dependent upon a single WLAN radio.

The primary method for providing RF high availability is cell boundary overlap. An overlap of 20 percent is recommended between cells. The purpose of the 20-percent overlap is to ensure that a VoWLAN handset can detect and connect to alternative APs when it is close to the cell boundary. This should allow a VoWLAN client to change AP associations with a minimum of interruption to a call, by minimizing the amount of data rate shifting, retransmissions and scanning needs at a cell boundary. This 20-percent overlap requirement means that APs are spaced closer together than the distance of two times the radius suggested by the cell boundary. The area of overlap between two circles, each of which has a radius equal to 1, is given by:

X = R² / 2R

In this equation, X is the distance between the centers of each circle. Solving for an area of 20 percent gives an X value of 1.374 for a standard radius of 1.4

This means that if the edge of the cell is 70 feet (21 m) from the AP, the next AP should be 96 feet (29 m) away from the first AP.

4 Other common values are 10 percent (1.611), 15 percent (1.486), 25 percent (1.269), and 30 percent (1.198).

www.CareerCert.info




AP and Client Power LevelsAP cell size is related to transmit powerAP power level should match client power levelSurvey should be done at lower power to account for coverage holes

The overlap is a percentage, which means that it is related to the size of the cell. Therefore, you need to determine what the optimal size of the cell should be. This size, in turn, depends on two factors:

the expected signal level at the edge of the cell

the required speed at the edge of the cell

The expected signal level at the edge of the cell depends mainly on the power level. An important factor to keep in mind is that the AP power level should match the client power level. The Cisco Unified Wireless Network solution can use a feature called Auto-RF, by which an AP can detect if WLAN clients are experiencing poor SNR values and increase its power accordingly in order to rectify the SNR issues.

If the VoWLAN handset has a maximum power of, for example, 40 milliwatts (mW), and the AP planning was based on an AP power of 40 mW, increasing the AP power to 100 mW to cover an RF hole does not help a VoWLAN phone in that hole. The client will clearly receive the signal from the AP, but the SNR of the client answer to the AP will not necessarily be symmetrically improved. In this scenario, the wireless network should probably be designed at 20 mW, to account for possible power increases for hole coverage of up to 40 mW.

www.CareerCert.info




Distance and SpeedSpeed, power level and distances are relatedCloser clients are at less risk of multipath issues

Example not to scale

The cell size is also direct related to the speed needed at the edge. Each client determines the RSSI of the AP, compares it to the SNR value, and decides at what speed the next frame should be sent. In a voice environment, high data rates at the edge of the cell are required. A wireless phone sending from the 1 Mb/s area encounters far more collisions than a client in the 54 Mb/s area creating retransmissions. The same wireless phone in the 1 Mb/s area also suffers more from multipath issues, and is more susceptible to changes in environmental conditions due to moving obstacles or sources of degradations affecting its received and sent frames. In addition, the slower the date rate, the slower the transmission, creating longer wait times for the next packet. These are a few of the reasons why the minimum speed allowed in the cell is usually higher than the 802.11 minimum.

When 802.11b and 802.11g are used, the minimum speed should be set to 11 Mb/s. No 802.11g speed is set to mandatory so that 802.11b clients can still associate. This configuration allows for fewer simultaneous calls in the cell, because the protection mechanism linked to 802.11g coexistence with 802.11b reduces the average speed of the cell. This configuration allows for fewer simultaneous calls in the cell, because the protection mechanism linked to 802.11g coexistence with 802.11b reduces the average speed of the cell. Where a pure 802.11g cell would allow 23 to 23 Mb/s throughput for 54 Mb/s cell speed, a hybrid cell degrades the average throughput to 8 Mb/s. For this reason, be aware that you should design for fewer concurrent calls, in order in order to avoid delay and jitter issues.

For each band, the AP communicates the speeds it supports in its beacons. With the 802.11b protocol, they are 1, 2, 5.5, and 11 Mb/s. With the 802.11g and 802.11a protocols, they are 6, 9, 12, 18, 24, 36, 48, and 54 Mb/s. Designing a minimum speed means forbidding lower data rates.

The above illustration is an example of the relation between speed and distance representative of an open environment. These values are not absolutes and may vary depending on the actual AP and antenna parameters.

www.CareerCert.info




Power Levels and Client DensityReducing cell size by matching client power and disabling low speed decreases the user number per cellEach user has better throughput and less risk of encountering interferenceCoverage holes might cause probing devices to disturb the WLAN clients

Reducing the cell size and disabling lower data rates decreases the user density per cell. A direct consequence is that each user benefits from a better connection, better access to the medium and better throughput.

You still have to be aware of the consequences of disabling lower data rates. Access point broadcast messages, such as beacons, are sent at the lowest mandatory speed. This means that beacons will be sent at 24 Mb/s. Clients far away from the AP that are attempting to discover the network may not hear this beacon. They might hear messages sent at 1 Mb/s, but this slow speed is disabled. These clients will try to discover the network by sending probe requests, at their lowest mandatory speed, which will probably be 1 or 6 Mb/s. The AP will receive these probe requests, but will not be allowed answers at these speeds. The clients will therefore not learn about the existing networks and will continue probing, thus colliding with some of the clients in the cell.

When designing such high-efficiency cells, always make sure that you do not leave coverage holes where clients may enter the network, such as lobby areas and public spaces, or potential new clients in these areas may interfere with the network.

When respecting these design recommendations, each client then has good access to the wireless resources. Cell size limitations are also linked to the application used. VoWLAN has strict cell design requirements to ensure that enough space is given to each call.

www.CareerCert.info




Minimum Data RatesA minimum 24 Mb/s speed is recommendedWhen using 802.11b/g, 11/12 as mandatory is possibleBeacons are sent at the lowest mandatory speedSize cell targets vary based on the standard:– 7-8 802.11b

conversations– 14 802.11g

conversations– 20 802.11a

conversations

Suppose wireless phones using the G.711 coder-decoder (codec). Each phone needs 64 kb/s both ways for callers to experience full-duplex communication. With the addition of the 802.11 overhead, the acknowledgements and other traffic-control-related requirements, about 200 kb/s is required per wireless phone. Suppose you design a cell offering only the highest speed, 54 Mb/s. The actual total throughput of the cell, after removing management frames (such as beacons and probes) and other overhead is around 22 to 23 Mb/s.

It may be tempting to deduce that 22 Mb/s means that 120 wireless phones needing 180 kb/s each can simultaneously communicate in the cell. This reasoning is far from valid. In a CSMA/CA environment, even with QoS enabled in the cell, each device picks up a random backoff timer before sending each frame. Because the timer is random, the more devices the cell contains, the more often collisions occur. When a frame is not acknowledged, the sender picks another backoff timer, and then resends the frame.

The 802.11 mechanism determines that if collisions occur, it implies that too many devices are trying to send at the same time. Statistically, if each device picks another backoff timer in the same range, at the scale of all the devices present in the cell, this does not change anything to this “overloading the cell capacity” situation. The same –too high- number of devices try to send at the same time.

To solve the issue, the 802.11 mechanism dictates that when collision occur, the sending devices pick a second backoff timer that must be in a larger range then the first one. If the first number was a random number between 1 and 31 slot units, the second will be a random value between 1 and 63 slot units, and so on. This means that every time a collision occurs, communication slows down.

www.CareerCert.info



As voice is a time-sensitive type of traffic, collisions must be minimized. It is often thought that any packet loss is unacceptable for voice, but this notion does not take into account the way most voice devices compensate for a missing voice sample when reconstructing the voice stream. For G.711 in wireless, a good objective is to achieve less than 1% loss—more than 5% loss will mean an unacceptable result. These values refer to real loss; that is to say, they refer to packets that never reach the receiving endpoints.

When a collision occurs and a packet is lost at Layer 2, the sending station picks a new backoff timer and then resends it. How many times can this process occur for a given packet? The answer is highly dependent on the sending device vendor.

In any case, 50 frames per second of streaming audio should be sent, and 50 frames per second received. At 54 Mb/s, sending a G.711 frame takes less than 0.034 milliseconds, so the main issue is linked to the contention due to the other devices sending, rather than to the time it actually takes to send the frame itself. In good conditions, several voice packets, representing, for example, 100 ms worth of voice sound, can be buffered in the sending device before being sent, one after the other, in a few milliseconds. They will be buffered on the other end before being played at the original voice speed, and no issue will be noticeable.

The problem occurs when the sending device cannot send within a reasonable delay. The backoff timer is increased by the contention due to the other devices sending in the cell. For example, if a device has to wait a total of 10 millisecond before sending a frame, then if this frame collides and if the phone has to wait 14 more milliseconds before having a chance to resend the same frame, the phone operating system may examine the frame to decide if it is better to keep on trying to send the same frame, or if it would be more efficient to drop this frame and send the next one instead. It might be better to have a void in the voice speech than to excessively delay other voice packets. If this frame is delayed, there is a good chance that the next one will be delayed as well, and this might completely interrupt the voice flow. The algorithm deciding which action to take depends on the vendor, but it is common to see that if a frame does not reach the AP within 30 milliseconds, it is dropped and the next one is sent. With 100 devices competing to get access to the cell, this delay is exceeded all the time. The maximum number of devices in the cell is therefore a combination of speeds and collision risks. This can be determined by capturing frames with a packet sniffer and monitoring collisions as new devices are added to the cell.

Cisco tests show that the following targets are reasonable in properly designed cells:

802.11a: 20 simultaneous voice conversations

802.11g: 14 simultaneous voice conversations

802.11b: 7 to 8 simultaneous voice conversations

Care must be taken to design the cell with minimal same-channel overlap to achieve these voice stream numbers. What is defined as a conversation is the communication flows between one wireless IP phone and one AP. A conversation is made of two voice flows, one upstream flow, from the phone to the AP, and one downstream flow, from the AP to the phone.

www.CareerCert.info



For the same reason, the minimum speed at the edge of the cells is usually set to 24 Mb/s. When defining the edge of a cell and deciding to limit it to 24 Mb/s, a good practice is to disable all speeds below 24 Mb/s (namely, disabling 1, 2, 5.5, 6, 9, 11, 12, and 18 Mb/s in a 802.11g environment, and disabling 6, 9, 12, and 18 Mb/s in a 802.11a environment). Any lowering of the data-rate below that used in the RF design extends the AP cell size, increases co-channel interference, and reduces call capacity. Higher data rates than the site survey rate can be enabled, but a VoWLAN client might not take advantage of these rates because some VoWLANs prefer not to shift data rates.

These parameters promote a good wireless network design for voice applications, with APs set at the power level of the clients, the roaming zone designed to be in the –67 dBm area, and 20 percent overlap between cells, in a normal environment with a low noise level.

www.CareerCert.info




Signal Attenuation

Object in Signal Path Signal Attenuation Through Object

Plasterboard wall 3 dB

Glass wall with metal frame 6 dB

Cinderblock wall 4 dB

Office window 3 dB

Metal door 6 dB

Metal door in brick wall 12 dB

Phone and head position 3 - 6 dB

Signal attenuation and signal loss occur even as a signal passes through air. The loss of signal strength is more pronounced as the signal passes through different objects. A transmitted power of 20 mW is equivalent to 13 dBm. Therefore, if the transmitted power at the entry point of a plasterboard wall is at 13 dBm, the signal strength will be reduced to 10 dBm when exiting that wall. This table shows the likely loss in signal strength caused by various types of objects.

Each site surveyed will have different levels of multipath distortion, signal loss, and signal noise. Hospitals are typically the most challenging environment to survey due to high multipath distortion, signal loss, and signal noise. Hospitals take longer to survey, require a denser population of APs, and require higher performance standards. Manufacturing and shop floors are the next hardest to survey. These sites generally have metal siding and many metal objects on the floor, resulting in reflected signals that recreate multipath distortion. Office buildings and hospitality sites generally have high signal attenuation but a lesser degree of multipath distortion.

User position also plays an important part in the signal level. Wireless signals are usually vertically polarized, and phones are designed to be used in a given range of positions. A phone lifted 90 degrees from its expected position may lose up to 6 dBm of RSSI. If the user’s head is between the phone and antenna, it may also impact the call quality. Last, the type of antenna is important. An internal antenna may be more affected by obstacles than an external antenna.

www.CareerCert.info




Coverage Requirements

When designing a voice over wireless network, the required coverage needs to be taken into consideration. Areas should be determined where voice coverage is mandatory, desired, or not supported.

Then, in each area where VoWLAN coverage is needed, the user density should be evaluated. The area type plays an important role at that step. For example, it is very unlikely that a medium-sized meeting room will need to be heavily covered. Although users will use data connections in the wireless network from a meeting room, surveys show that most users leave a crowded meeting room when placing or receiving a call. On the other hand, while corridors need only light coverage for wireless data, they will be heavily used by voice users. Many voice users walk along corridors as they speak, which has an impact on roaming issues. Other unusual places for wireless data also need to be covered, such as elevator shafts, stairwells, even bathrooms.

www.CareerCert.info




Elevator Shaft Coverage Solutions

Elevator shafts are a specific case in voice over wireless coverage. Users usually enter the elevator and leave the cell created by an AP on the floor. Then they travel vertically and this travel pattern must be taken into consideration when designing the network. Several possibilities exist. However, each must be checked for local regulatory compliance:

Rely on the floor coverage: In this configuration, no special equipment is set for the elevator shafts; users associate to the APs on the floors and roam from an AP on one floor to an AP on another floor as they travel. This is obviously the least reliable method, and results depend greatly on the building material and the distance between the closest APs and the elevator shafts. When designed and tested carefully, this solution does work; but without careful planning, users commonly loose the connection as they travel.

Install a directional antenna at the top of the shaft: With this system, users associate and stay connected to the same AP while they travel in the elevator. This system is limited by the range and asymmetric signal issues. When a user moves away from the highest floor, the distance from the antenna to the user increases. After a certain distance the antenna may not be able to hear the user’s phone well. The roof of the elevator cabin, usually metallic, increases this problem. However, this solution is still efficiency when the building is not too high. Always check local regulations before choosing this solution. For example, it is forbidden in many European countries to install any object inside the elevator shaft above the cabin, for safety reasons (falling objects).

www.CareerCert.info



Install an AP inside the cabin: This is possible in some configurations for which the cabin is equipped with an emergency phone. These phones are usually connected using a cable that runs in a tube along the cabin main cable. It is possible to add a second Category 5 cable to power and connect and AP, located in the cabin ceiling. This solution offers very good results, but presents several limitations. The first one is regulatory, as some countries forbid the addition of a second cable to the communication tube. The second is technical, as the overall length of the cable, from the AP to the switch, must not exceed 100m, which is sometimes challenging. The third is linked to the lifetime of the cable: as it runs along the elevator cable, it is rolled many times, and its lifetime is shorter than a cable in a fixed position.

Use a leaky cable: this solution is not recommended, and is not supported by the 802.11 protocol. It still offers good results when carefully planned. A leaky cable is a cable that is not isolated, through which 802.11 signal leaks. You can install an AP at the top of the elevator shaft and plug such a cable into the AP. The cable runs down in a corner of the elevator shaft. As users enter the cabin, they roam to the AP on the top of the shaft, as the “antenna” (the cable) is a few feet away. As they travel, and as the cable runs all the way down the elevator shaft, the “antenna” stays at the same distance from the user, a few feet, and the connection is maintained.

Whichever solution you choose, voice in elevators should be planned as a specific issue, and solutions should always been tested on site.

www.CareerCert.info



RF Design Guidelines This topic gives design recommendations for the 802.11b/g and 802.11a spectrum cell design.


2.4-GHz Network Design

A total of 14 channels are defined in the 802.11b/g channel set. Each channel is 22 MHz wide (even though 802.11g uses only 20 MHz of this 22-MHz space), but the channel separation is only 5 MHz. This leads to channel overlap, such that signals from neighboring channels can interfere with each other. In a 14-channel DS system, there are only three non-overlapping (and thus, non-interfering) channels: for example 1, 6, and 11, each with 25 MHz of separation. This channel spacing governs the use and allocation of channels in a multi-AP environment.

The 802.11b/g protocol also defines that channel separation should be 30 dB. Some countries, being allowed 13 of the 14 possible channels, determined that, given the power levels in use in their regulatory domains, it would be possible to use 4 channels instead of 3. They thus use 1, 5, 9, and 13. The consequence is a slightly higher noise floor in each channel. Although this noise level is perfectly acceptable for data, noise is one of the critical issues for voice deployment. Therefore, you should only use a design with three non-overlapping channels in VoWLANs.

The guidance for the Cisco Unified Wireless IP Phone 7921G VoWLAN handset is for a power level boundary of –67 dBm, and a separation between adjacent AP channels of –86 dBm. The –67 dBm requirement is to minimize packet loss, and the –86 dBm requirement is to minimize co-channel interference from other AP cells on the same channel.

www.CareerCert.info




2.4-GHz Network Design (Cont.)

Clear Channel Assessment can rely on Layer 2 or Layer 1 information.

Using three channels presents a challenge when trying to deploy APs, to ensure that an AP on a given channel cannot see the signal from another AP using the same channel. The AP coverage radius changes with the AP transmission power, the minimum allowed speed at the edge of the cell, and the boundary created by this radius is often considered the boundary of the AP.

The reality is somewhat more complicated because the AP influences the WLAN RF environment around it for a much greater distance than just the bit-rate boundary. This is because the RF energy from the AP, although too low to be demodulated in to a WLAN frame, is strong enough to cause an 802.11 radio to defer sending. In addition to the AP’s influence on the RF environment, the clients associated with that AP extend the range of the RF energy associated with that AP’s cell even further.

Before sending, each client performs a Clear Channel Assessment (CCA). If the CCA fails, the client does not send. On a theoretical level, a client counts down from a random value, and each time the counter decreases by one, the client performs a CCA. If another device starts sending at that time, the detected frame header contains the frame duration and the backoff timer is increased with the value of the frame duration. In most clients though, the CCA mechanism is typically triggered either by a simple raw energy level and physical layer convergence protocol (PLCP) header power levels, or by carrier detection. The CCA of an 802.11 radio does not vary with the bit rates being used and is not, generally, user-configurable. This means that if the client detects a noise, it defers sending its frame, even if the noise is not a readable 802.11 header.

www.CareerCert.info



The impact of CCA deferrals from radios that are not part of the local cell is called co-channel interference. As co-channel interference results in delays in sending frames, it causes increased jitter and delay. Although WLAN QoS prioritizes WLAN traffic, this occurs after the CCA and therefore prioritization does not overcome the jitter and delay introduced by CCA. This is the reason why the guidance for the Wireless IP Phone 7921G VoWLAN handset is for a power level boundary of –67 dBm and a separation between adjacent AP channels of –86 dBm.

The RF environment shown above is an example of the two boundaries created by the –67 dBm and – 86 dBm requirements, based on standard RF loss formulas for an open office environment. It provides an AP a client radius of 43 feet (13 m), with an AP co-channel interference radius of 150 feet (46 m) using standard antenna gain (2 dB) and an AP output power of 16 dBm (40mW). Different RF environments, AP powers, and antennas will result in different client and co-channel interference radii, but the principles discussed here will generally hold. Decreasing the power to 20 mW reduces the co-channel interference radius to 130 feet (40 m) and the client radius to 38 feet (12 m), and also reduces the co-channel interference proportionally to the co-channel interference generated by an AP.

www.CareerCert.info




2.4-GHz Network Design (Cont.)

Increasing the cell radius by allowing lower speeds also extends the client energy radius; this decreases overall cell speed and causes data rate shifting issues

The nominal bit rate for the Wireless IP Phone 7921G is approximately 24 Mb/s or greater, depending upon noise. The AP radius can be extended further by supporting lower bit rates. This is not recommended for the following reasons:

Lowering the bit rate extends the AP client radius, and therefore also increases the client co-channel interference radius (as the client can receive and transmit from a point farther from the previous AP cell edge), increasing the area that only has the VoWLAN call capacity of a single AP.

The lower bit rates reduce the overall call cell capacity, as packets with lower bit rates consume more time and transmit fewer packets.

VoWLAN call quality is sensitive to data-rate shifting. The decision to perform a data-rate shift is normally the result of being unable to send at the date rate previously used, which is determined by sending a frame multiple times without receiving an acknowledgement for it. This increases the delay and jitter experienced by a VoWLAN call.

www.CareerCert.info




2.4-GHz Network Design (Cont.)In open space, even though cell size is reduced by limiting allowed speeds, the AP footprint exceeds the cell edge. Co-channel interference renders the –86 dBmseparation objective difficult to achieve.Inner walls may help reduce AP footprint influence.In worst-case scenarios, only three channels are available for the whole floor.

When designing a WLAN network for voice, the AP power level is limited to match the client power level, and a margin is taken to account for coverage holes. This reduces the overall size of the cell and the RF footprint of the AP on the floor.

The AP layout within a building depends greatly on the building construction and shape, as well as the WLAN coverage requirements in the building. Due to differing effects of implementation-specific variables, there is not a single recommended deployment for the number of APs that should be deployed or a single solution for determining the effect of co-channel interference.

Keep in mind during your initial predeployment site survey or when migrating a network from data only to VoWLAN, that the AP footprint exceeds its effective data coverage. In the example shown here, the AP client radius is far less than the AP RF radius. This means that the APs using channel 1 (white) are effectively sharing some channel capacity. The two channel 1 APs increase the coverage over a single AP by two times, but do not increased the capacity by the same ratio and may not increase the capacity significantly in comparison with a single AP. The same is true for the APs on other channels. Due to co-channel interference, the call capacity of the floor is equivalent to something above the capacity of 3 independent APs, but not approaching the capacity of 6 APs. This is the primary reason for addressing VoWLAN call capacity in terms of the number of calls per channel, rather than the number of calls per AP. Channel capacity is the limiting factor.

In most real-life scenarios, obstacles such as walls reduce the effective RF footprint of each AP. In situations where inner walls do not isolate cells, the objective of achieving a –86 dBm separation between cells on the same channel may be difficult to achieve. Designing the cell based on client power is still preferred. It might be tempting to reduce the AP power level further, but it is often difficult to reduce the client power level. If the client power level remains the same, reducing the AP power does not solve the overlapping issue.

www.CareerCert.info




2.4-GHz Network Design (Cont.)Co-channel interference is a 3-D phenomenon.Interfloor material may not isolate APs.Using Auto-RF is recommended, but co-channel interferences must be taken into consideration when testing the VoWLANquality.

In a multi-floor building, the issue is worsened by the fact that RF energy can travel between floors. As part of RF planning, the channels are staggered from floor to floor to minimize the co-channel interference between floors.

As the signal path between the floors is different from that on the same floor (there is often reinforced concrete in the between-floor path), this must be taken into account when considering the co-channel interference radius of an AP. The loss from one floor to another must be measured before deciding how many floors are affected by the signal coming from a given AP. Co-channel interference can impact just a floor above and below, or several of them. In the example shown, the co-channel interference between floors is still significant and it is reasonable to assume that the capacity across the three floors may be the equivalent of four, five, or six APs, but not close to that of the 9 APs that have been deployed.

In a controller-based solution, the Auto-RF features help mitigating this issue by automatically assigning each AP to the best channel and transmit power level. Using Auto-RF is recommended in most cases, but keep in mind that it can only mitigate the issue, not solve it completely, if the building material does not isolate each AP from the radio footprints of the other APs.

www.CareerCert.info




5-GHz Network DesignUp to 23 “non-overlapping” channelsMinimum recommended data rate is 24 Mb/sRecommended cell edge RSSI –67 dBm802.11a is immune to ISM interferences and preferred for voice

The 5-GHz unlicensed band covers up to 300 MHz of spectrum and supports up to 23 channels. The 5 GHz band is actually a conglomeration of several bands: 5.150-to-5.250 GHz (UNII-1), 5.250-to-5.350 GHz (UNII 2), 5.47-to-5.725 GHz (UNII-2 extended) and 5.725-to-5.875 GHz (UNII-3). In each band, channels are 20 MHz apart. 802.11a uses orthogonal frequency-division multiplexing (OFDM), for which a channel is 20 MHz wide. This means that two consecutive channels on 802.11a overlap only slightly at their edges. They can be used for data networks, but you should use one channel of separation for better efficiency in voice deployments.

There are different limitations imposed on each of the Unlicensed National Information Infrastructure (UNII) bands. Depending on the band, restrictions include transmit power, antenna gain, antenna styles, and usage. The UNII-1 band is designated for indoor operation, and it initially required that devices use permanently-attached antennas. The UNII-2 band was designated for indoor or outdoor operation and permitted the use of external antennas. The UNII-3 band, originally intended for outdoor bridge products that used external antennas, can now be used for indoor or outdoor IEEE 802.11a WLANs as well. The channels in UNII-1 (5.150 to 5.250 GHz) are 36, 40, 44, and 48. The channels in UNII-2 (5.250-5.350 GHz) are 52, 56, 60, and 64, and they require Dynamic Frequency Selection (DFS) and Transmit Power Control (TPC). The channels in the new frequency range (5.470-5.725 GHz) are 100, 104, 108, 112, 116, 120, 124, 128, 132, 136, and 140, and they require DFS and TPC. The channels in UNII-3 are 149, 153, 157, and 161 (5.725-5.825), and they do not require DFS and TPC. Not all channels in a given range can be used in all of the regulatory domains. An extra channel, 165, is part of the industrial, scientific, and medical (ISM) band, but is not part of the UNII band and cannot be used for 802.11 deployments.

www.CareerCert.info




5-GHz Network Design (Cont.)

Data Rate (Mb/s)802.11a (20mW 1242 AP with 3.5 dBi gain diversity dipole antenna) Range

802.11g (20mW 1242 AP with 2.2 dBi gain diversity dipole antenna) Range

54 85 ft (26 m) 105 ft (32 m)48 150 ft (46 m) 180 ft (55 m)36 210 ft (64 m) 260 ft (79 m)24 230 ft (70 m) 285 ft (87 m)18 260 ft (79 m) 330 ft (100 m)12 280 ft (85 m) 355 ft (108 m)11 - 365 ft (111 m)9 310 ft (94 m) 380 ft (116 m)6 330 ft (100 m) 410 ft (125 m)

5.5 - 425 ft (130 m)2 - 445 ft (136 m)1 - 460 ft (140 m)

Operating in the unlicensed portion of the 5 GHz radio band, IEEE 802.11a is immune to interference from devices that operate in the 2.4 GHz band, such as microwave ovens, many cordless phones, and Bluetooth.

The 802.11a standard provides data rates of 6, 9, 12, 18, 24, 36, and 48, with a maximum data rate of 54 Mb/s, like 802.11g, though generally at shorter ranges compared to 2.4GHz network for a given power and gain. However, the increased number of non-overlapping frequency channels, up to 23, compared to the 3 non-overlapping channels for the 2.4GHz band, results in increased network capacity, improved scalability, and the ability to create microcellular deployments without interference from adjacent cells. This difference is sufficient for the 5 GHz band to be recommended for VoWLAN deployments.

As the 5 GHz cells are smaller than the 2.4 GHz cells, determining the position of dual band 802.11b/g and 802.11a APs might be challenging. As a general principle, do a site survey based on the most challenging client requirements. If your most challenging clients are voice devices using the 802.11a channel, do a site survey on this basis. Be aware that the 80.211b/g deployment will therefore most result in an increased interference level for your other clients. Reducing the power level of the APs on the 2.4GHz spectrum might help mitigate this problem.

The above table is a typical example of ranges and speed values for a common AP. Conditions in an actual deployment may be different depending on the RF and physical environments.

www.CareerCert.info




5-GHz Network Design (Cont.)UNII-2 and UNII-2 extended should be avoided, when possible, to mitigate 802.11h-related disruptions.

Although there are 23 non-overlapping channels in the 5-GHz band, it is generally recommended that you use the lower 4 channels and upper 4 channels of the 5 GHz spectrum as the base for VoWLAN, as they do not have DFS and TPC requirements. Then add to the base of eight channels by determining which other channels are unlikely to be affected by DFS and TPC. The timing requirements of DFS and TPC can adversely affect VoWLAN call quality.

If your region or location is such that you are certain DFS and TPC will not be triggered, then the use of specific channels should not be an issue. If you are not certain, you should investigate. The Cisco Spectrum Expert analyzer is a good tool for starting this assessment to determine whether there are any 5-GHz signals in the area that would trigger DFS and TPC. Monitoring the AP from the controller SNMP logs is another way to check if DFS and TPC are triggered.

DFS and TPC aim at avoiding interferences with airport radars. When a radar blast is detected, the AP must stop client data transmission within 260 milliseconds (Channel Closing Transmission Time). It then has 10 seconds to change its channel (Channel Move Time). During this interval, the AP continues to send beacons and can inform its clients of the next channel it is moving to (Channel Change Announce message). Once on the new channel, the AP must remain silent for 60 seconds to check if radar blasts are heard (Channel Availability Check Time). If blasts are heard, the AP moves to another channel. If no blast is heard, the AP can then resume its normal operations, and it must stay away from the channel it left for 30 minutes (non-occupancy period). In a worst-case scenario, more than 70 seconds (260 ms + 10 s + 60 s) can pass between the moment the AP receives a radar blast and stops sending traffic on a given channel and the moment the AP resumes its traffic on a new channel. No voice conversation can be maintained during such a long interruption. It is simpler to stay with the eight non-DFS channels, but every additional channel that can be safely deployed increases the capacity of the design. In addition to avoiding the DFS and TPC channels, it is also recommended that adjacent channels be avoided in the AP channel layout—to avoid interference from the sidebands in each channel.

www.CareerCert.info




5-GHz Network Design (Cont.)In open space, even though the AP footprint exceeds the cell edge, 8 non-overlapping / non-802.11h channels render the –86 dBmseparation objective easy to achieve.In worst-case scenarios, 8 channels are available for the whole floor; in best-case scenarios, up to 23 channels are available.

The general power levels and AP separation recommendations for VoWLAN in the 5 GHz implementation are the same as for the 2.4 GHz implementation: a power level boundary of -67 dBm and a separation between adjacent AP channels of -86 dBm. Given the lower noise floor in the 5-GHz bands, the overlap recommendation may be reduced to 15 percent. A 20 percent or higher overlap can still be used if desired. This amount of overlap provides a higher-availability design and takes into account that the use of the 5GHz spectrum is increasing; therefore, the noise floor can be expected to rise.

With 15 percent cell overlap, the distance x between APs is set as:

x = 1.486 R where R is the cell radius.

Although the AP RF footprint is larger than the cell size, the non-overlapping possibility means that each AP will interfere with few, or no, surrounding APs. It is very difficult to calculate if deploying 20 APs on 8 non-overlapping channels will create a capacity closer to 8 or to 20 APs, as this figure will depend on the building layout. It is still safe to assume at least eight times the capacity of one AP, which is already more than two times more efficient than the capacity of a 2.4 GHz deployment. In any case, a good practice is to avoid positioning neighboring APs on neighboring channels. Although 802.11a channels are said to be non-overlapping, a small overlap still exists. If you position one AP on channel 36 and its neighbor on channel 40, the performance of both channels will be degraded slightly. Cisco recommends leaving one channel of separation between neighboring APs (using 36 and 44 in this example), and using physical separation for APs on adjacent channels.

www.CareerCert.info



Combinations of WLAN Services This topic gives design recommendations for voice and data combinations in a wireless network.


Voice and Data ServiceSeparate voice and data– Separate SSID and

VLAN– Best to separate RF

Data characteristics– Bursty, with large

frames– Interrupts voice

Voice characteristics– Small frames at

constant rate– Requires QoS– Requires CAC to protect

voice from itself

Voice and data traffic in a wireless cell should be separated. These two types of traffic do not have the same behavior:

Data traffic is bursty, voice is smooth: When a user sends data over a web connection, for example, the user will generally send a large amount of data at once, based on TCP, then stop to process received information, then send another burst. Over the same course of time, a VoWLAN user will send a steady 50 frames per second, no more, no less. Data traffic bursts can delay the voice frames.

Data frames are large, voice frames are small: A typical voice frame is 240 B long in the 802.11 space, but the default maximum transmission unit (MTU) is 2346 B. Data frames commonly use the maximum MTU, and therefore take about 10 times longer to be transmitted. If a wireless IP phone needs to wait for several data devices to send before being able to send its own frame, it may exceed the 30 ms timeout.

Data traffic can be resent, voice traffic cannot: If a data packet is missing, it can be resent within a rather long interval (it takes more than 1 minute for a TCP window to slide down from 254 to 0 and timeout). When a voice packet is missing, it can be resent only within a short timeframe, after which it has to be dropped.

Data traffic and voice traffic QoS requirements are different: Voice traffic needs to be prioritized, in the wireless space as well as the wired space. Data traffic does not necessarily have the same need.

www.CareerCert.info



For these reasons, voice and data traffic should be separated. This can be done logically, by creating two different Service Set Identifiers (SSIDs), having two different QoS mappings. In this case, the isolation is artificial, as voice and data traffic still share (and compete for) the same wireless space. It is recommended that you use physical RF separation instead, where, for example, voice will be set on the 802.11a spectrum whereas data will stay on the 802.11b/g band.

On the wired side, voice and data should also be separated. They should be sent to different VLANs and have different QoS and security configuration parameters.

www.CareerCert.info




Voice and Mobility Services

The signal level requirements of 802.11 location-based services are similar to those on VoWLAN, but the AP placement requirements are different. The example here illustrates APs positioned close to what would be required in a Location-Based Service (LBS) deployment. In this environment, there are many APs deployed on the perimeter, as well as at the core of the building. An additional column of APs (four) might be required. The VoWLAN AP count and the Location AP count may not always be this similar. The AP placement requirements of LBS may result in the addition of more APs.

The addition of more APs for LBS will introduce an additional level of co-channel interference due to the additional 802.11 management traffic associated with additional AP; however, given the existing co-channel interference, the difference is not likely to be significant. The key point, as with the VoWLAN deployment, is that the addition of extra APs does not contribute to additional capacity in the 2.4GHz band due to co-channel interference. When a mixed environment is to be designed for both voice deployment and location services, it is possible to set some of the APs, particularly the ones at the perimeter of the building, to a specific mode, called Location Optimized Monitor Mode. In this mode, an AP does not accept any client and does not interfere with the other APs. It passively gathers location-related information to increase the location accuracy.

Another point that has to be taken into consideration is the impact of the RFID tags on the voice traffic. This might be important when voice is deployed on the 2.4-GHz band. Wireless IP phones will be impacted by traffic from other wireless clients, such as laptops. The RFID present some specificity at that respect.

To understand their impact, keep in mind that the purpose of the RFID tag is to provide the most accurate possible location information. It therefore tends to send its signal at a very low speed, in order for the signal range to be maximized. The more APs receive the signal, the more accurate the location will be. For the same reason, the RFID power level can be close to the maximum allowed for this spectrum.

www.CareerCert.info



This means that most RFID tags emit at a speed of 1 or 2 Mb/s. Depending on the RFID type, this behavior can have an impact on the cell. Some RFID tags emit a pure Layer 2 multicast message, which is detected and recognized as such by the APs. The APs relay the information to the controller for location purposes and do not take any further action. The send message being rather short, this type of RFID tag has little impact on the wireless network.

Some other RFID tags act as clients, and can associate to an SSID. They have a high impact on the wireless network. As soon as one sends a first message, at 1 Mb/s, which is an 802.11b speed, the AP triggers the protection mechanism by which 802.11g clients protect themselves from collisions. This immediately degrades performance in the cell, from 22-23 Mb/s in the 54 Mb/s area, to 8 Mb/s throughput. This directly impacts the number of possible concurrent calls in the same cell.

When you deploy this type of mixed environment, always make sure to use RFID tags that use a Layer 2 multicast mechanism. Also keep in mind that, although each RFID tag message is short, the interval at which the tag emits, and the number of tags present in the same area, should be taken into consideration. If a box of 100 tags is present in the cell, and if each tag is configured to emit every 5 seconds, you will certainly see a heavy degradation of the voice over wireless quality in this area. This degradation may be seen in the neighboring cells as well, as most RFID tags do not send on one channel only, but are configured to emit on channels 1,6, and 11.

www.CareerCert.info




Wireless CACCurrently only outgoing call admission control

Based on channel utilization from the access pointWill lock out after channel utilization exceeds the maximum threshold

No incoming call admission control

Incoming calls may force channel utilization to exceed the maximum threshold, which could potentially impact audio quality

The administrator also needs to make sure that the wireless phone initiates a call only if the wireless cell has enough resources for the call to occur in good conditions: to avoid having too many phones compete for the same resources, wireless call admission control (CAC) can be configured at the controller level. With wireless CAC, a phone is allowed to place a call only if there is enough bandwidth available in the cell. If too many phones use the same wireless resources, all calls suffer from quality degradation. Wireless CAC ensures that once the capacity limits are reached, no extra call will degrade the quality of the already ongoing conversations.

Wireless CAC is based on the AP utilization, and takes into account parameters such as the noise level to determine what space is left for outgoing calls. If a wireless phone needs to place a call in a situation in which there is not enough space, the phone receives a Network Busy message. This process is different from the call admission control specific to voice-control devices, which may allow or forbid a call based on other parameters.

Wireless CAC is only possible for outgoing calls, that is, calls initiated from the cell. When an incoming call is received from the wired side, it is not possible for the wireless infrastructure to send a “busy message” to the original voice traffic sender, which may be any voice device down to an analog phone on the other side of the globe. Therefore, for incoming calls, the wireless infrastructure applies the usual “best effort” policy, and forwards the traffic to the cell. This may result in one or several additional wireless IP phones starting voice streams in the cell. This is why a good design adds some margin to account for this scenario.

The same wireless CAC applies when the user moves and roams from one AP to the other. A configurable margin is added to allow extra users to enter a cell that is already fully utilized and continue their conversations without degrading the voice flow of other users. When a user enters a cell that does not have any room left on the “roaming margin,” the call is dropped.

www.CareerCert.info



Keep in mind that Wireless CAC only aims at protecting the wireless cell against excessive call numbers. Another form of CAC exists in Cisco Unified Communications Manager, by which calls can be admitted or refused based on voice-related parameters, such as dialed number, bandwidth on the WAN, and user ID. You can say that to succeed, a call initiated from the wireless space needs to pass to CAC tests: the wireless CAC, measuring 802.11-related information to determine if there is enough space to allow the outgoing flow, and the voice CAC, measuring voice-related values on the wired side.


VoWLAN AnalysisWireless analysis tools can usually evaluate wireless issues, but cannot determine call quality (MOS)Voice tools can usually determine call quality, but cannot always determine the cause a low quality

The design of wireless coverage for voice is one aspect contributing to call quality. Analyzing the effective call quality is usually required to ensure that the network meets user expectations. There are many tools to evaluate wireless network performances, but these tools are not always the same as the ones used to evaluate the calls quality.

Wireless-related tools can provide information about the noise level of the cell, the RSSI level at a given point, the average delay for a frame to get to the AP, and the comeback and loss rates.

Voice-related tools provide a more global view, in the sense that they focus on the whole network. At the same time, they also provide a narrower view, in the sense that they only concentrate on one factor: what the MOS5 level of a voice call is. These tools can usually evaluate if call quality is low, but cannot always determine why.

Wireless tools can evaluate the quality of the wireless network and its adaptation to voice, but cannot mark the voice quality level. Both types of tools are usually used in combination for VoWLAN analysis.

5 The MOS score relies on users marking the call quality. The term is used here as it was defined earlier. Most software tools evaluating voice call quality use automated evaluation systems, providing the same types of result but using a scoring system bearing a different name: for example, the Perceptual Evaluation of Speech Quality (PESQ), which compares the voice quality to predefined samples stored in a database, each sample having a predefined MOS value.

www.CareerCert.info



Note These tools help in the troubleshooting and analysis phases but are nothing without the troubleshooter’s intelligence. The case is well known of a large shipbuilding industrial environment in which voice over wireless was problematic: communication was poor and choppy, sound difficult to hear. Cells were all designed for 24 Mb/s minimum, and there were no more than 10 to 12 active calls at a time in each cell. The loss and retry level to the AP was very low, and the wired network showed no congestion. After weeks of wireless troubleshooting, it appeared that metallic dust coming from the melding phase was attracted by the phone microphone magnet and was creating interferences in the microphone itself… a simple piece of tape on the microphone holes solved the “VoWLAN” problem.

www.CareerCert.info



Voice over WLAN Security This topic describes the WLAN Security for Voice deployments.


Wireless SecurityAny wireless device possessing sniffing capabilities can easily eavesdrop on voice trafficVoice network can be a point of entry to the corporate wireless network if not protected

WLAN traffic is visible to any WLAN device within RF range, and is a shared access medium. This creates a number of security challenges:

How do you provide privacy for users of your WLAN, from non-users?

How do you provide privacy for users of your WLAN from each other?

How do you support privacy of multicast and broadcast traffic?

How do you identify which user is which on the WLAN?

A voice wireless client is, in this respect, just like any other wireless client. Its security is a concern, because any laptop equipped with free codec-enabled software can capture and replay a voice conversation.

www.CareerCert.info




Wireless Security (Cont.)

Each generation of WLAN security has addressed these challenges in a slightly different way. But the key mechanisms are based on the same strategies used to secure communication over an untrusted medium (in other words, Authentication, Authorization and Accounting (AAA) and encryption). The original 802.11 standard defined an encryption mechanism, Wired Equivalent Privacy (WEP), but did not define any AAA mechanism. The level of authentication offered in the 802.11 standard was at a group level, everyone in the group had to have the same encryption key. This key was used to encrypt unicast and multicast traffic.

The weaknesses in WEP and the demand for a solution drove the Wi-Fi Alliance to develop WLAN security improvements, based on an 802.11i draft. These improvements are defined as Wi-Fi Protected Access (WPA). WPA addressed the main weakness in WEP encryption by replacing it with the Temporal Key Integrity Protocol (TKIP), which reuses the core encryption engine of WEP (RC4). The reuse of RC4 allowed TKIP to be implemented in the majority of systems through a firmware upgrade, rather than requiring a hardware upgrade. In addition to TKIP, WPA implemented one other major improvement to WEP encryption, an additional message integrity check (MIC) mechanism. In addition to the encryption and message integrity improvements, WPA introduced cryptographic improvements in which the key shared between the WLAN client and the WLAN AP is not used directly for encryption, but instead it is used as the basis to derive the encryption key, which is rotated. The two versions of WPA are WPA-Personal and WPA-Enterprise. With WPA-Personal, a pre-shared key configured on the client and the wireless infrastructure is used as the basis to derive the other keys. With WPA-Enterprise, user authentication relies on a RADIUS server (AAA), relying on EAP mechanisms, such as Extensible Authentication Protocol-Transport Layer Security (EAP-TLS), Extensible Authentication Protocol-Flexible Authentication via Secure Tunneling (EAP-FAST), Protected Extensible Authentication Protocol (PEAP), or Lightweight Extensible Authentication Protocol (LEAP). The main key is derived for each client at the end of the authentication process.

www.CareerCert.info



While the security changes from 802.11i adopted by WPA are important, the key component in 802.11i was the incorporation of the Advanced Encryption Standard (AES) into WLAN security. This aligns its encryption mechanism with the new industry standard for encryption. The underlying mechanism of AES-Counter with CBC-MAC (AES-Counter describes the encryption mechanism, and CBC-MAC describes the frame protection mechanism) is very different from those of WPA and WEP. It generally requires a hardware upgrade to be supported. The hardware requirements to support AES encryption in WPA2 mean that migration from WPA is dependent upon hardware refreshes. In many cases, updating the network infrastructure is an easier task than updating the WLAN client infrastructure and a complete migration to WPA2 is dependent upon a generational change in the WLAN client infrastructure.

Wireless phones act and are authenticated exactly like other wireless devices. Depending on the model, a phone can support one or several authentication mechanisms. A good quality phone should provide the same quality for access to the network using an open SSID and for access using an encryption mechanism. Performance should be the same and no voice traffic degradation should occur.

The only time when a difference may be noticeable is during the authentication process, when using a strong mechanism such as EAP-FAST. The wireless infrastructure is usually built for wireless devices such as laptops, which process the EAP exchanges fast. The wireless phone CPU, being lighter, may be slower to process the exchange, and the EAP timeout sometimes needs to be extended on the wireless controller or APs to allow for wireless IP phones.

www.CareerCert.info



Voice over WLAN Roaming This topic describes how to manage roaming issues for voice over wireless deployments.


Roaming Decision

A wireless client typically decides to roam when its connection to the current AP degrades. Roaming necessarily has some impact on client traffic, because a client scans other channels for alternative APs, reassociates, and authenticates to the new AP. Although the roaming algorithms differ for each vendor or driver version (and potentially for different device types from a single vendor), there are some common situations that typically cause a roam to occur:

Maximum data retry count is exceeded: Excessive numbers of data retries are a common roam trigger. Each vendor has a different threshold, which is linked to the data rate. A given threshold triggers to shift to a lower data rate, another threshold to roam.

Low received signal strength indicator (RSSI): A client device can decide to roam when the received signal strength drops below a threshold. This roam trigger does not require active client traffic in order to induce a roam.

Low signal to noise ratio (SNR): A client device can decide to roam when the difference between the receive signal strength and the noise floor drops below a threshold. This roam trigger does not require active client traffic in order to induce a roam.

Proprietary load balancing schemes: Some wireless implementations have schemes in which clients roam in order to more evenly balance client traffic across multiple APs. This is one case in which a roam may be triggered by a decision in the WLAN infrastructure and communicated to the client via vendor-specific protocols. Cisco WLAN controllers are configured with a default set of RF roaming parameters that are used to set the RF thresholds adopted by the client to decide when to roam. The default parameters can be overridden by defining a custom set. These Cisco Compatible Extensions parameters are defined on the controller for each frequency band.

www.CareerCert.info




Scanning for a New Access Point

Wireless clients learn about available APs by scanning other 802.11 channels for available APs on the same WLAN or SSID. Scanning other 802.11 channels can be performed actively or passively as follows:

Active scan: Active scanning occurs when a client changes its 802.11 radio to the channel being scanned, broadcasts a probe request, and then waits to hear any probe responses (or periodic beacons) from APs on that channel (with a matching SSID). The 802.11 standards do not specify how long the client should wait, but 10 ms is a representative period. The probe-request frames used in an active scan are one of two types:

— Directed probe: The client sends a probe request with a specific destination SSID; only APs with a matching SSID will reply with a probe response.

— Broadcast probe: The client sends a broadcast SSID (actually a null SSID) in the probe request; all APs receiving the probe-request will respond, with a probe-response for each SSID they support.

Passive scan: Passive scanning is performed by simply changing the 802.11 radio of the client to the channel being scanned and waiting for a periodic beacon from any APs on that channel. By default, APs send beacons every 100 ms. Because it may take 100 ms to hear a periodic beacon broadcast, most clients prefer an active scan.

During a channel scan, the client is unable to transmit or receive client data traffic. Clients take a number of approaches to minimize this impact to their data traffic:

Background scanning: Clients may scan available channels before they need to roam. This allows them to build up knowledge of the RF environment and available APs so they can roam faster if it becomes necessary. Impact on client traffic can be minimized by only scanning when the client is not actively transmitting data, or by periodically scanning only a single alternate channel at a time (scanning a single channel incurs minimal data loss).

www.CareerCert.info



On-roam scanning: In contrast with background scanning, on-roam scanning occurs after it has been determined that a roam is necessary. Each vendor, and each device may implement its own algorithms to minimize the roam latency and the impact on data traffic. For example, some clients might only scan the non-overlapping channels.

There are some informational attributes that may be used to dynamically alter the roam algorithm:

Client data type: For example, voice call in progress

Background scan information: Obtained during routine periodic background scans

Ways in which attributes can be used to alter the scan algorithm include the following:

Scan a subset of channels: For example, information from the background scan can be used to determine which channels are being used by APs in the vicinity.

Terminate the scan early: For example, if a voice call is in progress, the first acceptable AP might be used instead of waiting to discover all APs on all channels.

Change scan timers: For example, if a voice call is in progress, the time spent waiting for probe responses might be shortened during an active scan.

www.CareerCert.info




ReauthenticationOpen and PSK reauthentication are fast enough not to be noticedRADIUS-based reauthentication can take one second

When a client wants to join a new AP, it needs to reauthenticate. The latency depends on the authentication mechanism:

Open authentication: the authentication and association process is a simple 4-packet exchange; it is fast and usually not detected by the user.

Pre-shared key authentication: this process adds a challenge phase, which is fast enough to be seamless for a voice conversation.

AP authentication: when a client roams using 802.1X with Dynamic WEP, WPA-Enterprise or WPA2-Enterprise, an EAP authentication generally must occur with the RADIUS server. Authenticating with a RADIUS server can take more than one second. A one-second interruption to latency-sensitive applications such as VoIP when roaming is unacceptable. To make this type of roaming possible, a specific solution must be in place. Cisco has developed fast secure roaming algorithms to reduce the roam latency.

The roaming process described here deals with the Layer 2 issues. At Layer 3, the client has an IP address, on which the Real-Time Transport Protocol (RTP) and Real-Time Transport Control Protocol (RTCP) flows are based. If both APs are in the same subnet, the client maintains its IP address and the RTP and RTCP sockets are still valid. If both APs are on different subnets, Layer 3 issues are added to existing Layer 2 roaming issues as the client changes its IP address, causing the IP communication to reset. Here again, Cisco has developed specific solutions to prevent this issue.

www.CareerCert.info



Cisco Compatible Extensions for VoWLANs This topic describes the main Cisco Compatible Extensions features for voice.


The Cisco Compatible Extension Program

No-cost licensing of technology for use in WLAN adapters and devicesIndependent testing to ensure interoperability with Cisco infrastructure latest innovationsMarketing of compliant products by Cisco and product suppliers under “Cisco Compatible” brand

Cisco Compatible Extensions Version EAP Types Supported

CCX v2 CCKM with LEAP

CCX v3 CCKM with LEAP, EAP-FAST

CCX v4 CCKM with PEAP, EAP-FAST, EAP-TLS and LEAP

When a client roams using open authentication (no keys) or using shared keys, authentication adds little roam latency. This is because no additional packets need to be exchanged between the client and the AAA server. When a client roams using IEEE 802.1X with Dynamic WEP WPA-Enterprise or WPA2-Enterprise, an IEEE 802.1X authentication generally must occur with a AAA/RADIUS server. Authenticating with a AAA/RADIUS server can take more than one second. A one-second interruption to latency-sensitive applications such as VoIP when roaming is unacceptable and therefore fast secure roaming algorithms have been developed to reduce the roam latency.

Fast roaming algorithms include Cisco Centralized Key Management (Cisco CKM is shown as “CCKM” on the controller configuration pages) and Proactive Key Caching (PKC). Cisco CKM and PKC allow a WLAN client to roam to a new AP and re-establish a new session key—known as the Pairwise Transient Key (PTK)—between the client and AP without requiring a full EAP authentication to a RADIUS server.

Both CISCO CKM and PKC are Layer 2 roaming algorithms in that they do not consider any Layer 3 issues such as IP address changes. In the Cisco Unified Wireless Network, clients are allocated IP addresses from subnets that originate at the WLC—not the AP. In this way, it is possible to group large numbers of WLAN clients for a given SSID into the same Layer 2 subnet. This maximizes the scope of the Layer 2 domain and the fast secure roaming domain. Additionally, multiple-WLC deployments support client roaming across APs managed by WLCs in the same mobility group on the same or different subnets. This roaming is transparent to the client, because the session is sustained and a tunnel between the WLCs allows the client to continue using the same DHCP-assigned or client-assigned IP address as long as the session remains active.

www.CareerCert.info



CISCO CKM is a Cisco standard supported by Cisco Compatible Extensions clients to provide fast secure roaming. CISCO CKM establishes a key hierarchy upon initial WLAN client authentication and uses that hierarchy to quickly establish a new key when the client roams.

CISCO CKM requires support in the client. Cisco Compatible Extensions provides client-side specifications for support of many client functions, including fast secure roaming. The table above summarizes the EAP types supported in each version of Cisco Compatible Extensions.

www.CareerCert.info




Cisco Compatible Extensions Scanning

While WLAN clients ultimately determine when to associate (or reassociate) to an AP, Cisco APs provide information to clients to facilitate AP selection by providing information (such as the channel load in the beacons and probe responses for each AP) or by providing a list of neighboring APs.

This Cisco Compatible Extensions information is divided in four different features:

AP assisted roaming: This feature helps clients save scanning time. Whenever a Cisco Compatible Extensions client (version 2 or later) associates with an AP, it sends an information packet to the new AP that lists the characteristics of its previous AP. The AP uses this information to build a list of previous APs, which it sends (via unicast) to clients immediately after association to reduce roaming time. The AP list contains the channels, Basic Service Set Identifiers (BSSIDs) of neighbor APs that support the client’s current SSID or SSIDs, and the time elapsed since disassociation.

Enhanced neighbor list: The enhanced neighbor list is an enhanced version of the “neighbor list” which is sent as part of the AP Assisted Roaming feature. It is always provided unsolicited by the AP to the client immediately following a successful association or reassociation. As the AP periodically checks to ensure its neighbor list is up to date, it may also send an unsolicited update to the corresponding clients. The enhanced neighbor list may include, for each AP in the list, information about AP timing parameters, the AP support for the client’s subnet, and the strength and SNR of the last transmission the AP received from the client.

Enhanced neighbor list request (E2E): the End-2-End (E2E) specification is a Cisco and Intel joint program that defines new protocols and interfaces to improve the overall voice and roaming experience. It applies only to Intel clients in a Cisco Compatible Extensions environment. Specifically, it enables Intel clients to request a neighbor list at will. When this occurs, the AP forwards the request to the controller. The controller receives the request and replies with the current Cisco Compatible Extensions roaming sublist of neighbors for the AP to which the client is associated.

www.CareerCert.info



Directed roam request: This feature enables the controller to send directed roam requests to a client in situations when the controller can better service the client on a different AP from the one to which the client is associated. In this case, the controller sends the client a list of the best APs that it can join. The client can either honor or ignore the directed roam request. Clients that are not Cisco Compatible Extensions clients and clients running Cisco Compatible Extensions version 3 or earlier do not take any action.

www.CareerCert.info




Cisco Compatible Extensions Scanning (Cont.)

Ping-pong effect occurs when a wireless client is at the edge of two cells and hops between them. Cisco Compatible Extensions allows an AP to determine “inner regions,” where scanning is not required, and “outer regions,” where a client moving away from the AP is instructed to scan for a better AP.

Another way that the Cisco Compatible Extensions feature adds value is by saving battery life and avoiding what is know as the “ping-pong” effect. When a client is at the edge between two APs, there is an area in which the signal is received from both APs. Even if the wireless network is properly designed, the client may decide to hop from the first AP to the second AP.

But as the RF environment is changing all the time, soon the first AP signal becomes better than the second one, and the client jumps back to it, and so on. The result is an unstable network connection. The client may realize very late that the signal from the first AP is weak, and start scanning for another AP when it is already losing the first AP signal.

Cisco Compatible Extensions includes many features to improve roaming and solve this kind of problem.

First, the client will not blindly follow the rule that would normally make it jump to the best AP, but will respond to messages from all the APs.

Second, the APs will detect each other and determine regions:

An inner region, comprising wherever the RSSI for the packets of the serving AP exceeds a certain threshold. Within this region, roams should not be required, so a client may suspend or markedly slow its scanning in order to reduce power consumption and minimize the time spent off-channel. The AP will send this information to the client (it will inform the client that the computer is in the inner region, therefore frequent scans are not necessary because there is good coverage).

An outer region, called the Transition Region, comprising wherever the RSSI for the packets of the serving AP packets falls below a certain threshold. Within the Transition Region, a roam may be required. The AP will inform the client to start scanning, as the cell edge is approaching.

With this feature, the client will always know when it gets to the edge of the cell. This solves the late scanning issue.

www.CareerCert.info



If the client stays at the edge, it will give a premium to an AP (it will prefer it) in order to avoid the ping-pong effect. If the client moves closer to an AP, the wireless infrastructure can detect that the second AP would be a better connection point than the first one, and send an instruction to the client for it to move.

With this feature, you can ensure that the clients are always connected in a stable manner to the best AP for them. This results in optimal speed for all the clients, and can also save up to 30 percent of battery lifetime, which is, of course, a very important factor for wireless IP phones.

www.CareerCert.info




Cisco Compatible Extensions: Roaming and Cisco CKM

Cisco Centralized Key Management allows the wireless infrastructure to cache the credentials, thus eliminating the need to conduct a full reauthentication when roaming.

When using central server-based authentication (802.1X), some issues occur when clients roam. When a wireless client leaves one AP to associate with a second AP, the normal process would be that the client should reauthenticate: if the client is known by the first AP to which it is associated, it is not known by the second AP to which it tries to jump. Therefore, all the process of validating the client identity should be restarted from the beginning, which implies that:

The client will be disconnected for a few hundred milliseconds because of the time it takes to reauthenticate with the central server and reassociate

The client might have to request a new IP address if the network infrastructure does not allow IP addresses to be re-used (to avoid duplicate IP address issues)

This means that the client connection is dropped in most cases. In modern networks, the network infrastructure can enhance roaming, because APs and controllers can communicate and anticipate client requests. The CISCO CKM feature allows the client and the wireless infrastructure to communicate on this point, removing the need for the client to be reauthenticated by the central server: the APs of the controllers will cache and transmit the credentials in a secure manner between each other following client movements. CISCO CKM requires Cisco Compatible Extensions version 2 or later6.

6 If your device is certified for Cisco Compatible Extensions and supports an authentication protocol, the Cisco CKM support for this protocol is included in your client Cisco Compatible Extensions version.

www.CareerCert.info




CCX: AP Specified Max Power

In a heterogeneous environment, a power difference between clients and an AP can prevent clients from being heard by the AP. With Cisco Compatible Extensions, the AP can detect and automatically configure the client optimal transmit power.

In heterogeneous environments, wireless networks often need to support different types of clients with different needs and also address challenges to determine the right density of APs to support those diverse clients. In a typical situation, the AP density must be higher (more APs used to cover the same area) to allow wireless phone coverage. In an optimal design, clients will have the correct transmit power to be in range of one AP and, when they roam, they will pick up a good signal from a second AP exactly when they start to lose RF signal coverage from the first AP.

This is an ideal scenario, but as clients are heterogeneous and networks are often a mix of wireless IP phones and wireless data devices, some clients will have a stronger transmit power, and will be heard clearly by the APs, where others will have their transmit power set to a lower value, resulting in communication in a single direction. The latter type of clients will hear the AP, but the AP will not hear them.

There are also some scenarios where the power level on the client is too high, and it is heard by quite a few APs. The result may be a flaky connection, as the client hops from AP to AP, or a slow network, as the client stays connected to a distant AP during roaming when it should have moved to a closer AP.

Setting the transmit power of each client device one-by-one to match the specifics of each network environment would be an impossible task, and no command permits this feature. However, with the AP Specified Max Power Cisco Compatible Extensions feature (version 2 and later), the AP will set the transmit power of each client automatically.

With the ability to identify the number of associated clients, cell sizes, and adjacent AP radio signals, the APs can determine the optimum transmit power required for the clients. The ability to dynamically set client output power during the association process will increase the overall performance of the wireless network and improve WLAN device battery life.

www.CareerCert.info




802.11rApproved in July 2008Allows for fast basic service set (BSS) transitions802.11r incorporates features of:– 802.11i, by which clients can preauthenticate to several

candidate APs– 802.11e, by which clients can inform the infrastructure about

their QoS need and reserve bandwidthCisco Unified Wireless Network will include 802.11r as clients cards start supporting itCisco CKM clients still benefit from Cisco CKM advantages

In July 2008, the IEEE committee published a new protocol for fast roaming and fast basic service set (BSS) transitions: 802.11r. The idea behind the 802.11r protocol is to specify a mechanism by which wireless devices can roam while still being continuously connected, thus enhancing the default 802.11 behavior explained previously. The target is to bring the roaming time under the 50 ms delay, which represents less than one lost packet in a normal environment. With such a low roaming time and one single lost packet, the roaming process should not be detected by human users. This delay should be achieved even with 802.1X authentication, which often requires a new dialog with the AAA server, which can, in turn, add to the roaming delay.

This is achieved by allowing a client to preauthenticate to candidate APs and reserve QoS resources before actually roaming. The 802.11r protocol is a way to obtain the features of two protocols:

802.11i, by which preauthentication is possible over several possible candidate APs, even in 802.1X environments

802.11e, by which QoS information elements can be sent to the AP for bandwidth reservation.

Cisco was part of the 802.11r development effort. As client cards start supporting the 802.11r protocol, the Cisco Unified Wireless Network infrastructure is integrating 802.11r support for roaming clients other than Cisco CKM clients.

A difference between Cisco CKM clients and 802.11r clients is that 802.11r clients authenticate via several APs, thus dialoging several time with the AAA server over the wireless and wired infrastructure. With Cisco CKM, clients authenticate once, and the wireless infrastructure takes care of credential transmissions.

www.CareerCert.info



WLAN Controller Configuration and Design This topic describes the controller positioning and configuration requirements for roaming efficiency.


Roaming and Mobility Groups

The mobility group is a global parameter configured on a controller. It defines membership to a group. Several controllers are usually associated with the same mobility group and know each other. They use this membership to exchange information about roaming clients. A controller can know controllers that belong to another mobility group. The two controllers are said to be in the same mobility domain. Roaming will not occur if the controllers are located in different mobility groups and do not know each other, but will occur if the controllers are in different mobility groups and in the same mobility domain.

For two controllers to be in the same mobility domain, they must know each other, which means that the built-in MAC address and management IP address of each controller has been entered in the other controller. When a client moves from one controller to another in a different mobility domain (that is, to a controller not known by the controller the client leaves), the client will have to reauthenticate, reassociate, and get new IP address information. The way mobility groups are organized in your wireless network determines if and how roaming will occur. Make sure that all controllers between which roaming should happen are in the same mobility group. Controllers in a mobility group must all have these characteristics in common:

Mobility group name

Version of controller code

Control and Provisioning of Wireless Access Points (CAPWAP) mode

ACLs

WLANs (SSIDs)

www.CareerCert.info



These requirements must be met for client roaming to function properly. Without them, a client may have to reassociate or reauthenticate.

www.CareerCert.info




Client Roaming within a Subnetwork

The Cisco Unified Wireless Network environment allows for roaming between APs associated with the same Cisco WLAN controller, as well as roaming between APs associated with different controllers. Roaming can also occur either as Layer 2 roaming, meaning that the client subnet does not change, or Layer 3 roaming, meaning the client moves from an SSID on one AP associated with one VLAN and respective IP subnet to the same SSID on a different AP associated with a different VLAN and IP subnet.

When designing a VoWLAN infrastructure, try to take roaming into consideration before deciding which AP should be associated to which controller. Roaming design takes roaming path into account: users do not roam through walls, but walk through corridors. Although internetwork roaming performances are high from the controller standpoint, network issues, such as the number of routing hops from one controller to the next, may delay the roaming exchange and the subsequent messages. Try to design the roaming paths so that roaming primarily occurs on the same controller. You can then leave roaming between controllers in the same subnetwork, if required, for less common roaming paths, such as when users move from one floor to another, or from one area of the building to the other. Try to avoid roaming between controllers in different subnetworks.

For the example in the illustration, when a client moves from the time and location of t1 to t2 and asks for reauthentication on a new AP, the query is sent to the controller to which the AP is connected. If this controller is the same as the one to which the AP that the client is leaving is associated, “roaming” is just a matter of registering that the client is connecting by means of another AP on the same controller. This internal operation takes less than 10 ms. This is known as intracontroller Layer 2 roaming.

When, at t3, the client roams to an AP connected to a different controller, the new controller exchanges mobility messages with the original controller, and the client database entry is moved to the new controller. This process takes less than 20 milliseconds and is transparent to the client. This is known as intercontroller Layer 2 roaming.

www.CareerCert.info




Client Roaming Across Subnetworks

A Layer 3 roam event requires more processing and controller coordination than a Layer 2 roam event. To the client, the process is seamless. The client will not get a new IP address or have to reauthenticate. The controllers create a tunnel that is used to trick the network and client into thinking that the client has not changed subnets.

As the client moves from one AP to another, control of the client passes from a controller on one subnet to a controller on a different subnet. The client maintains connectivity, and the approximately 30-ms event will not cause any client disruptions.

The controllers exchange mobility messages on the client roam. However, instead of moving the client’s entry to the client database for the new controller, the original WLAN controller marks the client with an “Anchor” entry in its own client database. The database entry is copied to the new controller client database and marked with a “Foreign” entry in the new controller. The client is reauthenticated to establish a new security context and the client database entry is updated for the new AP with which the client is associated. The choreography on the back end is totally opaque to the wireless client and the wireless client maintains its original IP address.

www.CareerCert.info




Symmetric vs Asymmetric RoamingOn controller code v.5.1 and earlier, you can configure roaming to be symmetric or asymmetric.On controller code v5.2 and later, roaming mode is not configurable—it is always symmetric.

When running controller code v5.1 and earlier, the controllers can be configured for two different types of roaming:

Symmetric: All traffic to and from the client is tunneled between the foreign controller and the anchor controller.

Asymmetric: Traffic from the wireless client is passed via normal IP routing to the destination, while returning traffic is passed to the originating anchor controller and then tunneled to the foreign controller before returning to the client.

The impact of choosing symmetric or asymmetric roaming for voice is directly related to the latency of the communication flow between both controllers. With symmetric roaming, packets always have to go to the anchor controller. This may cause a longer delay in the communication, which can be heard if both phones are actually in range of each other (callers will hear a delay between when they say the words and when they hear them on the other end). This effect does not create communication issues, as the phones buffer the voice traffic. It creates a slight “delay start” (voice starts being heard slightly after the communication begins), but it is usually not noticeable.

With asymmetric roaming, the packets transit time from the first phone to the second is not the same as the packet transit time on the way back from the second phone to the first one. Here again, the efficiency of the roaming process depends on the efficiency of the network infrastructure. End device buffers usually compensate for this difference, but the return path is more prone to jitter issues, as it travels through more internetwork devices.

When running controller code v5.2 and later, controllers are set for symmetric roaming only, and asymmetric roaming is not available as a configuration option. The above illustration is only relevant when controllers are running code release 5.1 or earlier.

www.CareerCert.info



Campus Network Design This topic describes the campus deployment of voice over wireless LANs.


Centralized vs Distributed Controllers Architecture

In a distributed model, controllers are close to the APs in each buildingIn a centralised model, APs connect to the controllers through the campus backbone

Within a Cisco Unified Wireless deployment, the primary design considerations are: AP proximity to the controller and the method of connectivity.

The flexibility of Cisco Unified Wireless LAN solution leads to the following choices about where to locate controllers:

Distributed controller deployment: In the distributed model, controllers are located throughout the campus network, typically on a per-building basis, managing the APs that are resident in a given building. The controllers are connected to the campus network using the distribution routers within each building. In this scenario, the CAPWAP tunnels between APs and a controller typically stay within the building. Each of the distributed controllers can be configured as a separate RF group and mobility group, so long as the WLAN coverage does not overlap between buildings.

Centralized controller deployment: In this model, controllers are placed at a centralized location in the enterprise network. This deployment model requires the AP and controller CAPWAP tunnels to traverse the campus backbone network. A centralized controller cluster is connected by means of a dedicated switch block to the campus core, which is typically located in the same building where the data center resides. The controllers should not be connected directly to the data center's switching block because the network and security requirements of a data center are generally different then that of a controller cluster.

www.CareerCert.info




Centralized ControllersCentralized configurationCentralized securitySingle connection point to security and traffic control devices

The general recommendation for VoWLAN in a campus architecture is to deploy the controllers at a central location within the overall campus environment. The distributed deployment model (which requires mobility groups and Layer 3 roaming) is well proven, but it is not recommended, because of added delay associated with Layer 3 roaming.

The best way to address Layer 3 roaming is to avoid deployment scenarios that would necessitate it. Currently, large mobility subnets are more feasible to implement due to the scaling capabilities of the Cisco Catalyst 6500 Series Wireless Services Module (WiSM) module. By centralized the controller infrastructure, you can make capacity management simpler and more cost-effective. Also, as WLANs become more mission-critical, centralized deployments make it easier to create a high-availability controller topology. Centralization reduces the number of locations where capacity management and high-availability issues must be dealt with.

The same principle applies when integrating the controller with other infrastructure components. Centralized controllers minimize the number of integration points and integration devices. For example, if a decision is made to implement an inline security device such as a NAC appliance, the centralized controller will have one integration point, whereas a distributed solution will have n integration points, where n equals the number of locations where controllers are deployed.

When planning any centralized controller deployment, consideration must be given to protecting the wired network infrastructure that directly connects to the controller. This is because the controller essentially attaches an “access network” at a location within the overall enterprise topology that would not otherwise be exposed to the ”access network” and its associated vulnerabilities. Therefore, all security considerations normally associated with an access layer network device must be considered. For example, in a deployment based on Cisco WiSM, features such as protection against denial of service and traffic storms should be considered. This is because of the large-scale role the Cisco WiSM plays in providing diverse WLAN services to large numbers of end users, while at the same time being directly connected to the backplane of a core multilayer, multifunction Cisco Catalyst 6500 switching platform.

www.CareerCert.info




Distributed controllers

Distributed WLCs are typically connected to the distribution layer router within the campus network. If this is the case, Cisco recommends against using connecting the WLC to the distribution layer by means of a Layer 2 link. It recommends against doing do for a number of reasons, including the following:

General best-practice campus design recommends Layer 3 access and distribution connectivity to provide fast convergence and simplified operation; inserting a WLC connected to Layer 2 breaks this model.

Layer 2 WLC connectivity requires the introduction of access layer features at the distribution layer, such as Hot Standby Router Protocol (HSRP), and access layer security features. This may be an issue if the distribution layer does not support all the preferred access switches, or needs to have its software version changed to support access features.

A WLC connected to Layer 3 allows the WLAN-related software and configuration to be isolated to a single device and connects to the distribution layer using the same routing configuration as other the access layer routing devices.

www.CareerCert.info




Traffic Load ConsiderationsCAPWAP control: 0.35 Kb/s per AP, on averageCAPWAP encapsulation overhead: 15 percent for voice traffic, 44 B per frameAll traffic is sent first to the controller

When deploying a Cisco Unified Wireless solution, questions often arise concerning:

CAPWAP traffic impact and load across the wired backbone and its impact on voice traffic

Minimum performance requirements to support a voice over wireless deployment.

Relative benefits of a distributed versus centralized WLC deployment in the context of traffic load on the network.

In examining the impact of CAPWAP traffic in relation to overall network traffic volume, there are three main factors to consider:

Volume of CAPWAP control traffic: the volume of traffic associated with CAPWAP control can vary depending on the actual state of the network. That is to say, it is usually higher during a software upgrade or during WLC reboot situations. Nevertheless, traffic studies have found that the average load CAPWAP control traffic places on the network is approximately 0.35 Kb/s. In most campuses, this would be considered negligible, and would be of no consequence when deciding between a centralized deployment model and a distributed one.

Overhead introduced by tunneling: a Layer 3 CAPWAP tunnel adds 44 B to a typical IP packet to and from a WLAN client. Given that average packets sizes in typical G.711 flows are less than 300 B (the exact value varies depending on which Layer 2 technology the packet is carried on), this represents an overhead of approximately 15 percent. In most campuses, this overhead would be considered negligible, and again would be of no consequence when deciding between a centralized deployment model and a distributed one.

Traffic engineering: any WLAN traffic that is tunneled to a centralized WLC is then routed from the location of the WLC to its end destination in the network. Depending on the length of the tunnel and location of the WLC, WLAN client traffic may not otherwise follow an optimal path to a given destination. In the case of a traditional access topology or distributed WLC deployment, client traffic enters the network at the edge and is optimally routed from that point based on destination address.

www.CareerCert.info



Note The scenario described above applies to APs set to Local mode. APs in Hybrid Remote-Edge Access Point (H-REAP) mode do not support wireless CAC. Cisco does not recommend using H-REAP for large voice deployments.

www.CareerCert.info




AP VLANs

APs should be on different subnets from the end users. This is consistent with general best-practice guidelines, which specify that infrastructure management interfaces should be on a separate subnet from end users. Cisco also recommends limiting the number of APs in the controller management interface subnet to no more than 16.

DHCP is generally the recommended method for AP address assignment, because it provides a simple mechanism for providing up-to-date WLC address information for ease of deployment. A static IP address can be assigned to APs, but requires more planning and individual configuration. Only APs with console ports permit static IP address configuration.

www.CareerCert.info




Campus Architecture

A classical campus network is organized into three layers: access, distribution, and core. In the past, it was common to run Layer 2 links between the access layer and the distribution layer, and to use Spanning Tree Protocol (STP) to create a redundant, loop-free topology. Current design best practice replaces this Layer 2 STP configuration with a Layer 3 configuration. Routing in the access layer allows for faster reconvergence around failure, and simplifies the network by eliminating the need for STP and HSRP.

In this architecture, multiple access layer blocks are connected to a single distribution block. In a production deployment, it is common that each access layer block connects to a dedicated distribution block, where policy appropriate to that type of access layer is applied. All router interconnections use fiber-optic transport due to the superior link failure detection capabilities fiber-optic links have when compared to wired Ethernet connections.

www.CareerCert.info




QoS in the Campus Architecture

For maximum efficiency, the QoS policy requested by wireless clients should be extended to the rest of the infrastructure, with the following recommended logic:

Access switches require the following QoS policies:

— Appropriate (endpoint-dependent) trust policies, and classification and marking policies

— Policing and markdown policies

— Queuing policies

Distribution and core switches require the following QoS policies:

— DSCP trust policies

— Queuing policies

www.CareerCert.info




Time Synchronization

An essential element of network management, troubleshooting, and security operations is to have all network elements (including routers switches and servers) synchronized to a common clock source.

Network Time Protocol (NTP) is most commonly used to synchronize clocks in network equipment. NTP provides a level of accuracy typically within a millisecond on LANs and up to a few tens of milliseconds on WANs.

By synchronizing the clocks across the network, it is possible to examine the exact sequence in which events occurred. This ability to analyze and correlate the sequence of events across multiple network elements makes it much easier to determine the root cause of network problems and security issues. In most production, the external time source used is redundant; dedicated hardware synchronizes via a Global Positioning System (GPS) receiver to a clock source that is itself directly synchronized to an atomic clock.

www.CareerCert.info



Voice Support in Mesh Environments This topic describes how voice is supported in a wireless mesh network.


Voice over Enterprise (Indoor) MeshOnly Indoor Mesh for v5.2AP-to-AP distance of 200 feetwith a cell radius of 100 feetHop count maximum of 2RAP handles 3 to 4 mesh access points

Starting with Wireless LAN Controller Software v5.2, there is very limited support for voice over Enterprise Mesh (also known as Indoor Mesh). Voice is only supported on Indoor Mesh Cisco Aironet 1130 Series APs and Cisco Aironet 1240 Series APs with controller code release 5.2. In addition to basic hop count and distance restrictions, the following are recommended:

Limit the coverage hole between cells to the minimum, typically targeting as low as 2 percent. When this number can not be achieved, ensure that coverage holes never exceed 10 percent.

Cell coverage overlap should be set to “voice values,” 15 to 20 percent.

RSSI and SNR values for both client and AP signals should be at least 15 dB higher than data requirements, following typical VoWLAN design recommendations.

An RSSI of –62 dBm is recommended on a 24 Mb/s 802.11a backhaul when universal access is configured and client traffic is present.

Packet error rate (PER) must be configured for a value of one percent or less.

The channel with the lowest utilization should be used.

While Radio Resource Management (RRM) can be used to implement the recommended RSSI, PER, CU, cell coverage, and coverage hole settings on the 802.11 b/g radio, RRM is not supported on the 802.11a radio.

www.CareerCert.info




Enterprise Mesh Configuration Overview

Voice shares the platinum queue with control traffic. When voice traffic is present, voice and control traffic are sent with the same platinum QoS level over the backhaul.

It is possible to enable CAC for WMM clients. There are some limitations to this feature. Wireless CAC is static and bandwidth-based. Load-based CAC is not supported. On the backhaul, there is no dynamic CAC and there are no QBSS load enhancements.

When configuring voice support for a mesh network, follow the following recommendations:

Disable all data rates less than 11 Mb/s for the interface to be used (802.11b/g or 802.11a).

Enable Dynamic Target Power Control (DTPC).

Set the EDCA profile to be voice-optimized for the interface (802.11b/g or 802.11a).

Disable the Low Latency MAC feature; it enhances voice performances in normal indoor conditions, but is not fully adapted to mesh network as of Wireless LAN Controller Software v5.2 .

Set the Max RF bandwidth allocated to 50 percent.

Set the Reserved Roaming Bandwidth to 6 percent.

Enable traffic stream metrics (TSM) to create a baseline and monitor the initial traffic.

Configure QoS as platinum on the WLAN, and allow Wi-Fi Multimedia (WMM).

Enable CISCO CKM (only) on the WLAN for fast roaming.

Edit QoS profiles for Voice for 802.11b/g and 802.11a.

Disable voice activity detection (VAD).

Note Check Cisco.com for updates to the mesh AP design guide, as the configurations and recommendations can change after the version 5.2 code release.

www.CareerCert.info





SummaryVoWLANs should be planned by taking into consideration client types and density, bands, and areas to cover or isolate.On the 2.4-GHz spectrum, planning is key to success, as only three non-overlapping channels are available, whereas on the 5-GHz spectrum, more available channels make deployments easier.Voice and data should be isolated on the wireless side, as well as on the wired side.VoWLAN clients should receive the same level of protection as data clients.


SummaryRoaming is a key consideration of VoWLAN planning, and the roaming path should be studied to prioritize intracontrollerroaming patterns.The Cisco Compatible Extensions program enhances roaming patterns by adding features that increase roaming speed and anticipate roaming.Controllers should be positioned so that voice traffic is not delayed when transiting from and to APs.The campus should be prepared for voice with proper device placement, QoS configuration, and other traffic optimization.Voice is supported in mesh networks, with design- and configuration-specific requirements.

www.CareerCert.info



www.CareerCert.info


Lesson 4

Verifying Voice Readiness

Overview Once a wireless network has been deployed for voice support, some tools are available to assess if the deployment matches the required criteria for calls in order to achieve the required quality level. Onsite verification is the logical first step. Cisco WCS also provides a Voice Readiness tool. The Voice Readiness tool allows you to verify, even before going to the facility, if the access point position will provide voice coverage. Once you are on site, a specialized tool can help you troubleshoot VoWLAN issues.

Objectives Upon completing this lesson, you will be able to verify if your wireless network is ready for voice deployment. This ability includes being able to meet these objectives:

� Use the Cisco WCS Voice Readiness tool

� Conduct a postdeployment site survey

� Use the AirMagnet VoFi Analyzer

www.CareerCert.info



Cisco WCS Voice Readiness Tool This topic describes the Cisco Wireless Control System (Cisco WCS) Voice Readiness tool.


Inspect VoWLAN Readiness

� Cisco WCS predictive tool

� Verifies a configurable RSSI level

There are several ways to verify if a wireless infrastructure is ready for voice communications. The first one is integrated in the Cisco Wireless Control System (WCS), and is called Inspect VoWLAN Readiness. This tool allows you to check your maps for which areas are within the acceptable Received Signal Strength Indicator (RSSI) level for voice.

By default, the threshold is defined at – 67dBm, but it can be customized. The tool is predictive: it relies entirely on the information you provide to Cisco WCS. If the information about floor sizes, walls, or other obstacles is not accurate, the tool cannot provide useful information. Before using the tool, make sure that the map reflects, as much as possible, the real environment conditions. Calibrating the floor maps can also help enhance the accuracy of the information provided by the tool.

As the Inspect VoWLAN Readiness tool is a predictive tool, it can be used in two situations:

� pre-deployment simulation: to help you determine the best AP positions and run “what if” simulations

� postdeployment verification or troubleshooting: to determine where voice coverage is acceptable and where voice quality suffers from the RF environment

The Inspect VoWLAN Readiness tool is conducted map-per-map. The tool is launched from the map you are reviewing, using the Inspect VoWLAN Readiness drop-down option.

www.CareerCert.info




A new screen appears with the VoWLAN readiness information. You can change the default values by selecting the applicable Band, AP Transmit Power, and Client parameters from the drop-down menus. By default, the region map displays the region map for the IEEE 802.11b/g/n band for an RSSI threshold based on a Cisco Unified IP phone. The new settings cannot be saved.

Depending on the selected client type, the RSSI values may not be editable:

� Cisco Phone: RSSI values are not editable.

� Custom: RSSI values are editable within the following ranges:

— Low threshold between –95dBm and –45dBm

— High threshold between –90dBm and –40dBm

The following color schemes indicate whether or not the area is Voice Ready:

� Green: Yes

� Yellow: Marginal

� Red: No

The accuracy of the green, yellow, and red regions depends on the RF environment and whether or not the floor is calibrated. If the floor is calibrated, the accuracy of the regions is enhanced.

www.CareerCert.info



Postdeployment Site Survey This topic describes how to conduct a postdeployment site survey to verify voice readiness.


Postdeployment Site Survey

Although a site survey should have been conducted before deployment, and despite the fact that the Inspect VoWLAN Readiness tool is an efficient verification tool, onsite verification is usually needed after deployment. Even if the predeployment site survey was conducted with great care, there are always differences between the expected and the actual coverage. Unexpected sources of interferences or changing conditions must be taken into account to ensure proper VoWLAN coverage.

The Inspect VoWLAN Readiness tool is a predictive tool. It can provide information about access point (AP) placement if used in a predeployment scenario, but cannot take into account elements not present in Cisco WCS. It shows the coverage as it should be from the information provided. The system has no graphical representation for unknown obstacles. It is certainly possible that actual coverage will be different from the predictive map, especially if the information provided was insufficient or inaccurate.

In a changing environment, unexpected events may also alter the coverage characteristics. People are made of 70% water, which absorbs wireless signals. Moving people, doors that open and close, partition walls, and desks that move around create changing conditions in the coverage pattern.

www.CareerCert.info




Postdeployment Site Survey (Cont.)

A postdeployment site survey provides more than just the RSSI level; it provides coverage, interference, and general performance information for your wireless LAN solution. This allows you to objectively measure the operability and functionality of the system you are deploying, to verify that it meets your technical requirements and is ready for production. The postdeployment site survey should be conducted using the voice device that you plan to use in the environment.

Ideally, the whole area should be surveyed. Each different surface in an indoor site will have a different effect on signal reflections and signal absorption. Each device that is running with a 2.4-GHz or 5-GHz frequency will affect the signal-to-noise ratio.

Elevator shafts block signals, supply rooms absorb signals, interior offices absorb signals, break rooms may produce 2.4-GHz interference, test labs may produce 2.4-GHz or 5-GHz interference and create multipath distortion and shadows, cubicles absorb and block signals, and conference rooms have high utilization requirements.

Expect that 5-GHz equipment will not radiate through walls as well as 2.4-GHz equipment.

www.CareerCert.info




Cisco Unified Wireless IP Phone 7921G Site Survey

Limited survey tool:

� Uses the RSSI value to determine cell size

� Presents a list of access points

� Accessed from the menu by choosing Settings > Status > Site Survey

The Cisco Unified Wireless IP Phone 7921G provides a basic site survey utility that is ideal for postdeployment verification.

The site survey utility can be used to actively and passively scan the wireless medium across all channels and locate APs that belong to the Basic Service Set (BSS).When you start the Site Survey utility, the phone disassociates from the current AP and remains disassociated for the duration of the operation.

To use the site survey utility, follow these steps:

Step 1 Configure the Wireless IP Phone 7921G with the same SSID, encryption, and authentication settings as the APs.

Step 2 Make sure that the phone associates to the WLAN.

Step 3 Choose Settings > Status > Site Survey. The phone displays a list of APs within range that have the same SSID and security settings as the phone. For example:

SSID: abcd (example value for the SSID)

Channel BSSID RSSI Channel Utilization

01 19:50 -38 50

06 cf:d0 -51 38

11 7b:b0 -42 61

To see more information about an AP, scroll to the desired line and press Details.

www.CareerCert.info



The following is an example of the details for a specific AP:

SSID: abcd

Channel:06

BSSID: 00:13:1a:16:cf:d0

RSSI:-51

CU:38

To verify the ability to roam between APs, follow these steps:

Step 1 Walk through all areas where phones are used and take readings. Approach areas from different directions to assure successful roaming conditions.

Step 2 Adjust AP and antenna placement and AP power settings to provide approximately 20 percent coverage overlap.

When you terminate the site survey, a report is generated from the phone web page. The phone web page can be access by connecting a USB cable to the phone.

www.CareerCert.info



AirMagnet VoFi Analyzer This topic describes the AirMagnet VoFi Analyzer and how to use it for VoWLAN verification.


AirMagnet VoFi Analyzer

� Automatically identifies and tracks wireless calls

� Measures call quality using MOS Score and R-Value

� Automatically diagnoses the voice problems

� Integrates with the PBX or Cisco Unified Communications Manager

� Visually correlates call-quality with voice metrics

� Provides professional reporting

� Validates QoS settings (802.11e)

� Supports encrypted networks

VoWLAN call quality does not rely solely on wireless parameters. Many other factors can impact the quality of VoWLAN deployments, such as issues in the phone itself, the RF environment, the WLAN, QoS settings, the PBX or VoIP gateway, and so on. This complexity can make voice problems particularly challenging and time-consuming to diagnose. The Cisco WCS VoWLAN Readiness tool and the postdeployment site survey can help you troubleshoot WLAN-related issues, but cannot detect problems that occur on the wired side. These tools also cannot provide any feedback on the call quality itself, such as mean opinion score (MOS) information.

AirMagnet VoFi Analyzer is a program dedicated to analyzing VoWLAN calls. It automatically analyzes each area along the call path and identifies the source of any Wi-Fi voice problem.

www.CareerCert.info




AirMagnet VoFi Analyzer (Cont.)

AirMagnet VoFi Analyzer displays the network in terms of calls and call quality. It scans all 802.11a/b/g devices, distinguishes between voice and data traffic, and automatically scores every call in terms of a WiMOS (wireless MOS) based on a variety of packet metrics, such as loss rate and jitter. The solution independently scores and tracks both sides of a call (AP-to-phone and phone-to-AP), allowing users to distinguish problems that are occur primarily on one side of the call. Each call is color-coded according to call quality, making it easy to visualize phone and call problems. In provides a full history of all calls on the network, and quickly reveals problems that are tied to a particular phone, channel, or given period in time.

The VoFi Analyzer also provides full visibility into all roaming events, allowing users to see which calls roamed, how often calls roamed, and the APs and channels that calls roamed to. This can help you identify problem phones that may be constantly roaming or “thrashing” between APs. It can also help you identify problems that are affecting an entire channel, and may be causing many phones to roam. Additionally, the VoFi Analyzer includes a “Follow Phone” mode that lets it lock on to a particular phone and follow the phone through any number of AP and channel roams, ensuring the deepest level of phone and call analysis.

www.CareerCert.info




AirMagnet VoFi Analyzer (Cont.)

AirMagnet VoFi Analyzer users can click on a call to gain access to detailed diagnostic information. The solution provides charts that correlate changes in call quality with about 50 critical call metrics. For example, a user could use these charts to see if poor call quality was related to changes in packet jitter or signal quality, or if it was related to a spike in the number of users competing for the same AP. These call charts also show when roaming events occurred, as well as when alarms were triggered, making it easy to place key events in relation to changes in performance. Users can choose from a library of prebuilt graphs or construct their own from dozens of metrics, such as CRC errors, fragmentation, active call count, and data utilization.

AirMagnet VoFi Analyzer provides details on how each call fits into the overall wireless environment. The application shows how many voice and data clients are competing for the same AP and channel resources. In addition, the analyzer displays the relative amounts of voice and data traffic and also tracks key 802.11e statistics such as video, best-effort, and background traffic. Finally, the solution gives a connection history for all devices on the channel, illustrating which devices have roamed the most.

www.CareerCert.info





Summary

� The Cisco WCS Voice Readiness tool provides predictive analysis of VoWLAN readiness by comparing AP distances and RSSI levels

� A local postdeployment site survey is usually required to verify, with an actual client device, coverage and call quality

� Wireless troubleshooting tools cannot determine MOS scores, and VoIP analysis tools cannot always identify where the qualityissue lies; the AirMagnet VoFi analyzer is a hybrid tool that analyzes VoWLAN and matches voice issues to WLAN- and non-WLAN-related sources.

www.CareerCert.info



Module Summary This topic summarizes the key points that were discussed in this module.


Module Summary

� In traditional voice architecture, PBXs and CO switches provide call-routing and call control services to analog phones.

� In a VoWLAN, the VoIP phone first has to associate to the wireless infrastructure before registering with Cisco Unified Communication Manager; considerations such as voice prioritization, delay, and jitter impact the voice quality.

� When designing a VoWLAN, great attention should be given to AP power level, cell size, overlap between adjacent cells, minimum speeds at the edges of cells, and the impact of building materials on co-channel interferences.

� Once a wireless network is deployed, the Cisco WCS Voice Readiness tool, a post-deployment site survey, and the AirMagnet VoFi Analyzer can help determine if the wireless network is voice-ready.

Migration from traditional voice networks to VoIP networks involved replacement of operators, semi-automatic switches, and PBXs with communication management software. Analog phones became IP phones, capable of encoding voice streams into binary digits and negotiating some communication parameters on their own.

The logic of building a voice over wireless LAN (VoWLAN) is not to add voice devices to an existing data wireless network, but to build the wireless infrastructure so that it matches the voice-specific needs: steady flows and consistently low delay to avoid jitter. Cells have to be smaller than for data, to increase their throughput and overall Layer 2 efficiency. A stricter control in the design ensures that wireless devices in the cell will not compete for bandwidth access. In other words, the wireless infrastructure is built to resemble the low-latency Ethernet network, and give voice traffic the bandwidth it needs to ensure smooth communication patterns.

www.CareerCert.info


© 2009 Cisco Systems, Inc. Voice Over Wireless Architecture 2-143

Module Self-Check Use the questions here to review what you have learned in this module. The correct answers and solutions are found in the Module Self-Check Answer Key.

Q1) Which of these best describes a CO switch? (Source: Describing Traditional Voice Architecture)

A) a device responsible for call control and management B) the corporate voice termination endpoint C) a device responsible for VoIP device interconnection D) the old name for the first two cable telephones

Q2) What is the PSTN? (Source: Describing Traditional Voice Architecture)

A) the connection point from the enterprise network to the telephone company network

B) the classical phone network C) a numeric-to-analog converting device D) a logical link between to analog phones

Q3) What is the role of a PBX? (Source: Describing Traditional Voice Architecture)

A) interconnect the VoIP networks of two companies B) replace the human central operator C) interconnect the corporate phone network D) convert analog signal to digital signal

Q4) What does the centralized model refer to when talking about a VoIP network? (Source: Describing Traditional Voice Architecture)

A) Each building has a main building switch through which the phones connect to the outside world.

B) All the phones are of the same brand and type. C) The call management software is centralized. D) Users cannot use mobile devices and must use fixed phones.

Q5) Which is better, the centralized model or distributed model? (Source: Describing Traditional Voice Architecture)

A) centralized, because control is easier B) distributed, because resistance to network issues is better C) Neither is better; the correct model choice depends on the enterprise

organization D) Neither is better; large networks should use the distributed model, but small

and medium-sized networks should use the centralized model

Q6) When a VoIP wireless phone is turned on, which step or steps are required before you can place a call? (Source: Describing Voice as It Applies to Wireless Networks)

E) The phone has to register with Cisco Unified Communications Manager before being allowed into the wireless infrastructure.

F) The phone has to associate to the wireless infrastructure and then register with Cisco Unified Communications Manager.

G) The phone just has to register with Cisco Unified Communications Manager in order to get an IP address and associated extension number.

www.CareerCert.info



H) The phone has to associate to the wireless infrastructure, but does not have to register with Cisco Unified Communications Manager until a call is received.

Q7) What is the role of QBSS? (Source: Describing Voice as It Applies to Wireless Networks)

A) used by a phone to inform the AP about roaming decisions B) used by an AP to communicate client information to other APs over the air C) used by a phone to evaluate jitter and delay levels in the cell D) used by a phone to determine the best AP to associate with

Q8) How does the wireless infrastructure determine how many calls can be placed in a given cell? (Source: Describing Voice as It Applies to Wireless Networks)

A) By relying on communication with Cisco Unified Communications Manager about voice client counts

B) by relying on AP CAC values C) by relying on wireless IP phone count values D) by running the WMM algorithm

Q9) Which of these is the default codec used in VoIP? (Source: Describing Voice as It Applies to Wireless Networks)

A) G.711 B) G.726a C) G.729u D) G.726u

Q10) What is the MOS of the G.729 codec? (Source: Describing Voice as It Applies to Wireless Networks)

A) 2.8 B) 3.7 C) 4.4 D) 5.0

Q11) Which of these is present on a Cisco Unified Wireless IP Phone 7921G? (Source: Describing Voice as It Applies to Wireless Networks)

A) SIP client B) SCCP client C) MGCP client D) SRVP client

Q12) Which of these values is the most important in designing wireless VoIP cells? (Source: Designing Wireless for Voice)

A) the RSSI value B) the SNR value C) SNR and RSSI are equally important D) the ratio of power level to speed at the cell edge

Q13) What is the recommended overlap between cells in a VoWLAN deployment? (Source: Designing Wireless for Voice)

A) 10 percent B) 20 percent C) 50 percent D) 80 percent

www.CareerCert.info



Q14) Which of these is true? (Source: Designing Wireless for Voice)

A) the higher the power level, the higher the SNR B) the higher the SNR, the better the RSSI C) the higher the power level, the higher the speed at a given point D) the higher the SNR and RSSI, the higher the data rate

Q15) What is the recommended minimum speed in an 802.11b/g environment? (Source: Designing Wireless for Voice)

A) 1 Mb/s B) 5.5 Mb/s C) 11 Mb/s D) 24 Mb/s

Q16) What is the maximum recommended number of concurrent calls in an 802.11a environment? (Source: Designing Wireless for Voice)

A) 7 B) 14 C) 20 D) 27

Q17) Which of these is used by a wireless IP phone when it is deciding whether to roam? (Source: Identifying and Describing Other Design Considerations)

A) Laptops Present information element B) Battery Client Count information element C) Battery level versus AP distance D) Data retry threshold

Q18) Which of these is the process used by a client to discover new access points? (Source: Identifying and Describing Other Design Considerations)

A) background scanning B) AP mining C) roam pinging D) probe catching

Q19) Which new protocol enhances roaming in a VoWLAN environment? (Source: Identifying and Describing Other Design Considerations)

A) 802.11e B) 802.11i C) 802.11k D) 802.11r

Q20) Which of these is a benefit of Cisco Compatible Extensions AP-assisted roaming? (Source: Identifying and Describing Other Design Considerations)

A) battery conservation B) hole coverage distribution C) AP load balancing D) bandwidth pre-reservation

www.CareerCert.info



Q21) Which of these best describes the AP-specified power feature using Cisco Compatible Extensions? (Source: Identifying and Describing Other Design Considerations)

A) The AP broadcasts its power level, thus allowing clients to deduce distance from the RSSI.

B) The AP turns puts clients into sleep mode when no traffic is buffered, thus saving battery.

C) The AP can change its power on a per-client basis, thus always using the optimal power level.

D) The AP can request a client power level, thus optimizing client signal.

Q22) Which one of these roaming strategies provides the best roaming efficiency for VoWLAN devices? (Source: Identifying and Describing Other Design Considerations)

A) Roaming should be intracontroller as often as possible B) Roaming should be at Layer 3 (internetwork) as often as possible C) Roaming should be intercontroller as often as possible D) Roaming should be at Layer 2 (intranetwork) as often as possible

Q23) Which of these is true? (Source: Identifying and Describing Other Design Considerations)

A) A distributed controller model is preferable in a campus type of environment. B) A centralized controller model is preferable in a campus type of environment. C) Distributed controllers provide a simpler security configuration model. D) Centralized controllers imply one single point of failure.

Q24) What overhead does CAPWAP control traffic typically add to network traffic? (Source: Identifying and Describing Other Design Considerations)

A) 44 B per frame B) 15 percent for data traffic C) 0.35 kb/s per AP D) 35 kb/s per AP

Q25) On which parameter does the Cisco WCS Voice Readiness tool rely to determine “voice ready” areas? (Source: Identifying and Describing Other Design Considerations)

A) AP interdistance B) predicted RSSI C) predicted data rate D) AP density

Q26) Which information can AirMagnet VoFi Analyzer provide that purely wireless tools cannot? (Source: Identifying and Describing Other Design Considerations)

A) The wired QoS policy B) RSSI thresholds C) global client count on the controller D) call MOS

www.CareerCert.info



Module Self-Check Answer Key Q1) A

Q2) B

Q3) C

Q4) C

Q5) C

Q6) B

Q7) D

Q8) B

Q9) A

Q10) B

Q11) B

Q12) C

Q13) B

Q14) D

Q15) C

Q16) C

Q17) D

Q18) A

Q19) D

Q20) A

Q21) D

Q22) A

Q23) B

Q24) C

Q25) B

Q26) D

www.CareerCert.info



www.CareerCert.info


iuwvn v1.0 student guide_vol1

Documents