Mobile QoS Primer

download Mobile QoS Primer

of 33

Transcript of Mobile QoS Primer

  • 7/27/2019 Mobile QoS Primer

    1/33

    Optimizing Quality of Service in Mobile Net works

    Why, Wh en, Wh ere and How

    Quality of Service in Mobile Networks

  • 7/27/2019 Mobile QoS Primer

    2/33

  • 7/27/2019 Mobile QoS Primer

    3/333Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Preface: How to Use This Document . . . . . . . . . . . . . . . . . . .4

    Introduction

    Life as A Network Operator: To Stand Still is to Fall Behind . . . . . . . . . . . . . .4

    Quality of Service: Todays Key to Competitiveness . . . . . . . . . . . . . . . . . . . .5A Picture of the Business Case for QoS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .5

    The Profit Impact of QoS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6

    Installation

    Installation Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

    QoS Considerations for Installation Personnel . . . . . . . . . . . . . . . . . . . . . . . .7

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7

    BTS Transmit Tests are the Starting Point . . . . . . . . . . . . . . . . . . . . . . . . .8

    Looking at Modulation Accuracy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9

    Protocol Analysis Supports Installation of New Features . . . . . . . . . . . .10

    Protocol Analysis Solution Requirements . . . . . . . . . . . . . . . . . . . . . . . .11

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .11

    Operations & MaintenanceOperations & Maintenance Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

    QoS Considerations for the O&M Department . . . . . . . . . . . . . . . . . . . . . . .12

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .12

    Portable Testers Provide Cost-effective Monitoring . . . . . . . . . . . . . . . . .13

    New Features Mean Added Responsibility for Protocol Testing Tools . .14

    SS7 Signaling Network is the Connection for Network-wide Monitoring .14

    Call Generator/Analyzer is a Demanding Subscriber . . . . . . . . . . . . .16

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .17

    Planning & Engineering

    Planning & Engineering Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

    QoS Considerations for the P&E Department . . . . . . . . . . . . . . . . . . . . . . . .18

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .18

    Drive Tests Provide an End-to-End View of the Network . . . . . . . . . . . .18Test Plant Debugs Problems in a Neutral Environment . . . . . . . . . . . . .20

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .21

    Quality

    Quality Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .22

    QoS Considerations for Quality Personnel . . . . . . . . . . . . . . . . . . . . . . . . . .22

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .22

    Drive Tests Deliver Objective Measurements . . . . . . . . . . . . . . . . . . . . .22

    Non-intrusive Measurements View the Whole Network . . . . . . . . . . . . .23

    Call Generators Duplicate Real-world Calling Situations . . . . . . . . . . . .23

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .23

    Marketing

    Marketing Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .24QoS Considerations for the Marketing Department . . . . . . . . . . . . . . . . . . .24

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

    SS7 Monitoring Profiles Subscribers . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .25

    Security & Billing

    Security & Billing Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .26

    QoS Considerations for the Marketing Department . . . . . . . . . . . . . . . . . . .26

    Measurement Challenges and Solutions . . . . . . . . . . . . . . . . . . . . . . . . . . . .27

    SS7 Monitoring System Guards Against Fraud Throughout the Network .27

    Alarms, Fraud Reports Help Detect Offenders . . . . . . . . . . . . . . . . . . . .28

    Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .29

    Appendix A: About Tektronix Role in Communications Test . . . . . . . . . . . .30Appendix B: Mobile Telephony Standards . . . . . . . . . . . . . . . . . . . . . . . . . . .30

    Table ofContents

  • 7/27/2019 Mobile QoS Primer

    4/33

    The mobile communications marketplace has experienced

    tremendous subscriber growth rates during the past

    decade. Hundreds of millions of mobile subscribers are

    today using digital mobile phones, and growth is expectedto continue unabated. For example, GSM network

    subscribership alone is growing by up to 10 million

    customers per month, with totals forecasted to exceed

    250 million GSM users as of the end of 1999. 1These digital

    phones interact with a network infrastructure made up

    of varied base station types and switching equipment.

    In turn, each individual network must interact with every

    other, delivering the subscribers calls as though they were

    traversing one seamless worldwide network.

    Today most mobile networks are implemented with

    2nd-Generation (as distinguished from 1st-Generation

    analog cellular) technology. Widely referred to as 2G,current mobile technology encompasses standards such

    as GSM in Europe and cdmaONE and IS-136 in the U.S.A.

    2G digital technology was prompted mostly by the need to

    provide more voice telephony capacity and better coverage

    in urban and high-density geographic areas. 2G technology

    came to market with improved Quality of Service (QoS) and

    broader roaming capability.

    But todays best 2G mobile performance is just a stepping

    stone for the next round of subscriber demands. Mobile

    users want enhanced data delivery, Internet access, email,

    and more. The evolving solution is called, aptly enough,

    3rd-Generation, or 3G, mobile telephony. Standards such

    as Wideband-CDMA are being defined and refined even as

    this document is written.

    Ultimately all these trends say one thing about the mobile

    network market: to stay competitive, mobile network opera-

    tors must continuously expand and improve their services,

    meeting subscriber demands while controlling operating

    costs. And as we will see, Quality of Service plays a major

    role in network competitiveness and business success.

    Life as a Network Operator:

    To Stand Still is to Fall Behind

    Today there are huge pressures confronting mobile network

    operators: business pressures, technical pressures, competi-

    tive pressures the list goes on.

    Hundreds of millions of mobile phone users around the

    world rely on digital mobile telephony to keep in touch

    with friends, family, and business associates. Increasingly,

    this contact involves services above and beyond simple voice

    connections. The business of running a mobile network is

    a juggling act that involves keeping up with subscribers

    coverage and service demandsbut not outpacing them.

    Technical pressures add cost and complexity to the

    equation. Standards are evolving and hardware capabilities

    are advancing. Hardware and software compatibility issues

    are becoming more difficult to handleor even to define

    as new technology arrives almost daily.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Ultimately all these

    trends say one thing

    about the mobile

    network market:

    to stay competitive,

    mobile network

    operators must

    continuously expand

    and improve their

    services, meeting

    subscriber demands

    while controlling

    operating costs.

    Introduction

    How to usethis document

    This Mobile QoS Primer is designed to provide an overview of the problems, solutions, and benefits of a coherent network

    QoS test and measurement strategy.

    The Primer is not intended to be a cookbook to provide step-by-step solutions for specific measurement problems, nor is it

    an authoritative reference for standards compliance. It will emphasize the real application needs of todays mobile networks,

    and it will discuss solutions in general terms. For the sake of simplicity in this primer, we will focus on one or two solutions for

    each type of network application. Bear in mind that most solutions have overlapping applications, and may find use in virtually

    every division of a network.

    Fundamentally, the Primer is divided into chapters that follow the organizational lines of the typical successful mobile

    network business. In each chapter we will look at the unique responsibilities of a single activity.

    The chapters will cover the entire range of organizational responsibilities and activities within a network, namely:

    Installation Planning and Engineering Quality

    Operations and Maintenance Security and Billing Marketing

    Note here that not all network operators have an organization specifically titled the Quality Department. In many cases the

    quality organization is simply part of the planning and engineering group. Note also that QoS is a broad concept that touches

    every aspect of network operation, not just those who have quality as their primary responsibility.

    Each organizations chapter is made up of four major subtopics:Overview Measurement Challenges and Solutions

    QoS Considerations Conclusion

    Most likely you will want to go directly to the chapter that pertains to your own area of responsibility. After youve finished

    reading that, though, we recommend that you peruse the other chapters to learn about the similarities and differences in

    QoS issues as they cut laterally across every network activity.

    1 GSMWorld Web site; http://www.gsmworld.com/news/press_releases_23.html

  • 7/27/2019 Mobile QoS Primer

    5/33

    Moreover, mobile users are always seeking improved

    service quality and reliability at lower cost. A mobile

    subscriber tends to think only in terms of his or her latest

    attempt to make a call. If there is service degradation of any

    kind, the subscribers perception of the networks product(connectivity and services) suffers.

    Competitive pressures are in some ways the most onerous

    of all. The mobile market is dynamic, with new competitors

    displacing old, mergers creating mega-networks, and every-

    one striving to gain a bigger share of the market pie.

    Competition has brought prices down. Every day, consumers

    are barraged with special deals, signing bonuses and low-

    price offers. The cost of marketing is rising, profit margins

    are eroding, and end-users are becoming more sophisticated.

    Today they look at value rather than price alone.

    Small wonder, then, that life in the network operator busi-ness seems like a marathon with a constantly accelerating

    pace. Business, technology, and competitive factors are

    driving network operators to seek product differentiators

    that will allow them to lead the race, recruiting and

    retaining loyal subscribers.

    Quality of Service:

    Todays Key to Competitiveness

    What can a mobile network operator do to protect existing

    businessand perhaps even grow faster than the market

    as a whole? Increasingly, Quality of Service (QoS) is thedifferentiator that mobile service providers are using to

    define their place in the market. Lets take a look at why

    and how this trend is progressing.

    Most mobile subscribers seem to understand that they

    must look past mere price considerations when choosing

    a mobile service provider. Complex rate structures and

    discount programs can make it difficult to compare vendors,

    and such comparisons often reveal that actual costs are

    fairly similar no matter which vendor provides the service.

    But the quality of the mobile service the subscriber receives

    is a daily reminder of the vendors competenceand compet-itiveness. The value of mobile service lies in its convenience,

    and convenience is drastically compromised when there is

    poor reception, dropped calls, or erratic service.

    Mobile subscribers are a notoriously fickle bunch, willing

    to change to a new network provider for the tiniest of

    advantages. Little does it matter whether the subscriber

    can actually gain much by churning from vendor to

    vendor; all that counts is the perception that he or she

    might get better service or save some money by switching

    to the network next door. Changing providers has become

    so easy that mobile subscribers quickly forget the incentives

    they were so happy about when they signed up.

    Complicating matters even more, this churning scenario

    plays out against a backdrop of network compatibility

    issues. For example, number portability makes it simple for

    users to change service providers, even while it presents

    new support challenges to all providers. Or, the subscriberslocal provider may be unjustly blamed for a roaming

    problem that is actually the fault of another network. Lastly,

    administrative complexities such as discounts and billing

    procedures grow more cumbersome with each exciting

    new rate plan, and security breaches, including large-scale

    organized fraud, are an increasing threat.

    The solution lies in ever-improving Quality of Service. But

    QoS is more than guaranteed bandwidth, a clear and

    continuous signal, and reliable roaming access. Certainly

    these are the key benefits the consumer of mobile services

    will receive. But the astute businessperson will see that

    well-implemented QoS also serves the best interests of thenetwork provider.

    QoS initiatives can:

    Attract new, more sophisticated subscribers

    Increase subscriber satisfaction, reducing churn

    Motivate subscribers to adopt new, billable services faster

    and to use these services more frequently

    Improve industry relationships among diversenetwork operators, service providers and carriers

    Provide a better understanding of subscribers

    network usage, thus the subscribers themselves

    Protect against fraudulent network use and piracy

    Ensure compliance with standards andgovernmental regulations

    Many network business have come to these same

    conclusions, and have begun carrying out QoS programs

    within their organizations. As used in this document, the

    term QoS program has a broad definition that encom-

    passes all of the efforts that support network quality. These

    are not limited to test and measurement issues; they may

    range from improvements in BTS installation procedures

    to integrating a network-wide SS7 monitoring system.

    A Picture of the Business Case for QoSThe success of any business rests on its subscriber loyalty.

    And that is the crux of the problem in the mobile network

    business. Loyal subscribers are cost-effective subscribers;

    they are users whose continued business does not depend

    on constant infusions of costly marketing energy or new

    discount incentives with each passing month. Loyal, repeat

    subscribers produce the reliable revenue stream that a

    network needs to fund modernization programs, design

    new service offerings, and support expansion.

    5I N T R O D U C T I O N

  • 7/27/2019 Mobile QoS Primer

    6/33

    As we explained earlier, Quality of Service is the key to

    competitiveness and therefore business success. Thats

    because it is the path to subscriber loyalty. Figure 1 summa-

    rizes the concept. There is an endlessly recirculating loop

    of dependencies. No single link in the loop can be ignored,so the challenge for mobile network providers is to find the

    right place to step aboard. It is the mission of this Primer

    to show you how QoS programs can help a network break

    out of the cycle of price wars and churning, and into the

    quality improvement cycle shown in Figure 1.

    The Profit Impact of QoS

    Promoting subscriber loyalty is just one of the benefits of

    a conscientious QoS effort. In fact, there is an even more

    direct, measurable benefit of particular interest to shareholders

    and executives. It is the immediate profit that accrues whenmobile providers deliver clean, reliable connections.

    Simply stated,good service quality stimulates increased

    service usage. And simple math tells us that more usage

    means more billable minutes of air time. It has been shown

    that mobile users will use their phones more often, and

    stay on the phone up to 20% longer, when their calls are

    not hampered by noise, echoes, or other interruptions.

    A term known as the Quality Index has been devised as

    a measure of signal quality. It is a composite figure of merit

    that encompasses variables such as Echo Loss, Echo Path

    Delay, and other components that affect the users percep-

    tion of voice quality. Normally the Quality Index is not

    measured directly; rather, it is based on the analysis of

    many discretely measured parameters of traffic activity

    and signal degradation.

    Figure 2 introduces the concept graphically. It plots the

    number of calls (vertical axis) vs. Quality Index (horizontal

    axis). The two graphed areas represent Quality Indices

    determined by measurements taken on calls originated

    from a specific area, terminated at a specific destination,

    and carried on two separate trunks. Here, Trunk B (Red)

    achieved a mean Quality Index of about 40 out of a

    possible 100. Trunk A (blue) performed somewhat better,

    reaching a mean of 60.

    Now look at Figure 3 to understand the impact of these

    quality differences. Its plain to see that users of Trunk B

    used their phones less. These are call duration graphs for

    the same two trunks monitored in Figure 2. For the

    purposes of this discussion we will ignore the vertical scales

    in Figure 3, which denotes number of calls (this variable is

    not significant unless service is so bad that subscribers stop

    using their mobile phones altogether!).

    Looking at the call duration, the mean duration for Trunk A

    users was more than two minutes longerthan those on

    Trunk B. The subscribers on Trunk A stayed on the phone

    longer. They were getting consistently better connection

    quality (as Figure 2 shows), which implies they were

    relaxed and able to take their time conversing with the

    connected party. As a result, they happily generated higher

    billings for the network provider!Quality Index is

    not measured

    directly; rather,

    it is based on the

    analysis of many

    discretely measured

    parameters of

    traffic activity and

    signal degradation.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    250

    150

    50

    050 100

    Numberof

    Calls

    Quality Index

    100

    200

    Trunk B

    Trunk A

    Figure 3: Call Duration

    Figure 2: Quality Index

    QualityofSer

    vic

    e

    Custo

    merLoyalty

    ReliableRevenu

    e

    Serv

    iceEn

    hancements

    Figure 1: Quality of Service Cycle

    NumberofCalls

    Call Duration

    400 800 1200 1600

    Trunk BMean Durati on: 367 sec.

    500 1500 2500 3500

    Trunk AMean Duratio n: 508 sec.

    NumberofCalls

    Call Duration

  • 7/27/2019 Mobile QoS Primer

    7/33

    Installation Overview

    The networks field installation technicians are the first to

    touch new equipment as it goes into service. Organizationally,

    the installation team may be made up of technicians

    gathered from other groups within the network, particularly

    the O&M group. Often these individuals work closely with

    a team from the equipment manufacturer, or they may be

    assigned to double-check an installation the manufacturer

    has completed.

    The classic installation task is installing and turning up new

    base stations that add capacity or coverage to the network.

    The installers work can make the difference between a

    trouble-free BTS turn-up with continuing reliability, and an

    elusive noise or dropout problem that costs time and

    money to correct.

    Other installation responsibilities include the addition ofnew software-based features in existing facilities. A case in

    point is the General Packet Radio System (GPRS), a set of

    enhancements that improves the GSM networks ability to

    deliver data. This capability currently is being added to GSM

    networks in many regions of the world.

    QoS is an obligation of every department within the network.

    Some groups (such as the Operations & Maintenance

    organization) are likely to include job titles and duties

    pertaining to QoS implementation. While installers may not

    have the same specific job descriptions, that doesnt exempt

    them from ensuring the highest possible quality in their area

    of responsibility. A conscientious installation effort is crucialto QoS throughout the network.

    QoS Considerations

    for Installation Personnel

    An installers thoroughness in applying and verifying high

    QoS standards when a BTS or new software is installed can

    prevent problems at that particular site and over a much

    larger area as well. After all, subscribers cant pinpoint the

    origin of a weak or distorted voice signal to one newly-

    installed BTS; they will tend to blame the networks service

    as a whole. Therefore new installations must be broughtonline at the same high QoS as the existing sites.

    Like so many tasks within the modern communications

    network, installation procedures must be carried out quickly

    and at minimum cost. When a new base station site is

    chosen and acquired, it is in the networks best interest to

    put in the BTS hardware as soon as possible. A new site

    means more capacity or better coverage, which are key

    competitive advantages.

    Naturally this time and cost pressure is in conflict with the

    need for the thorough verification measurements and

    compliance tests that support QoS. But that is a manageable

    conflict if the network equips its installers with the tools to

    do the job efficiently. That means instrumentation thatensures that the technicians can go to the BTS site, make

    consistent measurements without a lot of costly setup time,

    and quickly identify problems if they should arise.

    Measurement Challenges and Solutions

    Base station installation involves a variety of different tasks.

    Frequently, the network operator and the equipment

    vendor work closely together for the deployment period.

    But ultimately the responsibility is in the hands of the

    network personnel.

    Among the tasks are:

    Clearing the transmit and receive bands of existing usersIn some regions (especially in the case of new PCS networks

    in North America), the operator must find and relocate exist-ing private microwave systems already using the frequencies

    they need. This is a substantial cost that rests entirely on the

    shoulders of the network operator.

    Baseline performance checksAs sites are selected and base

    stations installed, technicians perform a baseline perfor-

    mance check on the BTS. This procedure includes antennaline performance measurements, interference tests, and a

    wealth of transmitter performance tests.

    Final offline compliance testsInstallation time is a goldenopportunity to perform full compliance tests on a base

    station operating in a network setting, but not carryingtraffic. It is the last chance to verify these characteristicswithout impacting subscribers use of the network.

    The other kind of installation, adding GPRS and other new

    features to existing base stations, is usually a matter of

    loading new software into the BTS, BSC (Base Station

    Controller), and MSC (Mobile Switching Center) system

    and verifying that the base stations are responding correctly.

    While this is less of a hands-on job than hardware

    installation, its no less important to do a thorough job in

    minimal time.

    One of the challenges in performing all these tests and

    measurements is doing the job cost-effectively. Field techni-

    cians obviously cant carry a lab full of equipment to every

    BTS site for an installation job. Installation needs have

    spawned a generation of compact, high-performance test

    and monitoring instruments, the best of which offer ease of

    use, accuracy, integration of multiple measurement functions,

    and automation. This makes it possible to affordably equip

    field teams with the flexible measurement capability they

    need to support QoS goals at installation time.

    7I N S T A L L A T I O N

    Installation

  • 7/27/2019 Mobile QoS Primer

    8/33

  • 7/27/2019 Mobile QoS Primer

    9/33

    Looking at Modulation Accuracy

    There is a second category of base station measurements

    that requires the facilities of a more full-featured spectrum

    analyzer. These are tests such as Error Vector Magnitude

    orin the case of CDMA systemsRho (waveform quality),

    that are designed to determine the modulation accuracy of

    the transmitter. To carry out such tests, as well as the all-

    important code domain measurements for CDMA and

    W-CDMA systems, it is necessary to use an instrument that

    includes modulation analysis capability.

    In recent years, optimized instruments containing both

    a spectrum analyzer and a modulation analyzer have

    emerged to address the most critical base station installa-

    tion needs. These tools can of course perform the essential

    spectrum measurements described earlier, and they can

    also provide a detailed transmitter analysis that suffices

    for initial compliance verification.

    Here again, the combination of two major test functionali-

    tiesspectrum and modulation analysisin one instrument

    gives the field installation technician greater convenience

    and mobility at lower cost, and requires training on only

    one instrument rather than two. And with two tools in one,

    the technician is more likely to have the right tool on hand

    to solve problems quickly.

    Imagine a technician who installs a BTS and makes the

    necessary RF spectrum measurements. Having determined

    that everything is within specifications to this point, he or

    she must still verify transmitter parameters such as errorvector magnitude, phase error, magnitude error, GSM ramp

    up/down power, and more. These can be complex measure-

    ments, and it is easy to imagine the technician wishing for

    some automated measurement and analysis capability!

    Fortunately the state of the art in spectrum/modulation

    instruments has advanced to the point where just such capa-

    bilities are available. Figure 6 shows a single measurement

    on a GSM transmitter. In this case, the transmitter happens

    to be frequency hopping. The instrument automatically

    completes a full range of measurements on the frequency

    agile signal from the transmitter. This includes spectrum

    performance, EVM (Error Vector Magnitude), and phase andfrequency errors to help diagnose any transmitter anomalies.

    Clearly, this level of automation allows technicians and

    engineers to quickly investigate service-affecting problems

    with a minimum of training.

    Some spectrum/modulation instruments take the process

    a step further, automating the full suite of conformance tests

    on BTS transmitters. An internal program automates every

    common measurement, with an on-screen display that

    lets the user select the tests by means of simple, on/off

    keystrokes. An example is shown in Figure 7. Such programs

    offer the added benefit of enhancing repeatability among

    widespread field teams.

    9

    If a field installation

    team can be equipped

    with an instrument

    versatile enough to

    bridge several test

    needs, then the team

    is freed from learning,

    carrying, setting up

    and maintaining a

    set of discrete single-

    purpose tools.

    I N S T A L L A T I O N

    Figure 6: A sample measurement on a hopping GSM signal.The cursor in each window represents the same s ignal sample.

    Figure 7: Automated Test Control Screen

  • 7/27/2019 Mobile QoS Primer

    10/33

    Protocol Analysis Supports

    Installation of New Features

    Currently, one of the most urgent network needs is to

    add capability and capacity for data transactions over the

    mobile network. Subscribers want to send and receiveinternet content and other data on their mobile phones.

    In addition, a new generation of mobile data appliances is

    emerging. It all adds up to intense pressure to add data

    services to mobile networks. A clear competitive advantage

    goes to network providers who come to the market first

    with reliable data services.

    Pending acceptance of new 3G protocol specifications such

    as UMTS Wideband-CDMA, many networks are installing a

    transitional protocol known as General Packet Radio system

    (GPRS). The new protocol uses existing BTS and other

    network elements as much as possible. To the Installation

    department, installing GRPS is a matter of loading new

    application software into certain network elements, running

    a series of tests, and documenting the result. As mentioned

    earlier, there is an entirely separate class of instruments

    known as protocol analyzers that supports protocol testing

    and verification during the installation of software-based

    network features and to a lesser extent, new BTS equipment.

    The protocol analyzer is another tool that has made great

    strides in the past few years. Todays leading field-portable

    analysis solutions can, under user-programmed control,

    address a wealth of protocol types, monitor several test

    points at once, and present a coherent, human-comprehen-

    sible picture of network activity in the protocol domain.

    The protocol analyzer is the cornerstone of QoS measure-

    ments during GPRS installation. Because GPRS installation

    is such a key issue at this time, we will use it to illustrate

    the broad range of tasks the protocol analyzer can perform

    (note that other tools such as call generators may be

    needed as well).

    As explained elsewhere in this document, GPRS is a

    packet-switched technology added in parallel to the existing

    networks circuit-switched architecture.

    Consequently there are four areas that must be verified

    at the time of GPRS installation:

    Physical layerWhile this is not explicitly a protocol test,physical layer access is essential because layer 2, which

    must be tested, uses the physical layer.

    Circuit-switched testingThis step includes viewing thesignaling protocols, and checking the switched circuits.

    A comprehensive procedure might require call generation,

    speech quality checks, and more. Other tests involve cover-age analysis with a call generator system, and testing of fax

    and data functionality.

    Packet-switched testingThis step involves testing of thenetwork layers up to and including layer 3 in the ISO OSI

    model. These network layers are handled by network nodesor network entities. The layers above are normally handled by

    user equipment such as workstations and mobile stations.

    In addition, packet switched testing can involve testing withreal applications at the user level. This might include check-

    ing functionality with a Web browser using the HTTP

    protocol, or checking file transfer using file-transfer protocol(FTP). It also applies to some mobile applications using the

    newly-developed Wireless Application Protocol (WAP).

    Inter-operationThe operation of circuit-switched andpacket-switched networks as they interact must be checked.

    Looking at the list above, it is clear that a simple GPRS

    retrofit procedure can be every bit as challenging as a full

    BTS install.

    From a specific protocol monitoring standpoint, the

    measurement tool must first and foremost be capable

    of providing a view of all the interfaces involved in any

    transaction of interest, as shown in Figure 8. The instrument

    can provide appropriate statistics to summarize the activity.

    Other measurements might be Mean Packet Delay and

    Mean Packet Size GPRS attach, transmit, and detach;

    PDP context activation, transmission, deactivation, and

    billing radius.

    Packet testing is also comprehensive: TCP timing problems;

    packet generation and verification; emulation of all lower

    layers up to IP on Abis, Gb, Gn, Gi; generation and checking

    of user layers, and more.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Figure 8: Protocol Analysis on The Gb Interface

  • 7/27/2019 Mobile QoS Primer

    11/33

    A third category of tests, conformance tests, is challenging in

    a different way. Conformance test suites normally consist of

    hundreds of test cases grouped together to test the differentstates of a protocol. By including a parameter list adapted to

    the specific implementation, one entire conformance test

    suite can be run automatically, without human intervention.

    Unlike other areas of protocol testing, there is no concrete

    standard specification for conformance tests available at the

    moment. Therefore, the common practice is to use confor-

    mance tests developed cooperatively among providers of

    GPRS functions. At this time certain conformance tests, i.e.

    ATS and ETS, are available for the BSS site and SGSN site,

    the NS, and the BSSGP layer. This is enough to ensure that

    the connection between SGSN and the BSS equipment

    from diverse manufacturers is working.

    Figure 9 depicts a part of the results of a typical conformance

    test cycle. It shows the interaction between the system under

    test (SUT) and the protocol analyzer with verdicts indicating

    whether a specific test case is successfully completed or not.

    Protocol Analysis Solution Requirements

    Selecting the right protocol analyzer for GPRS installation

    is the key to doing the installation efficiently. Moreover,

    the protocol analyzer enables the Installation department

    to meet the networks QoS goals by ensuring a clean,

    thoroughly-verified installation of the GPRS features.

    To do this the analyzer must provide all of the physical

    interfaces and layer 2 variants, and all protocols for GSM

    and GPRS. It should also be able to monitor and simulate

    multiple interfaces in parallel to truly capture the systems

    complex behavior.

    In addition to normal monitoring simulation, there is a

    requirement for emulation and packet generation and

    checking. A standard interface to user applications with IP

    data is also helpful. Lastly, the analyzer must offer a means

    of conformance testing.

    Conclusion

    Although a network Installation department is not expressly

    chartered with the responsibility of QoS, it nevertheless plays

    a key role in supporting the networks overall QoS goals.

    As the group tasked with installing new equipment and

    services, it falls upon the Installation department to make

    sure that new base stations and services like GPRS join the

    network in top operating condition. This is the foundation

    of a continuing pattern of reliability that keeps network

    quality at competitively high levels.

    Given these expectations, the Installation group relies on

    measurement tools that can assist the install process with

    automated features and the versatility to handle a broad

    range of installation measurements. Among these tools areRF spectrum analyzers and spectrum/modulation analyzers

    that help bring up the RF aspects of a newly-installed BTS.

    Another Installation department discipline is the addition of

    new capabilities to existing base stations and network

    elements. The best example is GPRS, a software-based

    feature that requires thorough verification of protocol

    behavior. Here, the Installation group calls on the versatile

    protocol analyzer to capture and display transactions across

    all layers of the protocol, as well as tests of the circuit-and

    packet-switched network elements.

    11I N S T A L L A T I O N

    Figure 9: Conformance Test Results

    Conformance test

    suites normally consist

    of hundreds of test

    cases grouped together

    to test the different

    states of a protocol. By

    including a parameter

    list adapted to the

    specific implementa-

    tion, one entire

    conformance test

    suite can be run

    automatically, without

    human intervention.

  • 7/27/2019 Mobile QoS Primer

    12/33

    Operations & Maintenance Overview

    In many respects, the Operations & Maintenance (O&M)

    organization is the heart of the mobile network enterprise.

    It is staffed by hands on technical personnel who have

    the expertise to install, troubleshoot, monitor, and maintain

    complex network elements such as switching and SS7

    signaling equipment. O&M work goes on at far-flung field

    sites as well as within the plant.

    The O&M organization is responsible for running the

    network day-in and day-out. O&M keeps the network

    up-to-date and up to the subscribers expectations for

    features, accessibility, and quality, as defined by Planning

    and Engineering. O&M is responsible not for envisioning

    the network of tomorrow, but for making todays network

    a practical reality. The technical team faces the challenge

    of debugging new equipment under intense time pressure.

    It is the O&M activitys job to understand the network as a

    whole; to evaluate network performance from a global

    standpoint. Inevitably, some elements of the network will

    be weaker and more error-prone than others. These the

    O&M group must characterize and correct where possible.

    In addition, the group must carry out the mechanics of the

    network operators competitive QoS programs, that is, they

    must implement the hardware and software procedures

    that maintain predefined network quality levels.

    We said it earlier: to stand still is to fall behind. The network

    O&M organization plays a key role in helping the network

    move ahead in meeting subscriber needs.

    QoS Considerations for the O&M Department

    Network O&M personnel face a paradox in executing

    their maintenance duties. Subscribers want more coverage

    and a growing variety of services. Yesterdays voice-only

    subscriber wants paging or data services today. The only

    way to deliver these improvements is to install new

    equipment and software.

    Yet these same subscribers also insist on uninterrupted

    high-quality connections. Unfortunately the need to

    constantly upgrade and enhance services can easily

    degrade mobile traffic in some way, unless precautions

    are taken. The impact may range from noise or distortion

    to outright discontinuities.

    Notwithstanding this challenge, the O&M organization also

    has the responsibility to measure network QoS on a day-to-

    day basis, and to respond to problems that have prompted

    subscriber complaints. Note here that O&M must address

    not only the networks provable quality problems, but also

    the subscribers perceptions of quality.

    Some examples of subscriber complaints are:

    Interrupted calls

    Calls requiring more than one dialing attempt

    before connection

    Inaccessible special services such as 800 numbersand the Short Message Service Center

    And of course, roaming limitations, noise, and generally

    poor audible voice quality

    As always (in the network business) O&Ms job must be

    done cost-effectively. In this context, the term cost-effective

    has implications about the cost of tooling up for the job,

    the cost of hiring and training skilled technical personnel,

    and even the cost of maintaining spares and software

    updates for the measurement and monitoring equipment!

    Measurement Challenges and SolutionsIn keeping with the need to minimize network traffic

    interruptions, O&M work is carried out on live network

    elements with non-intrusive measurements wherever

    possible. Typically this requires tools specifically designed

    for passive measurements, that is, tools that connect to

    the network but do not divert, delay, or degrade the

    signals therein.

    These instruments, which range in scale from fully

    integrated SS7 signaling monitoring systems to small

    portable protocol testers, observe key points of interest

    throughout the network. Integrated systems provide a

    global view of the network, while portable instruments

    help technicians spot more localized problems.

    Integrated SS7 signaling monitoring systems offer the

    advantage of continuous monitoring across many network

    test pointspotentially thousands of them. Because these

    systems view SS7 signaling activity rather than in-band

    events, they dont interfere with active network transactions,

    yet they can track all aspects of network activity down to

    the individual call level. Typically these SS7 monitoring

    systems are scalable to meet the needs of diverse network

    QoS strategies. They enable automated monitoring on a

    network-wide scale.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Operations &Maintenance

  • 7/27/2019 Mobile QoS Primer

    13/33

    Portable Testers Provide Cost-effective

    Protocol Monitoring

    The portable protocol tester is a valued QoS tool. Although

    the protocol tester is designed for on-the-spot troubleshooting

    and analysis, some of the more advanced protocol testerscan observe several interfaces simultaneously. This allows

    the instrument to evaluate the interaction of multiple

    network elements. This scheme is shown in Figure 10.

    Here, the protocol tester is observing four test points (the

    connections are shown encircling the line to symbolize

    the non-intrusive nature of the measurement), including

    the ultimate connection to the Public Switched Telephone

    Network (PSTN). By this it means it is possible to view the

    calls as they traverse all of the network elements following

    the base station (BTS). Test setups like these are useful for

    quickly assessing global network performance, comparing

    the various elements, and locating problem areas.

    In the interest of long-term cost-effectiveness, the protocol

    tester should be able to evolve and grow with the

    networks measurement needs. It should be designed

    around a powerful, flexible, and extensible operating

    system such as Windows NT. Its application software must

    adapt to many different network interfaces (far more than

    the four shown in Figure 10). Moreover, the instrument

    must offer hardware configurability and expandability. If it

    is to be used afield, it should be easily transportable and

    controllable via remote connection.

    Figure 11 depicts the results of a test (viewing just one

    test point) using a protocol analyzer as illustrated above.

    It shows a detailed analysis of messages taken between

    a BTS and a BSC (the Abis interface). The upper part shows

    the message flow itself (one message per line is shown),

    along with the most important parameters. These can

    be selected by the user. In this example only the IMSI

    is selected as a special parameter. The lower part of the

    window decodes every bit of the selected messages

    and helps pinpoint wrong values and settings.

    13

    Here, the protocol

    tester is observing

    four test points (the

    connections are

    shown encircling

    the line to symbolize

    the non-intrusive

    nature of the

    measurement),

    including the

    ultimate connection

    to the Public

    Switched Telephone

    Network (PSTN).

    O P E R A T I O N S & M A I N T E N A N C E

    HLR

    BTS BSC MSC

    PSTN

    Protoc0l Tester

    Figure 10: Network Monitoring With a Portable Protocol Analyzer

    Figure 11: Protocol Analyzer Results Screen

    The scenario below assumes a straightforward network

    connection made up of one BTS connected to one BSC

    connected to one MSC, and so on. When it is possible toisolate ones troubleshooting or monitoring efforts to this

    level, the protocol analyzer is in its comfort zone. But

    what about more complex arrangements?

    One answer is to use a switch matrix that connects the

    protocol analyzers basic complement of channels to many

    more network monitoring points. Through the switch, the

    analyzers resources can be shared among many groups

    of test points with no compromise in performance. By this

    means it is possible to monitor different network segments

    at different times of day, a feature that simplifies, for

    example, the analysis of changing traffic patterns during

    the business day.

    Another approach is to use an integrated, network-wide

    SS7 monitoring system, as explained later.

  • 7/27/2019 Mobile QoS Primer

    14/33

    New Features Mean Added Responsibility

    for Protocol Testing Tools

    One network interface that is of particular interest to

    network operators today is GPRSthe General Packet Radio

    System. GPRS is a transitional technology that enables

    current-generation GSM networks to better support data

    transmission and reception in addition to the normal voice

    activity. As such, GPRS is gaining wide acceptance and

    creating new test challenges for the O&M group.

    Existing GSM networks are circuit-switched structures with

    a packet-switched component for signaling. GPRS is a

    purely packet-switched network added in parallel to the

    GSM system. Some of the critical shared resources between

    the two are the base station system (especially the BTS),

    the visitor and home location registers, and the equipment

    identity register. GPRS base station systems have an addi-

    tional packet control unit (PCU) plus Serving GPRS Support

    Nodes (SGSN) and Gateway GPRS Support Nodes (GGSN).

    The GPRS system is normally connected to packet data

    networks like the Internet, X.25 network, or other GPRS

    networks, providing the needed data connections for

    mobile service.

    SS7 Signaling Network is the Connection

    for Network-wide Monitoring

    Consider a typical modern digital mobile network. It probably

    grew up gradually over the past 5-8 years, during which

    time the network operator installed several successive

    generationsoften from diverse manufacturersof BTS, BSC,

    and MSC equipment. Certainly this made business sense

    at the time, since each new round of improvements gave

    the networks more functionality at lower cost. But now the

    O&M organization is confronted with testing and trouble-

    shooting all these various makes and models, and theirinteractions as a network. It is a daunting responsibility.

    The key to meeting this challenge lies in one fundamental

    element of todays network architecture. Underlying the

    networks in-band functions is a supporting network of SS7

    signaling functions. The scale of the SS7 apparatus is all-

    encompassing; it is involved in every transaction; and it has

    a profound effect on subscriber-perceived Quality of Service

    issues such as dropped calls.

    Thus SS7 network measurements can provide a microscopic

    view into how subscribers use the network, what services

    they use, how often, and what problems they encounter.

    This kind of monitoring problem cries out for a solution that

    can be everywhere at once. No single fixed test site will

    provide all the information that is needed to monitor SS7

    network quality. Once it is understood that there must be

    many test points, where should they be? Ideally the moni-

    toring tool should be deployed in every link of the mobile

    network, although this is not really feasible for reasons of

    cost and complexity.

    Increasingly, O&M organizations are turning toward the

    integrated SS7 signaling monitoring system as a means of

    supporting QoS efforts. With its scalability, its widespreadprobing points, and its ability to capture network activity down

    to the individual call level, the SS7 monitoring system is the

    most effective permanent QoS monitoring solution available.

    Probably the most strategic SS7 test points are located at

    the Mobile Switching Centers (MSC). At these locations,

    there is a concentration of links to Base Station Controllers

    (BSC), networks (PSTN as well as other PLMN mobile

    networks), network databases (specifically the Home

    Location Register or HLR), and other MSCs. With monitoring

    instrumentation permanently based at MSC sites, a broad

    and comprehensive view of SS7 network activity is available

    for collection by a centralized system element.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    OtherPLMN

    SMS-GMSCSMS-IWMSC

    SM-SC

    GGSN

    PDN TEGGSN

    MSC/VLR

    E

    Gs

    CGd

    Gp

    D

    Gr

    GiGn

    Gf

    Gb

    Gn

    UmR

    HLR

    GSM

    Signalling Interface

    Signalling and DataTransfer Interface

    GRPS

    Other

    TE MT BSS

    GcA

    EIR

    SGSN

    SGSN

    Figure 12: GPRS Network Interfaces

  • 7/27/2019 Mobile QoS Primer

    15/33

    As stated earlier, the SS7 signaling structure is involved

    in every single network transaction. That makes it difficult

    to interrupt or intrude upon SS7 activity for the purpose

    of monitoring and troubleshooting. Intrusive monitoring

    techniques are inherently limited to connections at theperiphery of the network; they use controlled test calls

    between two test sets at opposite ends of the communica-

    tion path to characterize quality.

    A better approach is to use non-intrusive monitoring, in

    which instruments are woven integrally into the network.

    Note that this isnt the same as the built-in QoS monitoring

    features included in certain types of network equipment

    (such as MSCs). The performance of these built-in tools

    depends to some degree on the performance of the very

    same network they are observing. In contrast, integrated

    non-intrusive QoS tools are independent of these

    environmental problems.

    The combined need for integral, non-intrusive monitoring,

    a multiplicity of test points at various MSCs throughout the

    network, and a central unit connected via a WAN says one

    word to an experienced observer: system. Clearly the only

    acceptable solution for a measurement/monitoring problem

    of this scale is a carefully planned system made up of

    compatible elements designed to work as one. This is not

    the place for off-the-shelf instrument clusters or for small

    general-purpose tools.

    Figure 13 shows a non-intrusive QoS monitoring system

    integrated into a network environment as described above.Here, the QoS probes, which reside at the MSC installa-

    tions, connect through a Wide Area Network (WAN) using

    a TCP/IP connection. This is in turn connected to a Central

    Unit that is responsible for system-wide data collection,

    correlation and storage. In addition, the Central Unit is

    the focal point for management and control of the entire

    monitoring process.

    A network-wide non-intrusive measurement system can

    provide a wealth of critical data about network quality and

    traffic. The Call Detail Record (CDR) is a key deliverable

    of the network analysis system. Todays state-of-the-art

    network analysis systems can generate Call Detail Recordsfor every call or call attempt and store all this information

    in a relational database. An example of a Call Detail Record

    report is shown in Figure 14.

    Each row in the report represents a single CDR. The infor-

    mation can be filtered and displayed based on the values

    or ranges in one or more fields. For example, a filter might

    be used to view only the calls originating from one specificmobile subscriber (identified by the calling mobile number)

    in order to track a suspected quality problem with that

    subscribers connection.

    A CDR contains information about a specific call or

    attempt. This data ranges from calling and called telephone

    numbers, to duration (holding time, conversation time)

    to the outcome of the call (whether the call was successful

    and, if not, the reason for call failure).

    15

    The QoS probes,

    which reside at the

    MSC installations,

    connect through a

    Wide Area Network

    (WAN) using a TCP/IP

    connection. This is in

    turn connected to a

    Central Unit that is

    responsible for

    system-wide data

    collection, correlation

    and storage.

    O P E R A T I O N S & M A I N T E N A N C E

    Figure 14: Call Detail Record Report

    HLR

    To Central Unit

    PLMN

    PSTN Y

    PSTN X

    BSC

    MSC

    MSC

    WAN

    BSC

    MSC

    Figure 13: SS7 Monitoring Probes Acquire Signaling Data Via the MSC

  • 7/27/2019 Mobile QoS Primer

    16/33

    All this information provides valuable insight into

    subscribers behavior, leading to important inferences about

    the subscribers perception of service quality. For example,

    a very low average conversation time points toward poor

    transmission quality, which often causes users to abandontheir calls prematurely.

    Lets look at an example scenario demonstrating how an

    SS7 monitoring system can resolve a subscriber complaint

    quickly and easily.

    Suppose a subscriber complains about a call completion

    problem. He is having difficulty reaching his friends

    across town.

    Because the subscribers network provider has wisely

    chosen to install a comprehensive SS7 signaling monitoring

    system, the data needed to investigate the problem is

    already in the CDR database. Thats convenient! The O&Mgroup need only browse through the particular subscribers

    CDRs to spot trends like congestion or radio interface

    failures. The powerful search features of the relational

    database can accelerate this process. In addition, the

    operator can set up a user call trace to get a close-up view

    of protocol-related issues that may be causing the problem.

    Given all this detailed information, plus the map-based

    network views that the system provides, the problem is

    quickly localized. The intervention can be fast, focused,

    and cost-effective.

    The network analysis system can also deliver reports suchas the E.422 Report, an ITU-T standardized report for quality

    assessment at the interface between network operators,

    whether the networks are local, national, international,

    fixed, or mobile. The E.422 Report makes it easy to resolve

    any disputes between network providers who have

    reciprocal agreements to put through calls for one another.

    An example of the E.422 form produced by an advanced

    network monitoring system is show in Figure 15.

    The E.422 report takes information from the Call Detail

    Records and provides an assessment of the Quality of

    Service that calling subscribers obtain. Every call attempt is

    classified based on the outcome of the call. In particular,the E.422 report details the calls successfully put through,

    and specifies, for the unsuccessful calls, how many are

    due to the customer behavior and how many are due to

    the network. Thus it is possible to determine whether call

    failures are due to interworking problems or to protocol

    failures among calls to a specific destination.

    Lastly, the SS7 monitoring system makes available a host of

    day-in, day-out statistical views such as the Gauge display

    in Figure 16. The gauge display provides a continuous

    near-real-time view of the health of the network. Significant

    statistical counters are regularly updated and ranked inorder to provide worst-case values of measurements all

    across the network. This and other SS7 monitoring system

    displays are designed to simplify the comprehension of

    complex measurement data, ensuring efficient, error-free

    interpretation of the networks behavior. By this means, it

    is possible to focus and prioritize network interventions to

    resolve critical performance issues that can impact Quality

    of Service.

    Call Generator/Analyzer

    is a Demanding Subscriber

    Of course, the O&M groups tasks are not limited to simply

    viewing network activity and reacting whenever there is a

    problem. A comprehensive QoS strategy usually includes

    active stress testing of network elements in simulated

    real-world situations. Significantly, the most realistic stress

    situations are not the normally initiated and terminated

    calls; rather, they are abnormal circumstances that occur

    all too often in networks.

    Examples include:

    Ping-pong handoffswherein a call bounces back andforth erratically between two adjacent cells

    Overloadingwhich can cause delays in voice mail prompts,

    in turn causing callers to disconnect before they are awareof being connected

    To create these and other aberrant network situations,

    a test system that can actually generate mobile traffic is

    needed. The system must interact with the mobile network

    through the air interface, and with fixed networks through

    common PSTN or ISDN lines. It must not only generate

    traffic; it must also acquire and analyze the resulting

    network behavior in order to perform true end-to-end tests.

    To create aberrant

    network situations, a

    test system that can

    actually generate

    mobile traffic is

    needed. It must not

    only generate traffic;

    it must also acquire

    and analyze the

    resulting network

    behavior in order

    to perform true

    end-to-end tests.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Figure 15: E.422 Report

  • 7/27/2019 Mobile QoS Primer

    17/33

    This call generator/analyzer can be installed as part of the

    network itself, managed under the auspices of the O&Morganization. Whats important to note here is that the

    system is a surrogate for all the different elements in

    the networkBTS, BSC, and MSC alike. The instrument

    emulates the behavior of mobile subscribers who use any

    and all of the available network services. Among these are

    speech and data, Short Message Services, supplementary

    services such as call waiting and call barring, and voice

    mail. The system also can produce heavy traffic, a matter

    of great interest to the O&M organization. Lastly, the most

    advanced mobile call generation systems can perform

    a host of RF measurements and reporting.

    To be truly effective, such a system must be easy to use.

    How does the O&M engineer design calls that are both

    realistic and thorough? Some systems have graphical

    Call Modeling tools that simplify the process. With these

    instruments, the user selects from a library of pre-writtenbuilding blocks that include such terms as Cell Select, Wait

    on Hook, Generate DTMF, and many others, as shown in

    Figure 17. The blocks are concatenated to form a full call.

    Calls can be combined to build up an exhaustive test that

    challenges the network units capacity and services.

    Conclusion

    Much of the responsibility for a mobile networks quality

    of service rests on the shoulders of the Operations &

    Maintenance organization. In turn, the O&M organization

    places its trust in expert personnel using the latest test,measurement and monitoring equipment.

    In this chapter we have seen how the protocol analyzer can

    be used to observe and analyze the quality of live network

    traffic without disturbing the networks most valuable

    asset, its subscribers. We have also discussed the scalable,

    network-wide capabilities of the SS7 signaling monitoring

    system. In addition, we have discussed the mobile call

    generation system, which can be used to troubleshoot or

    certify myriad mobile network elements under worst-case

    call and traffic conditions. This is a reliable way to ensure

    (again, without interrupting subscriber traffic) that newly-

    installed or upgraded BTS, BSC, and MSC units will deliverexcellent Quality of Service.

    17O P E R A T I O N S & M A I N T E N A N C E

    Figure 16: Gauge Display

    Figure 17: Call Generator/Analyzer Setup Screen

  • 7/27/2019 Mobile QoS Primer

    18/33

    Planning & Engineering Overview

    In the Planning & Engineering (P&E) organization, the

    networks future takes shape. Here, nothing is more constant

    than change itself. Operators pressured by competition must

    steadily introduce new services and enhance the old ones.

    P&E personnel work hard to provide new network features

    in a timely and cost-effective manner. Its a process that

    involves some new design work, some upgrading of existing

    infrastructure elements, and inevitably some concessions

    to the limitations of older installed equipment. Networks

    become ever more complex as the mixture of old and new

    equipment, varying brands, and diverse software revision

    levels begins to grow.

    New and improved services are essential for competitive

    reasons, but bigger is also better. That is, successful networks

    are constantly expanding their capacityadding more base

    stations, more mobile switching centers, and so on.

    From the viewpoint of an engineer working in the P&E orga-

    nization, this creates an environment thats always growing

    and is in migration from one technology to another, or

    transitional. Frankly its not a stable platform to stand on,

    yet the P&E engineer must somehow ensure that all of the

    variables come together in a network system that works

    reliably and smoothly. Moreover, it is the responsibility of

    the P&E organization to deliver network elements (both

    hardware and software) that are supportable by Operations

    & Maintenance personnel without requiring lengthy (and

    costly) new training cycles.

    Functionally, P&E is charged with:

    Planning of the network topology

    Evaluating and selecting network elements (BTS, MSC, andso on) that will provide the best performance and QoS at

    the least cost

    Increasing the networks capacity

    Planning and designing the SS7 signaling network

    Helping to resolve network problems such as coverage gaps

    and dropped calls

    Operating and managing the test plant

    QoS Considerations for the P&E Department

    In a sense the networks ability to compete in its market

    begins with the P&E group. After all, this is the organization

    that determines what equipment will be sited where.

    Therefore QoS is a mission that stands or falls on the

    choices made in Planning & Engineering.

    Choosing a base station site is among the first big decisions

    that must be made when adding network coverage or

    capacity. Planning for base station siting is commonly done

    with special software planning tools. These applications

    weigh volumes of geographical and morphology data to

    predict the coverage that a given BTS site will achieve.

    Even with these powerful predictive tools, QoS issues can

    make site selection a complex equation. For example, the

    most accessible real estate for the base station might be

    compromised by conflicting signals from other providers,

    which would very likely impact QoS. The site with the bestelevation might be hampered by proximity to tall build-

    ingsa one-way ticket to fading and interference. The reality

    is that predicted base station coverage and actual coverage

    can differ significantly. And gap between prediction and

    reality is expected to grow even wider as new high-speed

    data services are introduced. Therefore a thorough evalua-

    tion of the site environment, both before and after BTS

    installation, is a necessity.

    Measurement Challenges and Solutions

    After sites have been selected and base stations installed,the P&E groups work still is not finished. P&E is responsible

    for an ongoing process known as network tuning and

    optimization. This involves continued monitoring of site

    and network performance long after the initial turn-up.

    Why is this necessary? There are a host of reasons: new

    structures or competitors base stations may be erected

    near a BTS, causing interference; an antenna may get

    damaged by wind or snow; transmission lines can become

    impaired by moisture ingress or mechanical damage. Any

    and all of these problems will affect QoS in both measured

    and perceived terms.

    Drive Tests Provide an End-to-End View of the Network

    The cornerstone of network optimization is the drive test.

    The drive test is exactly what its name implies: a series of

    quality measurements conducted from the vantage point of

    a moving vehicle. This job is the province of an integrated

    instrument designed to take these moving performance

    snapshots. As always, there is a variety of solutions in this

    instrument category. Until recently, most of these have

    offered rather basic measurement capabilities: Rx (receive)

    quality and level measurements, certain call statistics, and

    downlink information correlated to the receivers position.

    Increasing utilization and the presence of competingnetworks in almost every region of the industrialized world

    is changing the nature of the drive test. As complex

    network services become more common, and as QoS

    concerns (and therefore, subscriber churning) continue to

    mount, older 1st-Generation drive test methods simply

    cant provide enough data. They dont take readings on

    both the downlink (in the direction of the mobile

    subscriber) and the uplink (in the direction of PSTN

    subscribers and other mobiles) simultaneously. Typically

    they lack capability to monitor more than one network at a

    time, even though there may be several networks operating

    in the same area. Lastly, P&E needs extensive details about

    network performance as it relates to the callers physical

    location. That means more information recorded and corre-

    lated accurately to geographic coordinates.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    In a sense the

    networks ability

    to compete in its

    market begins with

    the P&E group.

    After all, this is the

    organization that

    determines what

    equipment will be

    sited where. Therefore

    QoS is a mission that

    stands or falls on the

    choices made in P&E.

    Planning &Engineering

  • 7/27/2019 Mobile QoS Primer

    19/33

    The solution of choice for comprehensive drive testing is a

    tool optimized for end-to-end testing of the network. This

    type of analyzer offers a wealth of important features that

    help the drive testing process deliver results more quickly

    and economically, with less margin for error.

    Among these characteristics are:

    Automatic generation of wireless-to-wireless, wireless-to-POTS,

    and POTS-to-POTS calls, including voice, fax, and data calls

    Automatic generation and receipt of SMS messages

    Configuration and automatic execution of complex

    test sessions

    Analysis of the information exchanged with protocollayers 1 and 2 of the mobile network.

    Uplink and downlink transmission quality tests

    Simultaneous management of multiple mobilenetwork interfaces

    And more

    A typical end-to-end analyzer system consists of two units

    a fixed base station unit that connects to the PSTN, and a

    mobile station with up to four links, as shown in Figure 18.

    These components can automatically call one another and

    exchange voice and data content. The analyzer takes

    measurements on the call process during these calls,

    including transmission quality measurements in the voice

    band. The test system also acquires the data obtained by

    decoding the messages exchanged between the mobile

    unit and the mobile network. Once the drive test is

    concluded, the mobile unit can transfer its stored results

    via the mobile network connection it has just finished

    testingto the fixed unit, which gathers all the relevantinformation in one place for analysis.

    Looking at Figure 18, note that the analyzer system uses the

    signal from the Global Positioning System (GPS) to track

    the exact position, moment by moment, of the mobile

    station. This information, too, is stored in the instrument,

    tightly correlated with the measurements occurring at the

    same instant.

    19

    P&E is responsible

    for an ongoing

    process known as

    network tuning and

    optimization. This

    involves continued

    monitoring of site

    and network perfor-

    mance long after

    the initial turn-up.

    P L A N N I N G & E N G I N E E R I N G

    Figure 18: End-to-End Analyzer Setup

    PSTNGPS

    NetworkY

    Mobile Stations

    BTS

    Drive TestSystem-stationaryunit

    BTS BTS

    NetworkX

  • 7/27/2019 Mobile QoS Primer

    20/33

    From the standpoint of the P&E organization, the next step

    provides the true benefit of using an advanced analysis

    system. The system has, built in, a powerful software appli-

    cation that displays an actual map of the drive test and its

    results. Figure 19 is an example of the systems output.

    With all the foregoing talk about measurements, maps, and

    stored results, weve begun to see exactly what the SS7

    monitoring system can do to characterize network coverage,

    as well as quality in general terms. But what about its

    capacity to objectivelymeasure the Quality of Service?

    Its essential to take the opinionating out of the discussion

    about Quality of Service. Anybody can place a call and

    describe its quality as poor or good. But until you can

    support words like poor, acceptable, and excellent

    with actual measurements, you simply cant build a QoS

    program. Quantification gives you a starting point and a

    way to track the success of improvements.

    Quantification begins with accurate measurement

    of the following key parameters:

    AccessibilityThe ratio of successful calls to those

    terminated abnormally or routed incorrectly.

    ConformityThe extent to which a call provides aclean, clear, uninterrupted connection. The conformity

    measurement produces a Mean Opinion Score basedon a hypothetical subscriber profile, see Figure 20.

    ContinuityThe percentage of calls that fail during a

    simulated conversation between the mobile and fixedunits of the monitoring tool.

    As we have said elsewhere in this document, virtually

    all networks are made up of elements both new and old,

    from diverse equipment manufacturers. Even though most

    vendors of such equipment strive to test their new products

    thoroughly, it is not feasible to test for the innumerable

    combinations of base stations, MSC, base station controllers,

    and so on. Consequently the network operator faces the

    risk of unpredictable interactions and malfunctions once

    the equipment is installed.

    Test Plant Debugs Problems in a Neutral Environment

    The accepted solution for this problem is the test plant,

    a self-contained miniature network designed for the

    purpose of testing and debugging network elements. The

    Planning & Engineering group is normally the organization

    responsible for the test plant. The test plant operates under

    controlled conditions and is a focal point of measurements

    that help predict the QoS impact that new network

    elements might have. The test and verification proceduresuse many of the same measurement tools deployed in the

    actual operating network.

    In the test plant, the goal is to find one tool that represents

    the widest range of equipment in the field, including

    elements that the future may bring. Increasingly the

    solution is a call generation system that interfaces to the

    network through the normal air interface and with the fixed

    network through conventional PSTN or ISDN lines.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Figure 19: The Drive Test System Provides A Map View

    Figure 20: Conformity Measurement Result

  • 7/27/2019 Mobile QoS Primer

    21/33

    The RF connection is usually made through coaxial cables

    connected (via splitters or branching units) to one or morebase stations under test, as shown in Figure 21. The

    adjustable RF attenuator immediately preceding the BTS is

    used to simulate varying radio propagation conditions.

    With a call generation system as the heart of the test plant,

    its possible to create a set of tireless, demanding auto-

    matic subscribers who use every imaginable service the

    network can offer. The call generator creates controlled,

    lab quality traffic on mobile and fixed networks and

    reports detailed call analyses, including failure causes

    and QoS levels.

    Conclusion

    Planning & Engineering organizations have a huge

    responsibility in the typical networks business, where

    services are always in flux and the size and complexity of

    the network are constantly changing. The P&E group must

    oversee network expansion, the introduction of new

    features, and evaluation of new hardware and software

    elements. No wonder, then, that P&E engineers are seeking

    QoS measurement solutions that can get the job done

    quickly and efficiently.

    Drive test systems support wide-ranging network expansion

    efforts, and can do double-duty as QoS evaluation tools for

    the existing network. Their ability to pinpoint geographical

    locations and correlate them with network behavior at

    those locations makes them indispensable P&E tools.

    Similarly, the call generation system is key to test plant

    activities in which there is a need to simulate real users

    demands on the network. This system creates controlled

    yet exhaustive traffic for evaluating network elements and

    predicting how newly-installed equipment will interact

    with that which is already in place.

    21P L A N N I N G & E N G I N E E R I N G

    PC/Laptop

    BTS Under TestTo PSTN

    Test Plant System

    BranchUnit

    Attenuator

    Figure 21: Test Plant

    The RF connection

    is usually made

    through coaxial

    cables connected

    (via splitters or

    branching units) to

    one or more base

    stations under test,

    as shown in Figure 21.

  • 7/27/2019 Mobile QoS Primer

    22/33

    Quality Overview

    To achieve the best possible QoS while remaining competi-

    tive in terms of cost and services, quality efforts must bridge

    organizational boundaries within the network. Seamless

    cooperation among P&E, O&M, Marketing, and all the other

    entities is essential. To foster this level of cooperation,

    many networks assign dedicated personnel to the task of

    overseeing QoS across network departmental lines. Others

    establish a fully-constituted quality department with a

    charter to develop, implement, and support QoS programs.

    Subscriber satisfaction is the path to increased market

    share, higher revenues, and rapid subscriber growth for

    the network. But quality programs cannot be pursued

    blindly without regard to cost. A pragmatic investment in

    technical infrastructure can increase ROI, reduce costs,

    and boost profits. Conversely, an excessive investment can

    drive up costs (and therefore rates), yet may not provide a

    perceptible improvement from the subscribers viewpoint.

    Moreover, constantly upgrading and enhancing services is

    very likely to interrupt mobile traffic in some way. And thats

    all it takes to prompt the dreaded churn phenomenon.

    The Quality activity supervises QoS programs that maintain

    this equilibrium. A good QoS program acts as a virtual

    subscriber who can be replicated throughout the system to

    artificially stress it and identify the lapses in service that

    would be perceptible by real subscribers. If the QoS program

    discovers deficiencies that will remain safely beneath the

    subscribers threshold, they can be tolerated in the interest of

    holding investment costs and operating costs to a minimum,

    and keeping profits up. If the program reveals deficiencies

    that produce perceptible lapses in service, then technical

    investments can be stepped up proportionally.

    The Quality activity supports the networks business in other

    ways, as well. For example, QoS is an article often specified

    in Service Level Agreements (SLAs) between network oper-

    ators and between operators and carriers. And regulatory

    bodies in some regions require network providers to

    publish QoS data. The Quality activity is the focal point

    of compliance with these provisions.

    QoS Considerations for Quality Personnel

    From the perspective of the typical network user, that is,

    the individual who simply wants the convenience of a

    mobile phone for person-to-person conversations, voice

    quality is a key criterion of his/her network providers QoS.

    Problems with voice transmissions are easy to hear and are

    very distracting to the par ties on either end of the conversa-

    tion. The subscriber doesnt know about base stations and

    protocols; only that its hard to understand the voice

    coming through the mobile set. Unresolved complaints

    about voice quality quickly lead the subscriber away from

    the network.

    As a result, voice quality is a major point of emphasis for

    network quality efforts. Several methodologies are accepted

    for managing voice quality in the network:

    Drive tests

    INMD (In-service Non-intrusive Measurement Devices)

    Traffic generation and analysis

    Measurement Challenges and Solutions

    Many network operators choose to manage voice quality by

    conducting tests that simulate real voice calls, and duplicate

    as closely as possible the circumstances that impact voice

    qualityproblems such as interference, fading, and others.

    Given these techniques, it is no surprise that drive tests are

    a widely-used method of checking voice quality, as well as

    other phenomena such as dropped calls, handover failure,and more. Normally the drive test system includes a fixed

    station connected to a PSTN network, and mobile station

    that (as the name implies) is driven around in a car or

    truck. The mobile station has access to a GSM network.

    With this combination of equipment, it is possible to set

    up PSTN-GSM or GSM-GSM connections.

    Drive Tests Deliver Objective Measurements

    Drive tests are a powerful tool for monitoring voice quality.

    Innately, they perform end-to-end testing of the whole

    transmission/reception path. They deliver call-based analysis,

    so specific calls can be analyzed. The drive test system

    provides geographical details about the location of the

    mobile unit at any given time, which means voice quality

    issues can be correlated with physical obstructions that

    might be the source of the problem. A further benefit of

    drive testing is that it makes it possible to compare

    competing networks performance.

    Using the drive test system to conduct GSM-GSM tests,

    both the uplink and downlink path are measured. To further

    localize the source of the problem, the system can be set

    up to run PSTN-GSM measurements. In this case only the

    uplink or the downlink is measured.

    To achieve the best

    possible QoS while

    remaining competitive

    in terms of cost and

    services, quality

    efforts must bridge

    organizational

    boundaries within

    the network.

    Seamless cooperation

    among P&E, O&M,

    Marketing, and

    all the other entities

    is essential.

    Q U A L I T Y O F S E R V I C E I N M O B I L E N E T W O R K S

    Quality

  • 7/27/2019 Mobile QoS Primer

    23/33

    Drive testing has a few limitations, however. It is not fully

    automated. Its two-station architecture requires a live

    user at either end, which is costly in terms of personnel.

    Moreover, drive testing is considered an intrusive

    methodology, since it actually uses the channels under testrather than simply monitoring them.

    Some modern drive test solutions eliminate the need for

    a skilled technician at the mobile end; the operation at the

    end is simplified to the point where unskilled personnel

    can be used.

    The most advanced drive testing tools produce an objective

    quality measurement known as the Mean Opinion Score

    (MOS). This is based on a model (known as the E-model)

    defined by ETSI that accounts for all possible impairments

    in a mobile network. The MOS takes the opinion out of

    drive testing and presents a figure based on analysis of real

    acoustic phenomena such as clipping, echo, noise, etc.

    Non-intrusive Measurements

    View the Whole Network

    In-service non-intrusive Measurement Devices (INMD) are

    a more sophisticated way of getting the same jobvoice

    quality analysiscarried out effectively. The INMD measure-

    ment systems name tells it all: the solution monitors active

    (in-service) network elements, yet it does so without itself

    sending and receiving information through those elements.

    Importantly, the INMD does not require any call simulation

    or generation; it simply tracks the network during actual

    use by subscribers.

    The INMD system is installed integrally with the network

    itself. It is made up of a central management/analysis

    element linked via WAN to remote site monitoring units

    connected at key points in the network, especially access

    trunks to subscriber premise and interconnection trunks

    between networks. The INMD system captures and analyzes

    key voice band transmission parameters such as echo,

    noise, clipping, and more. In addition, it makes measure-

    ments in both directions (incoming and outgoing) on

    answered calls.

    The embedded INMD system performs automatedmeasurements constantly, with no incremental cost of opera-

    tion other than routine management and maintenance of

    the monitoring system. And importantly, the system delivers

    measurement results that correlate closely with actual mobile

    use perceptions. For example, when the INMD system

    detects a poor echo reading, it almost always corresponds

    to an objectionable artifact that mobile users can hear.

    The INMD approach does not provide a means of

    comparing one networks QoS performance with that of its

    competitors, however. That sort of activity is better handled

    by drive tests or by traffic generation and analysis.

    Call Generators Duplicate

    Real-world Calling Situations

    Traffic generation is another powerful tool to support

    network voice quality control. The solution architecture

    involves call generator nodes monitoring strategic pointsin the network, or even the signals from other networks.

    Like the INMD system, these probes are managed from

    a central system linked to them via WAN. Tests can be

    automated, minimizing the personnel resources needed

    to oversee this class of QoS measurements.

    The traffic, or call generator is designed to simulate real-

    world conditions. It is uniquely capable of reproducing

    complex phenomena such as ping-pong handovers,

    allowing the operator to characterize the response of the

    network under these circumstances. By generating many

    calls simultaneously, the system can stress the network

    and observe the results. It is also the best tool for testing

    billing (e.g., rates changes when crossing boundaries or

    time zones).

    The call generator is the right solution when in-line network

    elements must be tested from end to end. It allows the

    operator to simulate very complex and challenging call

    procedures, and it faithfully models realistic user calling

    patterns. The call generator is, however, an intrusive tool.

    By definition, using the call generator consumes some of

    the very network capacity it is testing for.

    Conclusion

    A network operators quality function, whether it is

    organized as a full-fledged Quality Department or as a

    company-wide floating resource, exists, ultimately, to

    ensure subscriber satisfaction. As such, the Quality function

    has to become a surrogate network user who expects

    perfect conne