A Practical implementation of high-speed communication using digital subscriber line technology

Rochester Institute of Technology

RIT Scholar Works

Theses

Thesis/Dissertation Collections

1997

A Practical implementation of high-speed

communication using digital subscriber line

technology

Todd Engle

Follow this and additional works at:

http://scholarworks.rit.edu/theses

This Thesis is brought to you for free and open access by the Thesis/Dissertation Collections at RIT Scholar Works. It has been accepted for inclusion in Theses by an authorized administrator of RIT Scholar Works. For more information, please [email protected].

Recommended Citation

A PRACTICAL

IMPLEMENTATION

OF HIGH-SPEED

COMMUNICATION

USING

DIGITAL

SUBSCRIBER LINE TECHNOLOGY

by

TODD R. ENGLE

THIS THESIS IS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS

FOR THE DEGREE OF MASTER OF SCIENCE IN INFORMATION TECHNOLOGY

ROCHESTER INSTITUTE OF TECHNOLOGY

DECEMBER 1997

Professor A'ishaAjayi,Advisor

Master of Science in Information Technology

Capstone Project/Thesis Approval Form

Student

Name:_--'-~/---:::::o....;;d;.:..c\=·_f\~.:...-...~=::;;;...:.h...l...gJ~e""",-<,----

_

Student ID#:

_

ProjectfThesis Committee:

Name

Chair

Committee member

Graduate Program Chair

revised: 4/9/97

Wallace Library

Rochester Institute of Technology

Thesis Reproduction Permission Statement

Permission From Author Required

Title of thesis:

A Practical Implementation of High-Speed Communication

using Digital Subscriber Line Technology

I, Todd

R.

Engle, prefer to be contacted each time a request for

reproduction is made.

If permission is granted, any reproduction will not

be for commercial use or profit. I can be reached at the following address:

4 Folkestone Lane

Penfield, NY 14526

Phone:

(716) 377-3995

COPYRIGHTS

Allrights reserved.Nopartofthispublication_{may be}reproducedor utilizedin any form or

by

anymeans,electronic or_mechanical,

including

photocopying,recordingor

by

_{any information}

storage and retrieval _system,withoutpermissionin writing fromtheauthor.

Analog

Devices is a registeredtrademarkof

Analog

Devices Corporation.

Ascendis aregisteredtrademarkofAscend Corporation.

Aware isa registeredtrademarkofAware Corporation.

GTE is a registeredtrademarkofGTE Corporation.

Microsoft isa registeredtrademarkofMicrosoft Corporation.

PulsecomandWavePacerare registeredtrademarksofPulse Communications Corporation.

TeleChoice is a registeredtrademarkofTeleChoice Corporation.

Time Warnerisa registeredtrademarkofTime Warner Corporation.

3Comis aregisteredtrademarkof3Com Corporation.

U.S. Robotics is a registeredtrademarkof3Com Corporation.

Allother names andbrands arethe_propertyoftheirrespective owners.

Todd R. Engle Page 2 CapstoneThesis

ACKNOWLEDGEMENTS

Iwantto thank the

faculty

and staff oftheDepartmentofInformation

Technology

atRochester

Instituteof

Technology

_{for making}this

learning

experienceso_rewardingand enjoyable.Iwould

liketogive a specialthanks toProfessor A'isha

Ajayi,

_mythesisadvisorand aninstructorfor

several of_mycourses. Professor Ajayihas alwaysbeen an_extremelysupportive_teacher, andhas

providedme withthe_opportunitytoachieve to_{my fullest}potential.

Ialso wanttothank_my

family

andfriends fortheir

help

and support overtheyears. _{I specifically}

wantto acknowledge_myparentsand grandparents for encouragingmetocontinue_myeducation.

Dueto_{my family's many}examples,Ihave learnedthateducationisa

life-long

process.

Finally,

Iwantto thank_mywife,

Michele,

for

"being

there"

forme

during

thisentire process.

She hasserved as_myeditor,proofreader, and chief supporter.Isharethisachievement with

Michele,

mywife and_{my best friend.}

ToddR. Engle Page3 Capstone Thesis

TABLE

OF CONTENTS

1.

INTRODUCTION

9

1.1 DefinitionofKey Terms 11

1.2 DefinitionofHigh-Speed Communication 12

1.3 Multimedia: The FutureofCommunication 15

1.4 Requirements 16

1.4.1 High Data Transmission Rate 17

1.4.2 Global Deployment 17

1.4.3 Wide-Scale Deployment 17

1.4.4 Short-Term Deployment 18

1.4.5 Low Cost To Consumers 18

1.4.6 Low Cost To Suppliers 18

1.4.7 High Growth Potential 18

1.5 Objectives 19

2.CURRENT PROBLEMS 20

2.1 Technology Issues 20

2.1.1 Infrastructure Limitations 21

2.2 Economical Issues 22

2.2.1 Consumer Resistance 22

2.2.2 Supplier Resistance 22

2.3 Social Issues 23

3. POTENTIALSOLUTIONS 24

3.1 Integrated Services Digital Network

(ISDN)

24

3.1.1 Advantages ofISDN 25

3. 1.2 DisadvantagesofISDN 25

3.2 Digital Subscriber Line Technologies

(xDSL)

27

3.2.1 Asymmetric Digital Subscriber Line

(ADSL)

27

3.2.2 High Bit Rate Digital Subscriber Line

(HDSL)

27

3.2.3 Symmetric Digital Subscriber Line

(SDSL)

28

3.2.4

Very

High Bit Rate Digital Subscriber Line

(VDSL)

29

3.2.5 Advantagesof xDSL 30

3.2.6 Disadvantagesof xDSL 30

3.3 Cable Modems 31

3.3.1 Introduction 31

3.3.2 AdvantagesofCable Modems 33

3.3.3 DisadvantagesofCable Modems 33

4.PRACTICALSOLUTION

(XDSL)

34

4.1 SelectionCriteria 34

4.2xDSLandCableModems 34

Todd R. Engle Page 4 CapstoneThesis

5. PRACTICAL IMPLEMENTATION OFXDSL 35

5.1DIAL-UPXDSL 37

5.1.1

Reliability

andFault Tolerance 37

5.1.2

Flexibility

38

5.1.3

Interoperability

39

5. 1.4 System Usage 40

5.1.5Deployment Costs 41

5.1.6EconomicsofDeployment 42

5.1.7GlobalDemographics 43

5.2 End-to-end Interoperability 43

5.2.1 Two LevelsofProviders 44

5.2.2 PPPoverATM over xDSL 45

5.3 End-User Configuration 47

5.3.1 POTS Splitter 47

5.3.2 Hybrid High-Speed Modem 49

6.CONCLUSIONS 50

6.1 xDSL: The Only Practical Solution 50

6.2 Phased Approach 51

6.3 Dial-UpxDSL 51

6.4 Standard Protocols 52

6.5 Hybrid Modem 52

6.6 POTS Splitter 53

6.7 Promotionof xDSL andHigh-Speed Communication 53

6.8 Summary 53

7.DEFINITIONS AND ABBREVIATIONS 55

8.BIBLIOGRAPHY 61

9.ENDNOTES 64

ToddR. Engle Page 5 CapstoneThesis

LIST OF TABLES

Table 1-1: ComparisonofDifferent Transmission TimesatVarious Transmission Rates 13 Table 5-1: Remote Access andInternet Access Usage Patterns 41

ToddR. Engle Page6 CapstoneThesis

[image:10.564.65.515.96.256.2]

LIST

OF FIGURES

Figure 1: Communication System Overview 10

Figure 2: Access Network Architecture 11

Figure 3: ISDN Architecture 24

Figure4: ADSL Architecture 27

Figure5: HDSL Architecture 28

Figure6: SDSL Architecture 29

Figure 7: VDSL Architecture 29

Figure 8: Data Over Cable Reference Architecture 31

Figure 9: Dedicated ADSL Implementation Model 35

Figure 10:

Dial-Up

xDSLArchitecture 37

Figure 1 1: xDSLModemandPOTS Splitter 48

ToddR. Engle Page 7 Capstone Thesis

ABSTRACT

Thisthesiscreates a planforthepracticalimplementationofhigh-speedcommunicationfor

residences andbusinesses.

By

implementing

low-cost,

high-speedcommunicationon a global

scale, tremendousbenefits can occurinareas such asInternetcommunication,interactive

multimedia,telecommuting, anddistance learning. Despite several successfultrialsof various

high-speedcommunication _{technologies,many barriers}remainbefore deploymentcan occurto

thegeneral public. Thisthesisproposes a plantobridgethe _{gap between}theoretical teststudies

and globalimplementation.

Thisthesisevaluatesthreecommunication systems as potential solutionsforhigh-speed

communication and selects one systemasthesolution.Thethreecandidatesystems areDigital

Subscriber Linetechnologies

(collectively

referredtoas_xDSL), Integrated Services Digital

Network

(ISDN),

and cable modem.Thechosen

technology

_{solution, xDSL,} allowstwisted-pair

copper wire(i.e. telephone

lines)

tobeusedfor high-speedcommunication. Thechoice of xDSL

asthe

technology

solutionis basedon_many

factors,

all of which correspondtopracticality. The

intentofthis thesisis notthepromotion ofxDSL; rather,the_primaryobjectiveistocreate a plan

to _quicklyand_{globally implement}a

low-cost,

high-speedcommunicationinfrastructure for

residences andbusinesses.

Todd R. Engle Page 8 _{Capstone Thesis}

December1997 _{M.S.inInformation}

1.

Introduction

Throughoutthe world,businesses and residencesdependuponnetworkstoprovide

information,

education and entertainment._{The popularity}ofthe InternetandtheWorld Wide Web has created

atremendous_{demand for many}network-related_services, such asinteractivemultimedia,

telecommuting, video-conferencing,anddistance learning.

Unfortunately,

themost_widelyused

systemforremotenetwork access _analogmodems over voice-gradetelephonelines does

not providetheperformance required

by

today's users and applications.Users are

becoming

increasingly

frustratedwith _veryslow access_times,andbusinesses arelimited intheservices

they

can_currentlyprovide customers.

A_{cost-effective,}high-speedcommunication systemforresidences andbusinesses capable of

globalimplementation canhavea_{revolutionary impact}ontoday'sworld (Broadband Services

Expert

Group,

1994). Just as construction oftheinterstate

highway

systeminthe 1950s

facilitatedtheexpansion oftheUnited

States,

theconstruction ofinformation highways inthe

1990s will generate economic and social changes wellintothenext_century

(3Com,

1995). The

globalimplementation ofhigh-speedcommunication will allowindividualstointeractwithout

regardtogeographicalboundaries.

From 1995 to

1997,

several major_{corporations,} such as

Microsoft,

GTE,

U.S.

West,

andTime

Warner,

conducted_{many data}trialsofhigh-speedcommunication

(GTE,

1997). The datatrials

successfullydemonstratedthecapabilities of varioustechnologiesfor high-speedcommunication

tohomesandbusinesses.Whilethedatatrials_{focused primarily}onthe technical detailsof

high-Todd R. Engle Page 9 Capstone Thesis

speed_{communication, many} moreissuesneedtobeaddressedand resolvedbefore

implementation

can occurto the generalpublic. Withouttheresolution ofthese

issues,

the

availabilityofcost-effective,high-speed communicationto thegeneral public_may not occur until

theyear2000orbeyond

(King,

1996). Apracticalplanneeds tobecreatedwhichfocuses onthe

real-worldissuesof

deployment,

inordertopromotetheimplementationofhigh-speed

communicationinthe_very nearfuture.

Thisthesisproposes a practicalplan_{for globally}

implementing

a_{cost-effective,}high-speed

communicationinfrastructure for businesses and residences.Three high-speedcommunication

technologiesareevaluated,and a single_{technology, Digital Subscriber Line}

(xDSL),

ischosen as

the

technology

solution.Theplan specifies a_complete,end-to-end systemfor high-speed

communication,fromthe end-user, tothenetworkaccess_{provider, to the} service provider:

Service

Providers

Network

Access

Provider

Corporate

Networks

Customer

[image:13.565.117.447.385.599.2]

Premise

Figure 1: Communication System Overview

(U.S.Robotics, 1997)

Todd R. Engle

December1997

Page 10 Capstone Thesis

1.1

Definition

of

Key

Terms

Customer_premise, asshownin Figure

1,

includesresidences,homeoffices and smallbusiness

offices. Customer Premise Equipment

(CPE)

represents _any

device,

locatedwithin thecustomer

premise, thatispart of a communication system.For_example, aresidence isacustomer_premise,

atelephoneis atype of customer premise_equipment, andthe telephonenetwork represents a

communication system. Forthis thesis,CPEwillinvolve_modems,personalcomputers

(PCs),

and otherdevices_{necessary for}

implementing

high-speedcommunication.

Accessnetworkisthecommunication systembetweenthecustomer premise and an access

network_provider,such as atelephoneor cable company.Thetelephonesystem's access network

is usuallyreferredto asthelocal

loop,

and_{predominately}consistsoftwisted-paircopper wire

connectingthetelephonecompany'scentral office

(CO)

toa customer premise. The connection

betweenaCOand a customerpremiseis commonlycalledthesubscriberline. Whilethis thesis

specifies acomplete end-to-end_system,it focuses primarily ontheaccess _network,customer

premiseequipment,andthesubscriberline. In Figure

2,

atypicalaccess networkisshown:

CustomerPremise Equipment

Telephone

Personal

Computer

Modem

Voice

Switch

))

A

Customer Premise Subscriber Line TelephoneCompany Central Office

Figure2: AccessNetwork Architecture

ToddR. Engle December 1997

Page 11 Capstone Thesis

[image:14.565.103.461.488.643.2]

Two

terms,

upstreamand

downstream,

are used

frequently

in thisthesis.Upstreamrefers to the

directionofdataflowfromthecustomer premisetotheaccess network provider.Downstream

refers to thedirection ofdata flow fromtheaccess network providerto thecustomer premise. In

situations wherethesubscriberline'supstreamdata flow isnot equalto thedownstream data

flow,

thesubscriberline isconsideredtobe asymmetric.

Conversely,

a symmetric subscriberline

represents an equaldata flow for boththeupstream anddownstreampaths (i.e.channels).

As stated_{previously, the}proposedimplementationplanispractical. Thedefinitionsoftheword

practical arefollowedthroughout the thesis.Theplanis

"relating

_to,or manifestedinpractice or

action: nottheoretical or

ideal.1"

The details for implementation are"capableof

being

puttouse

or account:

useful.2"

Finally,

the_underlyingthemeofthe thesisis "disposedtoaction as opposed

to speculation or

abstraction.3"

1.2

Definition

of

High-Speed Communication

Sincethe term"high-speedcommunication"

isrelative,itmustbe defined inthis context.In

1987,

high-speed datacommunication(tohomesandbusinessesovertelephone

lines)

wouldbe

definedas 9,600 bitspersecond(9.6 kilobitsper_second,orkbps). Thisthesisdefines high-speed

communication as millions ofbits

(megabits)

persecond (Mbps).Thereasonfor

defining

a

multi-megabittransmissionrate is duetothedata intensive demandsof current andfuture

network services. The

following

tablecompares differenttransmissiontimesat various

transmissionrates, andhelps todemonstratetheenormous potential ofhigh-speed

communication:

Todd R. Engle Page 12 Capstone Thesis

[image:16.564.49.520.79.304.2]

Table 1-1: Comparison ofDifferent Transmission Timesat Various Transmission Rates

File Types FileSize Transmit Time @ 56 kbps Transmit Time @ 64 kbps Transmit Time @ 384 kbps Transmit Time @ 1536 kbps (1.5

Mbps)

Transmit Time @ 2048 kbps (2

Mbps)

Digitized Photo 1 Megabit 17.9 sec 15.6sec 2.6sec 0.7sec 0.5 sec

Movie

Clip

2 Megabits 35.7 sec 31.2sec 5.2 sec 1.3 sec 0.9 sec

X-ray

Image 40 Megabits 11.9 min 10.4min 1.7 min 26 sec 19.5 sec

Large Computer

Program

500 Megabits 2.5hours 2.2hours 21.7 min 5.4min 4min

Themaximumdownstreamspeed oftoday's _analogmodemsis 56,000 bitsper second(56

kilobitspersecond,or56 kbps). Asshown

by

Tablel-1,

theperformance difference between56

kbpsand 1.5Mbps isextreme.Theperformancedifference between 56 kbpsand2 Mbpsiseven

greater.Whilegreater performance isalwaysdesired

by

the end-user, practicallimitationsmust

be

defined,

dueto technologicaland economical concerns.

Anincrease inaccess network performance willhavea_{wide-ranging impact in many} areas. The

possibilityexistsfortheaccess networkto_actuallyexceedtheperformanceof other areas of a

communication_system,such as a modem's interfacetoa personal computer(U.S.

Robotics,

1997).

Additionally,

theaccessnetwork's capabilities _mayalso exceedthemaximumdatarate of

a service provider'snetworks._{A study}oftheInternethelpstodemonstratethis case.

ToddR. Engle

December1997

Page13 Capstone Thesis

The

Internet,

andotherIP (Internet

Protocol)

_{systems,generally function in}anAvailable Bit Rate

(ABR)

mode, andarethereforegraceful about_{accommodating}various and_varyingspeeds.

Furthermore,

mostInternetservers

today

operateat56

kbps,

and power servers seldom operate at

speeds aboveTl (1.544

Mbps),

limiting

usefuldataratesto 1.5 Mbps forsometime

(Maxwell,

1997).

Forend-users_{performing Internetcommunication,}a6 Mbps transmissionrate ontheaccess

networkis notpractical, sincethe_majorityoftheInternetservers operate at 1.5Mbps orless.

Since Internetcommunicationis a

key

reasonfor

implementing

high-speedcommunicationto

homesand

businesses,

thecapabilities oftheInternethavegreat significancein thespecification

oftheaccess network'stransmissionrate.

Therefore,

this thesisproposes an initial

implementationofhigh-speedcommunicationwith a practicallimitof2 Mbpstransmissionrate

ontheaccess network.

Casual Internetusers_{may greatly}appreciatefasterresponsetimeswithhigh-speed

communication,butsomeindividuals may havemore criticalneeds.Table 1-1 showedthe

various responsetimes_{involved in retrieving}a40 Megabit

X-ray

atdifferenttransmission rates.

Ifa_physician,for_example,needs toreview this

X-ray

(whichis stored at a remote

hospital),

the

physician_{may be}abletowait20secondsforthe

information,

buta 12minute

delay

_may notbe

acceptableincritical situations.

Thisexampledemonstratesonebenefitofhigh-speedcommunicationinthe area of_{telemedicine,}

and also showstheenormous potential of

implementing

high-speedcommunicationtothe

general public.Withsome applications and_services,

however,

a multi-megabittransmissionrate

Todd R. Engle Page 14 Capstone Thesis

isnot_simply abenefitoracritical need it isa requirement. Anexample ofthis type of service

is multimedia.

1.3

Multimedia:

The Future

of

Communication

Multimediais a_{very important}and_noteworthy service which requireshigh-speed

communication. Multimediais definedasthesimultaneoususe of multipleformsof

communicationsmedia,such as_{text, graphics, voice, sound,}and still and_{moving images}

(3Com,

1997). Multimediaappealsto thevariousfundamentalmechanisms ofhuman_{communication,}

including

observation and

imitation,

spoken

language,

and written_{communication, creating}a

multi-dimensionalcontextfortheexchangeofinformation. Theuse of multiple communication

pathsincreases interestand_{significantly improves}retention

(3Com,

1997).

Multimedia is commonlyclassified

by

thenature ofthecontent andthemethod of consumption

by

users.Thetwomain categories are stored playback and real-timeinteractivemultimedia.An

exampleof stored playback multimediais playingatape on a video cassette recorder.

By

contrast, a video conferenceis an example of real-timeinteractive multimedia.Whetherthe

information isstored or

live,

thegoal of multimediaistocommunicateinformation quickly and

effectively

(MPEG,

1997).

Multimediaispart of afundamentaltrendincomputing, anditwill_{eventually be deployed}

everywhere

(3Com,

1997). Multimedia is

being

deployednot_{only in}workgroups,butthroughout

thecorporatenetwork, andis _reachingintothe

daily

livesofthegeneral population. The

multimediaapplications

being

deployedover networks

today

rangefrom distance

learning

(e.g.

Todd R. Engle Page 15 Capstone Thesis

computer-basedseminars and_{degree programs)} tocollaborative_{computing (e.g.} video

conferencing

andtelemedicine).Torespondto this_trend, networks mustsupport multimediain

allforms.Thefundamentalrequirementtosupport multimediatransmission is high-speed

communication, rangingfrom hundredsofkilobitspersecondtotensof megabits per second

(3Com,

1997).

Duetothecontinuousnature ofmultimedia

data,

itrequires _networkingtechnologies that

minimizeboth

delay

andlost information. Forstored-playback_multimedia, delaysof afew

seconds(betweentheissuanceof a command andthe

delivery

of

information)

istolerated

by

users,as

long

astheinformation is deliveredsmoothly.

By

contrast,inreal-timeinteractive

multimedia_{applications,} video and audiotransmissionmustbereceivedina

timely

_manner, or

effectivehuman interaction isnot possible. Significant delaysare not acceptable.

The MPEG

(Moving

PictureExperts

Group)

standards

body

is currently

developing

a standard

for interactivemultimediaapplications,entitledMPEG-4. The MPEG-4standard willbecomean

InternationalStandardin

January

1999,

andit specifies adatatransmissionrate

"up

to4Mb/s"

(MPEG,

1997). Interactivemultimediaapplications such as_{videoconferencing, videophone,} and

telemedicinerequirethehighbitratemode of

MPEG-4,

and

they

demonstratetheneedfor high

speed communication.

1.4 Requirements

Thisthesiscreates a set of requirements forhigh-speedcommunication, whichismodeled after

thelistof requirementsfromtheGTE Data Trial

(GTE,

1997)

andPulsecomCorporation's

Todd R. Engle Page 16 Capstone Thesis

systemsolutionfor high-speednetworks

(Pulsecom,

1997). Theimplementationplan adheres to

these_{requirements,}with_everyrequirement

being

metfortheimplementation plantobe

considereduseful. The

implementation

plan alsofollows a set of goals and objectives

(Pulsecom,

1997),

which areconsideredtobe both beneficialand

desirable,

butnot essentialforsuccessful

implementation.

1.4.1 HighDataTransmissionRate

As_{previouslynoted, the}proposedsolution mustinclude a multi-megabit perseconddata

transmission rate.Aminimumlimitof one megabit per second

(Mbps)

willberequired asthe

transmission rate.Atechnicalsolution with atransmissionratelessthan one megabit per second

will notbeselectedasa possible solution.

1.4.2 Global Deployment

Thesolution mustinclude deploymenttocountries and regions otherthan theUnited States. For

a communication plantobe

truly

practical,itmust recognizetheneedforworld-wide

communication._{The many issues relating}toglobaldeploymentmustbe addressed, such as

geographical

limitations,

political_{restrictions,}and economics.

1.4.3 Wide-Scale Deployment

Theplan must _{specify immediate}

implementation,

whichisto_{be commercially}used

by

the

general public.

Todd R. Engle Page 17 Capstone Thesis

1.4.4

Short-Term

Deployment

The solutionhas atime-to-marketrequirementof six months orless. A fiveyear plan isnot

acceptable.

1.4.5 Low Cost To Consumers

Apractical solution would notexpect consumerstoendurelarge_expense,eitherforone-time

equipment_purchases,orforcontinuous servicefees. Themaximumlimits forconsumer costis

set at$300forone-time equipmentpurchases, and$50per month forcontinuous servicefees

(ADSL

Forum,

1997).

1.4.6 Low Cost To Suppliers

Thesuppliers cannotimplement high-speedcommunication unless it is

financially

beneficial.

Dueto the _varyingand subjectivedegrees offinancial benefit forthevarious _{suppliers, this}

requirement cannotbeconstrained

by

specific numbers.

Therefore,

it is leftas a general guideline

fortheimplementationplantofollow.

1.4.7 HighGrowth Potential

Thesolution musthavelong-term future benefits. Itmust allowfor_{growth,specifically increased}

speed and new services.

Todd R. Engle Page 18 Capstone Thesis

1.5

Objectives

Thegoals and objectivesoftheplan are _generallydefinedas:

1. Utilizationof

existinginfrastructure: equipment,personnel and procedures

2. Utilizationof_existing

billing

andauthentication infrastructure

3. Selectionof ahighspeedservicewhich outperforms otherdatacommunicationtechnologies

4. Reductionoftraffic onPublic Switched Telephone Network

(PSTN)

voice switches

by

moving datatohigh-speed backbonenetworks

Theaforementionedrequirements and objectiveshaveone commonlink: theneeds oftheuser.

Gene Frantz

(Dunn, 1997),

business developmentmanagerfor Texas

Instruments,

states:"When

welookat_{emergingmarkets,} Iresorttoan

interesting

concept and

try

todecidewhatdoesthe

user_reallywant.And itappears thatwhathe wantsis adatarate of about 6 Mbpsor

better,

at a

cost of$150fortheboxat

K-mart,

and a_monthly service no morethan $20fromthe telephone

company."

To achievetheseextensive"customerrequirements"

as suggested

by

Mr.

Frantz,

severalissues

mustberesolvedbefore

low-cost,

_universal, high-speedcommunicationbecomesa reality.

Todd R. Engle Page 19 Capstone Thesis

2.

Current

Problems

Many

current problems mustbe solvedbefore high-speedcommunication canbe implemented.

Theproblemsexistin many_{areas, specificallytechnology,economics,}and sociology.Thearea of

sociology refersto the_{many issues surrounding}a society's_abilitytocomprehendand utilize

technology,in additionto theimpactof

technology

onthelifestyleof an individual.

Many

current approaches to theimplementationofhigh-speedcommunication are too_{narrowly focused}

on_technology, anddonot address_many other critical areas.

According

toAlbert Ankhoa

(1996),

customers "carenot about embedded_plant, new plant or

whethertheirinformation and entertainment services comefromtwistedpairs, coaxial_cables,

wireless, or satellitelinks. Whatconsumers wantis affordableprice and service

reliability."

While Ankhoa iscorrect on cost and_reliabilityconcerns oftheconsumer,the listmustbe

expandedtoinclude

high-speed,

availability,and minimal configuration.

2.1

Technology

Issues

Formost residences and_many

businesses,

datacommunication is accomplished overtelephone

linesvia voice-grade

(analog)

modems.Voicegrade modems operate atthesubscriber (i.e.

customer)endof voice grade linesandtransmitsignalsthrough thecore_switchingnetwork

withoutalteration; thenetworktreats them_{exactly like}voice signals.Whilethis_processing

methodhas many

drawbacks,

italsohas beena greatbenefit. Despite relatively slow_speeds,

Todd R. Engle Page20 CapstoneThesis

voice-grademodemscanbeconnected anywhere atelephonelineexists,andthereare_{nearly 800}

million suchlocations(ADSL

Forum,

1997).

Themost practicalimplementationplanof_any

technology

would_specifytheleastamount of

necessarywork. Foranimplementationofhigh-speedcommunication,atremendousbenefit

wouldbethe _abilitytousethe_existingnetwork oftelephonelines formulti-megabit

transmission.

Unfortunately,

thecurrenttelephonenetworkhas limitationson speed.

2.1.1 Infrastructure Limitations

The speedlimitationsoftelephonelines are notdueto thephysicalcharacteristics of_{the wire;}

rather,thelimitationsexist withinthetelephonesystem.Thetelephone_{company has filters} atthe

edge oftheaccess network which restrictthevoice channelto_{only 3.3 kilohertz}

(kHz)

bandwidth

(ANSI,

1995). Without

filters,

copper accesslinescan passfrequencies into

megahertz

(MHz)

_regions,albeit with substantial attenuation(ADSL

Forum,

1997). Despite

technological

limitations,

the_{existing infrastructure}oftelephone lines hasthepotentialto

becomepart of ahigh-speednetwork.

Cable Television

(CATV)

alsohas significantinfrastructure limitations.

Only

about5%of

CATVcompanies'

cable facilities have beenupgradedto

two-way

digitaltransmission_capability

(Ankhoa,

1996). The onlypractical plantoimplementhigh-speed datacommunication,ina_very

shorttime

frame,

istoreducethe

time-consuming

and expensive taskof new wireinstallations.

Todd R. Engle Page 21 Capstone Thesis

2.2

Economical

Issues

Apracticalimplementationofhigh-speedcommunicationmust considertheeconomical issues. Consumerand supplier resistancetonew

technology

is discussed inthe

following

sections.

2.2.1 Consumer Resistance

Thecustomercostisacriticalissue. As farasthecustomersare_concerned,

they

havethechoice

tobecouch potatoes ortointeractwiththecontents at will

(Ankhoa,

1996). Ifcustomer costis

too

high,

consumers will continuetouse _analogmodemsforthenextdecade. Theentire plan

mustbeconstructed with customer costinmind.

2.2.2 Supplier Resistance

Suppliers are resistanttocapital expenditures without short-term return oninvestment. The

implementationplan cannotdedicate largecapitaltoinfrastructure

improvement,

becausethatis

not practicalinthecurrent economy.For_example,thecurrenttelephonenetwork usestwisted

pair copper wiretoconnect most residences and_{many businesses} tothetelephonecompany's

centraloffice(CO). Thecopper wireinfrastructure is

limited,

anddoesnothavethecapabilities offiberoptic cable.

However,

replacingthecurrentaccess network withfiberoptic cable(referredtoas

FTTC,

or

"fiberto the curb")is a_costlyproposition.

Industry

experts estimatethat_creating anFTTC

network will cost about$1,500per currenttelephone customer

(Ankhoa,

1996). Thereplacement

of_{the existing}800million worldwidetelephonelineswould cost morethanthree-quartersof a

Todd R. Engle Page 22 Capstone Thesis

trilliondollars

(Analog

Devices,

_1997).

2.3 Social Issues

Personalcomputersareseldomlocated inahomeadjacentto thetelevision ortelevisioncoaxial

cabling._{Personal computers, especially} ones

desiring

network_access,

typically

sit near a

telephoneline

(Ankhoa,

1996). Ifahigh-speedcommunication

technology

does not utilizethe

telephone

line,

theresidentialcustomer will needtoeitherinstall new_wiringor relocatethe

computer.Additional workforconsumersisviewedas adefinite drawbacktosuccessful

deployment.

Apractical planforhigh-speedcommunication would attempttofollowthemodel of

communication whichhasexisted_{for approximately} two_{decades. Consumers readily}understand

thebasicsystem of voice-grade modems andtelephones._{The simplicity}ofthismodelis alarge

contributing factorto thepopularityofdatacommunicationtoday.

Any

technology

solution

whichdeviates fromthismodel wouldneedtohave significant advantagestowarrantthechange.

Ifatelevisionis used as a networkinterface device (i.e.the television screenisthecomputer

screen), major changes must occurinthe social patterns oftheend-user. Theinterruptionof

normal social patterns of consumers would_greatlyreducethe _{opportunity for}successful

deployment.

Todd R. Engle Page 23 Capstone Thesis

3.

Potential Solutions

3.1

Integrated

Services Digital Network

(ISDN)

IntegratedServices DigitalNetwork

(ISDN)

isan all-digitaltelecommunications

technology

which _{simultaneously}transmitsvoice conversationsanddataoverthesame pair of copper

telephonewires.Two importantaspects ofISDN and

key

differentiators from analog

telephonelines are

flexibility

anddatatransferspeed.

ISDN ispart of an_evolutionaryprocess,startedinthe

1950s,

toupgradethe telephonenetwork

from analogtodigitaltechnology. Theultimate goal of

ISDN,

designedover adecade ago

by

internationalstandards_{organizations,} allows usersto access_voice,

data,

fax,

video and graphics

overthesame phone lineand_enjoytheerror-free performance ofdigitaltechnology. ISDN Basic

Rate Interface (ISDN

BRI)

operates over most ofthe telephone_{wiring in}placetoday:

iv

0JM**

iveii

f$fc

ft***

V&BI

Figure3: ISDN Architecture

(Microsoft, 1997)

Todd R. Engle

December1997

Page24 Capstone Thesis

[image:27.564.75.484.452.613.2]

ISDN brings_{the speed,}

flexibility

and_reliability ofdigitalcommunicationsinto homesor small

businesses,

delivering

the total bandwidthof a 144 kbps digital line intothreeseparate channels

(Microsoft,

1997). Twoof_{the channels,}calledB

(Bearer)

_channels, operate at64 kbpsand are

always usedto_carryvoiceordatatraffic.Athird channel, theD

(Data)

_channel,isa 16kbps

signalingchannelusedto_{carry instructions}whichtell the telephonenetworkhowtohandleeach

oftheB channels.ThecombinedBRIisoften referredto _simplyas "2B+D."

3.1.1 AdvantagesofISDN

Basedonspeed_{alone, ISDN is significantly better}than _analogcommunications.

Top

datarates

for ISDNarefourtimesfasterthanthelatestmodem

technology

over_analogphonelines (28.8

kbpsupstream).ISDN_{is nearly}ninetimesfasterthanV.32bis modems,whichtransmitdataat

14.4 kbps.

3.1.2 DisadvantagesofISDN

Unlike analogtelephoneservice, ISDNserviceis not_universallyavailable. A localtelephone

companymusthave installedthe_necessary equipmentintheend-user's central office.

Additionally,

because ISDN is adigitalservice,it is very sensitivetooutsideinterference. An

end-user'sresidenceorbusinessmustbe locatedwithin alimited distance ofthe telephone

company'scentral office

(typically

18,000 feet). As a_result, eveniftheequipmentis installedat

thecentral_office, ISDN maynotbea communication option_{for many}users.

ToddR. Engle Page 25 Capstone Thesis

From ahigh-speedcommunication_perspective, ISDNis disadvantageous forseveral reasons.

First,

ISDNBRIprovides atransmissionrate of_{only 128}

kbps,

whichisnot sufficientfor many

data-intensiveapplicationsofthepresent and

future,

such as multimedia.

Secondly,

ISDNwill

usetwochannels through thePublic Switch Telephone Network

(PSTN)

foradatacall,

contributingto thecongestion ofthePSTN

(Ascend,

1996).

Finally,

ISDN has inherent

complexity, dueto theintegrationof voice anddata. Eventhoughmost users_{employ ISDN}

mainly forthedataservice, the "integratedservices"

remain as partofthe system.

Conclusion: ISDN may have begunthemigrationofdigitaloverthelocal

loop,

butxDSLis the

technology

which will_eventuallyreach_{every home} and office

(Ascend,

1996).

Todd R. Engle Page 26 Capstone Thesis

3.2 Digital Subscriber Line Technologies

(xDSL)

3.2.1

Asymmetric

DigitalSubscriber Line

(ADSL)

Asymmetric Digital Subscriber Line

(ADSL)

provideshigh-speedcommunicationtoalmost_any

homeorbusiness with atelephoneline. ADSL suppliesthree separate

frequency

channels over

thesamephoneline.Phoneconversations are carried on one_channel,whiledata fromtheservice

providertotheuseristransferredon anotherlineat speeds_{ranging from 16 kilobits-per-second}

(kbps)

to9million-bits-per-second(Mbps). Thethirdchannel runsdataupstreamfromtheuser

to theservice provider at speeds_upto640kbps.

Server

/

Core

I

Network

Existing

telephone

line

ADSL

ADSL

1 S tn f\ Mbps

^

Internet

<

ui tn640 kbps

Figure4: ADSL Architecture

3.2.2 High Bit Rate Digital Subscriber Line

(HDSL)

Using

the_existingcopper

infrastructure,

HighbitrateDigital Subscriber Line

(HDSL)

provides

full duplex Tl (1.544

Mbps)

orEl (2.048

Mbps)

datatransmission across_existingtwistedpair

Todd R. Engle December 1997

Page 27 Capstone Thesis

[image:30.564.54.422.339.513.2]

copper without repeaters. _{Not only is HDSL}thefastestandleast costly solutionfor

deploying

Tl/El

lines,

italsoprovidestransmission _qualitycomparabletofiber.

Existing

telephone

line

544nr7 04S Mbps

[image:31.564.52.422.127.310.2]

4 1 544 nr 7 04R Mbps

Figure 5: HDSL Architecture

3.2.3 Symmetric Digital Subscriber Line

(SDSL)

Symmetric Digital Subscriber Line

(SDSL)

_{simultaneously}provides symmetric

(bi-directional)

high-speed,

variable rate communications andPOTS (Plain Old Telephone

Service)

on a single

phoneline. Ittransmitsdataat 160 kbpsto 2.084 Mbps. This

technology

issuitablefor

applicationsthatrequire a symmetricdatarate.

SDSL,

_alongwith

ADSL,

relievesthecongestion

onvoice networks. Voiceanddatawill notbetransmittedthrough thecircuit switched_network,

as_{currently done}with_analogmodems andISDNservices.

Instead,

telephonecompanies will

routevoiceto thecircuit-switched networks anddatato thepacket-switchednetworks.

Todd R. Engle December 1997

Page 28 Capstone Thesis

Existing

telephone

line

60 kbps tn 7 Mbps

160kbps tn 7 Mbps

Figure6: SDSL Architecture

3.2.4

Very

High Bit Rate Digital Subscriber Line

(VDSL)

With

Very

high bitrateDigitalSubscriber Line

(VDSL),

thedatarateincreases asthecustomer's

locationgrows closerto thecentral office. A datarate of 13 Mbpscanbe providedtoa subscriber

5,000 feet away fromthecentral_office,while a rate of26 Mbpscanbeprovidedtoa subscriber

3,000 feetaway. 51 Mbpscanbeachieved at adistanceof 1,000feet fromthecentral office.

Existing

telephone

line

n Mbps a oooft

)

76 Mhrs noon ft

)

51 Mbps (1 OOPft

)

Figure7: VDSL Architecture

Todd R. Engle December 1997

Page29 CapstoneThesis

[image:32.564.50.423.94.279.2] [image:32.564.54.405.477.645.2]

3.2.5

Advantages

of xDSL

A

key

advantageforxDSL

technology

isthatitcan operate over_{existing POTS (Plain Old}

Telephone

Service)

_wiringwithout_affectingthevoice_{transmission, eliminating}theneedfornew

cable

installation.

_ThexDSL

technology

providesadedicatedservice over a singletelephone

line,

whichis more securethana shared medium.

Also,

sincethesingletelephoneline isnot

sharedwith otherusers, theavailabletransmissionrateofthelinewillbeconstant.

However,

the

mostsignificant advantageof xDSL isthe_availabilityof800milliontelephonelinesworld-wide.

3.2.6 Disadvantagesof xDSL

Whilethe_existingcopperinfrastructureis a

key

advantageforxDSL,it may alsoprovetobea

limiting

factor in

implementing

xDSL on alarge scale._{Three underlying}problemsofthe

infrastructureare noted

by

CarusoandRendleman (1996).

First,

xDSL cannotbeprovisioned on

coppercircuitsequipped with loadcoils,which phone companieshave installed inthepastfor

residences far froma central office.Thisproblem affects as much as 20percent of all copper

lines.

Secondly,

xDSL'shigh frequenciespose considerableinterferencepotentialtoother copper

lines inthesame cable sheath.

Finally,

length limitations on_{existing local loops}meanthatusers

andcarriers will needtoadjustfor longer

loops,

_particularlywithhigherspeed xDSL.

Theseproblems are_notable,but

they

alsohavesolutions.

Bellcore,

which claimstohave

invented_xDSL, statesthat_{line quality} should notbea significantfactor ifequipment vendors

and providers stickto thestandard and_staywithintheguidelines of 1.5 Mbps at 18,000 feet.

Regarding

loadcoils,adigital

loop

carrier

(DLC)

devicecanbe used.Copper lines connectto the

DLC,

which connectsto a central office.Iftheuseris within 18,000feetofthe

DLC,

theusercan

still get xDSL.

ToddR. Engle Page 30 _{Capstone Thesis}

December1997 _{M.S.inInformation}

3.3

Cable

Modems

3.3.1

Introduction

Acable modemisadevicewhich allowshigh-speedcommunicationthroughthecableTV

(CATV)

network. Forconsumers,a cable modem will

typically

have two connections,oneto the

cable wall outlet andtheothertoa computer. Cablemodem

technology

hasthepotential to

produce a30 Mbpstransmissionrateto theconsumer(i.e.

downstream),

and a"low

Mbps"

upstream rate

(Maxwell,

1997). The

following

diagramrepresentsthecablemodemarchitecture,

asdefined intheCable ModemtoCustomer Premises Equipment Interface Specification

(Lipoff,

1996): ;Personal !Computer Cable Modem Coax Distribution Network O/E Node Fiber O/E Node O/E Node

Distribution HuborHeadend

upstream splitter andfilter bank .data Combiner CableModem Termination System Demod. Mod. remote dialup access server generic headend switch

Figure 8: Data Over Cable Reference Architecture (Lipoff, 1996)

Todd R. Engle

December1997

Page31 Capstone Thesis

[image:34.565.51.510.313.639.2]

Cablemodemsoperateoverhybrid fiber/coaxial

(HFC)

networks, comprised offiber feeder from

a

head-end,

andbranch coaxialcablesinstalledfromtheOptical Network Unit

(ONU)

to

customerlocations. _{In manyconfigurations,}as_manyas 100users_mayshare a single cableline.

Forcommunicationtooccur overCATVnetworks via cablemodems,thecustomer'scable

modem communicateswiththecablecompany'scablemodem, whichis locatedattheONU

location.

Most CATVsystems

today

are not

HFC;

rather

they

aretree andbranchcompositions of coaxial

cable, sometimes_servingas_manyas 10,000customers froma singlehead-endunit, with one

wayamplifiersthatpreclude_anyupstream data flow

(Maxwell,

1997). Since

1993,

_{many CATV}

lines have been installedwith

two-way

_{amplifiers, creating}an upstream pathfrom 5to45 MHz.

However,

thesheer size ofthesenetworks andthenoise and channel problems with so _many

subscribers attachedtoa commonline makehigh-speedupstreamchannels unattainable after

afewsubscribershave joinedtheline

(Lipoff,

1996).

Theprocess of_upgradinga coaxial systemfromunidirectionalto_{bi-directional may be}

accomplished

by

_{physically replacing}amplifiers. Theprocessof_{upgrading from}coaxtoHFC

requires a greatdealof work:

installing

fiberoptic_cables,

installing

ONUs,

_{rerouting any}coax

not convenientto the

ONU,

and_replacingthe_{few remaining}coax amplifiers. Cablecompanies

are_veryresistanttocapital expenditures unless ahigh ROI (returnon

investment)

exists

(Maxwell,

1997).

Todd R. Engle Page 32 Capstone Thesis

3.3.2

Advantages

ofCableModems

Onesignificantadvantage of cable modem

technology

for high-speedcommunicationis that

transmissionrates donotdependuponcoaxial cable

distance,

asamplifiersinthecable network

boostsignal power_sufficiently

(Ankhoa,

_{1996). Another very important}advantageisthe

potential30 Mbpstransmissionratetotheend-user.

3.3.3 DisadvantagesofCable Modems

Several distinct disadvantagesexist_{for using}cable modems as ahigh-speedcommunication

technology.

First,

variationincable modem_capacity will_occur,dependenton noiseintheline

andthenumber of simultaneous users_seekingaccessto a sharedline.

Secondly,

alimited

number ofhomes are wiredforcable, _{especially homes}outside oftheUnited States.

Additionally,

very few businessesthroughouttheworld are wiredforcable

(Ankhoa,

1996).

Athirdand most significantdisadvantage isthe lackofbi-directionalcommunication. To

providebothupstream anddownstreamchannels,cable companieshavetwooptions: either make

major(and_expensive)enhancementsto the cable

infrastructure,

or utilize a separate connection

forupstream communication _possiblythroughan_analogmodem overthe telephonenetwork.

Finally,

cable modem standards are still_evolving,which willfurther limitthe

interoperability

of

equipmentfrom differentvendors. Withall ofthese

disadvantages,

cable modems(andthecable

infrastructure)

cannot competewiththe_manyadvantagesof xDSL

technology

andthetelephone

network.

Todd R. Engle Page 33 Capstone Thesis

4.

Practical

Solution

(xDSL)

4.1

Selection

Criteria

Theselectionof xDSL asthepractical

technology

for high-speedcommunicationwasbasedon

the requirementsfrom Section 1. Dueto the800milliontelephonelines inexistenceworld-wide,

any

technology

thatcanutilizea_widelyavailableinfrastructure has adistinctadvantageover

otherformsofcommunication.

4.2

xDSL and

Cable Modems

800milliontelephonelinesexistintheworld_today, with70% connectedtoresidences,andthe

balanceconnectedtobusinessesand_payphones. _{Despite emerging}technologiesfornon-copper

basedsystems, theworld-wide copper

loop

plantisstill projectedtoexceed900millionlines

by

theyear2001. IntheUnited

States,

approximately 80%oftheselinescan accommodateADSL at

approximately 1.5

Mbps,

and50% can support rates of6 Mbpsor more(ADSL

Forum,

1997).

Many

other countrieshavemorefavorablecablelength statistics.

However,

some countries or

regions_{have very}old

loop

plants, andthepercentage of copperlinesthatare _actuallyusablefor

megabitaccess _{may be}wellbelow 80%

(TeleChoice,

1997). Inthesesituations,variablerate

ADSL (Rate-Adaptive

ADSL,

or

RADSL),

with speeds _{slightly below 1}

Mbps,

_mayenable

connectiontoall users.In summary: xDSListhechosen

technology

forseveral_reasons, withthe

primaryreason

being

the_availabilityofthemedium.

ToddR. Engle Page 34 Capstone Thesis

5. Practical Implementation

ofxDSL

Forsuccessfuldeploymentof xDSLtechnologies, the

industry

needsto incorporatethemost

practical implementationmethods.

Currently,

dedicatedxDSLimplementations haveone central

office xDSL modemdedicatedtoeach subscriber.This implementationmethodhas had negative

effects on _{scalability, reliability,}and_cost,andhaspreventedthedeploymentofhigh-speed

communicationvia xDSL. The

following

diagramrepresentstheimplementationmodel of

currentADSLsystems:

ADSL implementation_Today

OmTele*CentralOffle*

Modamts[indicatedlo Each HouseholdOr Business

ATMrHighspeed B4cW>omUr*

DatiCsnltr

[image:38.564.186.405.301.609.2]

HulU{ll**K

Figure 9:DedicatedADSLImplementationModel

(NetSpeed, 1997)

Todd R. Engle

December1997

Page35 Capstone Thesis

Thecurrentdedicatedmodel of xDSLis _simplynotpractical forwide-scale

deployment,

and

another model mustbeconstructedforxDSLimplementation.

Interestingly,

thecorrectmodelto

follow isprovided

by

today's telephonesystem.Theparadigm_currentlyused withvoice switches

and _analogmodemsallowsthesedevicesto

dynamically

allocatebandwidth. The

oversubscription of portsinacentral office enables a singledeviceto serve several users.

CurrentxDSL systems are

fairly

_costlyto

implement,

largely

dueto thelackof volume

deployment. Withmass-market

implementation,

service providers willbe abletoreducethecost

toconsumers.

However,

untillarge-scale deploymentoccurs, theeconomics ofdedicatedxDSL

systemsare notbeneficialtoconsumers or service providers.

Therefore,

the _onlypractical

solutionis _{non-dedicated,}

"on-demand"

xDSL_systems, referredtoas

dial-up

xDSL.

Todd R. Engle Page36 Capstone Thesis

5.1 Dial-UpxDSL

Personal Computer A Personal Computer B Fax xDSL Modem xDSL Modem relephone Telephone_Company Central Office

f^t

1&

u

Telephone Line (singlepair twistedwire) POTS Splitter

^-&

Q

Telephone Line (singlepair twistedwire) POTS Splitter POTS

6

Voice Switch POTS Splitter

It

xDSL multiplexer

u

' '."_"T POTS

Splitter XDSL multiplexer

Telephone_Company Central Office

Figure 10:

Dial-Up

xDSLArchitecture

5.1.1

Reliability

andFault Tolerance

Serviceproviders and customersareaccustomedto the_redundancyof_existingnetworks.With

dedicatedxDSL, the one-to-one natureofxDSL connections will

likely

require manual

interventiontore-establishlinks

following

a modemfailure.

By

_providinga_one-to-many

ToddR. Engle December1997

Page 37 Capstone Thesis

[image:40.565.53.513.100.433.2]

association with

dial-up

_{xDSL,consumers} willnot sufferserviceoutages ifasingle central office

modemfails.

5.1.2

Flexibility

As previouslymentioned, the_{many forms}of xDSL

(HDSL, ADSL, SDSL,

VDSLand

RADSL)

are still_evolvingandwill drivetheneedfor flexiblesystems. Serviceproviders will not want

decisionsmade

today

tolimit futureoptions orthe_abilitytoprovide newfeaturesand services

quicklyand costeffectively.WhilexDSL modems arebuiltto support a singleline_code, it's

unlikelythatcentral offices or even individualservice providers will

deploy

a singleformof

xDSL.

However,

thealternative method_{for achieving compatibility}

designing

dual linecode

modems is

truly

price prohibitive

(TeleChoice,

1997).

Fromatechnicalstandpoint,

dial-up

xDSL relies onthe transmissionof"request for

service"

tonesthatare generated

by

customer modems. Theremotemodem signalsitscentral office

counterpartthat_{it is ready}toform abroadbandconnection.This interaction isperformed atthe

physicallayerand canbe initiated

by

softwaredrivers thatgenerate a"starttone"

whendata

needstobesent.Theoff-hookdetection

by

softwareinitiatesadigital dialtone that thenforms

connectionson an as-needed

basis,

_allowingseveral usersto_effectively share a pool of modems

atthecentral office.

Todd R. Engle Page 38 Capstone Thesis

5.1.3

Interoperability

Thedigitaltonescanbedesignedtosignalthespecificlinecodefromthecustomer's_modem,

enabling service providerstodetectthe

incoming

modemtypeand route ittoaline-code

compatible modem inthecentral office system.Theinitiationoftheconnection can occur

transparently

to_{the user,}becausetheconnectionis digital. Connections are established without

the

time-consuming training

sequences associatedwithtraditional_analogmodems.

Quickconnectimplementations meanthatcustomerequipment can storetheparametersto train

onthelineasopposedto

training-up

foreach connection.This implementationcan reduce

connectiontime tofractionsof a_{second, allowing}subscriberlinesto_{be brought up} anddown

transparently

totheuser

(NetSpeed,

1997).

Sincenetwork connections are muchfasterthanvoice-grade

(analog)

modems,itencourages

growth oftelecomservices

by

users.Inthe situation where an end-user wantsinformation from

the

Internet,

the

delay

_{in establishing}themodem connectionto theISP

(approximately

one

minute) may discouragetheuserfromeven_accessingtheservice. _{The ability}tohave sub-second

connectiontimewould provide afastresponsetime toanimpatientworld.

Thedynamicnature ofthecommunicationis a

key

aspect.Whennodataarepresent, the

subscriber'sphysical connection istorn

down,

butalogicalconnection (i.e.virtual _circuit)

remainsintact. Theresultisthattheuser never perceivesdisruption in service, and service

providers canimplementmore cost-effective strategies.Forapplications and servicesthat

Todd R. Engle Page 39 Capstone Thesis

warrant

dedicated

access,connectionscanbeconfigured tobepermanent. The dedicatedaccess

architecture_mayappealto businessesandeducational institutions.

5.1.4 System Usage

A studyofInternetusagepatternsindicatesthatmost usertransmissionsareintermittent innature

(NetSpeed,

1997). Users downloadtheitem

they

_require, reviewit forsome_time, andthen

request anotherdatatransferwhen

they

are_{ready for}anotheritem.

Dial-up

xDSL enables service

providersto takeadvantageofthestatistical nature ofInternetaccess and gain ahigher levelof

average utilizationfromeach_{modem, thus}

lowering

deploymentcosts.

Inprivate_networks,remoteofficesandtelecommutersalso use_{WAN bandwidth sporadically}

(NetSpeed,

1997). Largeroffices_{may have}more constant need ofthroughputbecausea greater

number of people are_{placing demands}ontheshared network resource.

Dial-up

xDSL can also

providededicatedconnectionstomeettheneeds oflargercorporatelocations.

Itis

likely

thatbothresidential andbusinessusers willbeableto_{statistically}share xDSL

bandwidthsincetheirpeak usagehoursareatdifferenttimesin a givenday. Thestatistical

sharingofbandwidth worksandis demonstrated inthe

following

table:

Todd R. Engle Page40 Capstone Thesis

[image:44.564.44.521.101.433.2]

Table 5-1: Remote AccessandInternetAccess Usage Patterns

(NetSpeed, 1997)

User

Application

Usage Pattern

Peak Usage Hours

Residential Web InternetAccess connectsforonehour Activeat night (after5

Surfer or

less,

_extremely

bursty

data transmissions

p.m.) and weekends

Power Internet User Internetaccesswith connectsforone or Activeat night(after5

someuse of_streaming

less,

mixture of

bursty

_p.m.) and weekends

applications andcontinuous transmissions

Telecommuter Remote LANaccess Connects for

long

Active

during

afew

(e-mail,

file

transfer)

periods (3+

hours),

weekdays eachweek,

bursty

data occasional_evening

transmissions use

Small Office/Home LAN

interconnect,

Connects for longer Active

during

normal

Office Shared Internetaccess periods (8+

hours),

transmissions

fairly

continuous

business hours

5.1.5 DeploymentCosts

Because

dial-up

xDSL enables modem_pooling,it allows service providerstoreach more

subscriberswithfar fewermodems.Ifproviders_{conservatively}estimatethatxDSLconnections

are active_{only 50}percent ofthe_time,service providers can cut central office equipment costs in

half

by deploying

fewermodems.

ToddR. Engle

December1997

Page41 Capstone Thesis

As

demonstrated

in_{the table above,} residentialusers are_primarily activeinthe_evening

hours,

whereasbusiness userswilluse xDSL

during

theday.

Dial-up

xDSL'smodem_poolingallows a single modemto pull_{double duty:}

Serving

oneset ofsubscribers

during

the

day

andanotherin

theevening. _{The alternative,}

deploying

amodemfor everysubscriber, reducesthe _scalabilityof systems anddrives costs_up,reducingthebusinesscase_{profitability}and return on investment for xDSLservices.

By

_utilizingtheoversubscriptionmodel enabled

by dial-up

xDSL,service

providers can cost

justify

xDSLdeploymenteveninan_early market.

5.1.6 Economics ofDeployment

Many

businesscasesforxDSLservicesaredeveloped

by initially

_viewingthenumber ofhomes

orbusinesses

being

served.Thismodel often calculates return oninvestment

(ROI)

on a per modembasis.

Every

modemdeployedmustbecost-justified

by

therevenuesthatcanbe

recoupedfromanindividualsubscriber

(TeleChoice,

1997).

Dial-up

xDSL accelerates thexDSLbusinesscase

by

_{effectively enabling}asingle modemto "pass"

orbe

logically

deployedtomultiplelocations. This ismuch closertothebusinesscasefor

standard voice_services, with whichlargetelephonecompanies are_{very familiar. With} this sharedinfrastructuremodel,service providers can expect a quicker return ontheirinvestment.

Equipmentcosts can alsobe driven down ifvolumes areforcedup. Ifthepriceis sufficiently

low,

then _usingadedicatedmodemforeach xDSL customer_maynotbecost prohibitive. If

enoughservice providers launchedxDSL-based_services, itcouldspark a_{large early demand}and

Todd R.Engle Page42 CapstoneThesis

allow the_economyof scale_{necessary for}equipment vendors and service providerstosupport

lowprice points

(Ankhoa,

1996).

In additionto

lowering

equipment_costs,

dial-up

xDSL cuts operational costs inthe areas of

space, power, and reliability.

By

_sharingmodems across subscribers, theper watt consumption

by

subscribercanbereduced

by

50percent

(NetSpeed,

1997). Whenoversubscriptionisoffered

atthemore

likely

10: 1 ratio,the reductioninpower and costiseven moredramatic (U.S.

Robotics,

1997).

5.1.7 GlobalDemographics

Dial-up

xDSL

technology

enables service providers and corporationsto

bring

_upconnections

when users requirethem. Datanetworks canbe engineeredtomeet average trafficdemandsmuch

likevoice andISDNnetworks are createdtoday.

Additionally,

service providers can

deploy

systems_accordingtodemographics.

By

_monitoringuser_{activity for}a given_servingarea,

providers candesignsystemstomeettheneeds ofthecustomersinthatarea. Thecurrent

dedicatedmodel resultsin costlyover-engineering,and cannotscaleto the_varyingneeds of

world-wide subscribers.The onlypractical modelforwide-scaledeployment is

dial-up

xDSL.

5.2 End-to-end

Interoperability

Acurrent problem with xDSLimplementation isthelackofaddressing. An_{end-user, through}a

single xDSLconnection,needs accesstomultiple _services,such astheInternetor corporate

networks.AsynchronousTransferMode

(ATM)

is receivingwidespread support asthesolution

Todd R. Engle Page 43 Capstone Thesis

to xDSL's_addressingproblems

(Taylor,

1997).

However,

notall service providers agree with_{establishing "ATM}over

xDSL"

as a standard

addressingscheme. Most Internet Service Providers

(ISP)

andcompetitive Local Exchange

Carriers

(LEC)

prefer either

IP,

Point-to-Point Protocol

(PPP)

orFrame

Relay

over xDSL. To

solve_{this problem,}

Microsoft, Alcatel, Cisco,

Fore

Systems,

U.S. RoboticsandWestell

Technologies have joined forcestoarrive at an official standard(U.S.

Robotics,

1997). These

vendorshave collectivelyproposed an architecturethatwould usePPPoverATMover xDSLfor

interoperability

betweenmodems,routers and access multiplexers.

The

key

tosuccessful xDSLdeployment isend-to-end

interoperability

(3Com,

1997). With

interoperability,

a competitive multi-vendor environment canexist,whichleadstomore

affordable solutions.Throughthestandard

bodies,

significant progresshas beenmadein

establishingphysical layerxDSL

interoperability

(ADSL

Forum,

1997). To facilitaterapid

deployment,

thestandardization of upper-layer protocols must also occur.The fastest solutionfor

achievingstandardizationistobase theentire _addressingarchitecture on_existingstandards:

namely,PPPoverATMover xDSL.

5.2.1 Two LevelsofProviders

The deploymentof xDSL will_{closely involve}the telephonecompany.Insome

deployments,

the

telephonecompanies_mayrestricttheir_operatingscopeto theaccess network orthebackbone

network.In suchcases,explicit agreementontheinteractions between different levelsof

providersis mandatory. Forremote corporate_access, thedivisionof_{responsibility}is equally

Todd R. Engle Page 44 Capstone Thesis

critical sincethecorporationis almost_actingas a service provider.

Therefore,

it is importantto

distinguish

between an access network provider and aservicenetwork provider intheend-to-end

interoperability

model.

Thenetwork modelforxDSLwill notdeviate fromthe

dial-up

model.In_{this model, connectivity}

is explicitlyestablished at severallevels: thecall level

(dial-up),

thelink levelandthenetwork

level. Notethatboththe_setupandtherelease ofthesession at eachlayerare essentialforusage

metering and

billing by

theproviders involvedatthedifferent layers

(Microsoft,

1997).

5.2.2 PPP overATMoverxDSL

Theproposed xDSLend-to-end service

interoperability

modelis basedon an end-to-endATM

networkbetweenthecustomer premise networks andthe service provider networks

(ISP,

content

provider and corporate networks).The ATMendpoints includeallthe devices atthecustomer

premise (suchas aPC or_router)andtheservice provider network(anaccess server or a_router)

thatterminate theend-to-endATMnetwork. The ATMover xDSL architecture preservesthe

high-speedcharacteristics inthexDSL environments without_changingprotocols.

With ATMover_{xDSL, the}residentialandsmallbusinessofficecustomershave accessto

Internetenvironments. ATMover xDSL provides seamlessconnectionsfrom remote usersto_any

ATM distributionnetwork,to_{any ATM}

backbone,

to_anycorporate

intranet,

orto theInternet. In

addition, ATMprovidesdirectconnectiontoInternet/intranetservers,such as a_{security server,}

anInternetcontent_cachingserver, ora video server. ThisenhancesInternet services,intermsof

performance,loadsharing,andredundancy.

Todd R. Engle Page45 Capstone Thesis

5.2.2.1

PPPoverATM

Once ATM layer connectivity isestablishedbetweenthecustomer premise andtheservice

provider_{network, the}session_setupandrelease phases atthelink level and networklevelcanbe

established_{using PPP. The}definition of astandardfor PPPoverATMwillincreasetheuse of

ATM as anaccesstechnology. Otherconnection modelsfor ATM (such asClassical IPover

ATM)

targetcampus environments andlackthesecurity, session,and auto-configuration

functionality

thathigh-speed remote accessnetworks willdemand

(Microsoft,

1997).

Essentialoperationalfunctionscanbe deliveredover_{ATM using features}well establishedin

PPP:

Authentication

Layer 3 addressauto-configuration

Multipleconcurrentdestinations

(i.e.,

multiplePPP _sessions)

Layer 3

transparency

(e.g.both IPandIPXare_currently supported on

PPP)

Encryption

Compression

Billing

and usage_metering

By

adaptingthePPPsuite_{to xDSL,} the

delivery

of aninter-operableservice architecture willbe

accelerated. PPPoverATM iseven more valuablebecauseit adheresto thesame service model

used

by

theISPbusiness

(3Com,

1997).

Todd R. Engle Page46 Capstone Thesis

Easy

migrationfrom_{existing ISP}Access Infrastructure: SincexDSLis only replacingthe

underlying layerfor carryingthePPP_{session, themajority} oftheISPinfrastructureismaintained

(Microsoft,

1997).

Multi-Protocol Support: PPPsupports multiplelayer 3 protocols.

Security:

Security

isenabled

by

theconnection-orientednature ofATMplusPPP'sencryption

andauthentication mechanisms.SinceeachATMcircuit_onlycarriesaPPPsession,the_security

provided

by

PPPappliestotheentireATMconnection.

Multiple Service Classes: Multipleservice classes are enabled

by

thenature ofATMtosupport

differentservice classes.Individual user profiles canbe definedas part of_policymanagementin

thenetwork.

5.3

End-User

Configuration

5.3.1 POTSSplitter

Amodem andPOTS (Plain Old Telephone

Service)

splitteris installedattheuser's premiseto

transportvoice anddataoverthelocal loop. The POTS splitteris ahardware devicewhich

separates high

frequency

(xDSL)

andlow

frequency

(POTS)

signals at network end and premises

end.

The

functionality

ofthePOTS splitteris_essential,becauseit directsvoice callsto thevoice

switch andforwards datatraffic to thexDSLAccess Multiplexer

(DSLAM)

inthecentral office

Todd R. Engle Page47 Capstone Thesis

(CO). Withouttheseparation of

frequencies,

end-userswould nothavesimultaneous voice and dataservices. -Jk-Personal Computer A Fax

k#

xDSL Modem

n

POTS Splitter elephone Telephone Line (singlepair twisted wire) Telephone_Company Central Office

Figure11:xDSLModemandPOTS Splitter

Alargeobstacletowide-scaledeploymentof xDSL(andhigh-speedcommunication)isthe

installation ofthePOTSsplitter. Currentimplementationplans of xDSLrequire service

providerstoinstall aPOTS splitterina customerpremise,whichisanexpensive and

time-consumingactivity. End-userswouldpreferaninstallationprocesswhichisequivalentto today's

analogmodems.

Fortunately,

a solutionto thisproblem exists.

Aware

(1997)

demonstratedthefirst "splitterless"

versionofxDSL

technology

suitablefor PC

modem applications.

Delivering

downstreamspeedsof_upto 1.5 Mbpswithouttheneedfor

POTS splitters atthecustomer_premise,

Aware'

s

technology

is animportant steppingstone

betweenthe_analogmodemsavailable

today

andfull-rate ADSLat6Mbps.

Todd R. Engle December 1997

Page48 CapstoneThesis

[image:51.565.49.512.117.308.2]

Aware'

s modemcanbe pluggedinto any_existingphone

jack,

andno modificationto the

customer premise's_{internal wiring}isrequired.

By

_removingthe POTS splitter,implementation

and acceptanceof xDSL will_{be significantly} accelerated.

5.3.2 Hybrid High-Speed Modem

Untilrecently,xDSLmodemshave had several

limita