Bits, Signals, and Packets

AnIntroductiontoDigitalCommunications&Networks
M.I.T.6.02LectureNotes
HariBalakrishnan
ChristopherJ.Terman
GeorgeC.Verghese
M.I.T.DepartmentofEECS
Lastupdate:November2012
MIT6.02DRAFTLectureNotes
Lastupdate:September13,2012
CHAPTER1
Introduction
Theabilitytodeliverandexchangeinformationovertheworldscommunicationnetworks
hasrevolutionizedthewayinwhichpeoplework,play,andlive.Attheturnofthecentury,
theU.S.NationalAcademyofEngineeringproducedalistof20technologiesthatmadethe
mostimpactonsocietyinthe20thcentury.
1
Thislistincludedlife-changinginnovations
suchaselectrication,theautomobile,andtheairplane;joiningthemwerefourtechnolog-
icalachievementsintheareaofcommununicationradioandtelevision, thetelephone, the
Internet, and computerswhose technological underpinnings we will be most concerned
withinthisbook.
Somewhat surprisingly, the Internet came in only at #13, but the reason given by the
committeewasthatitwasdevelopedtowardthelatterpartofthecenturyandthatthey
believedthemostdramaticandsignicantimpactsoftheInternetwouldoccurinthe21st
century.Lookingattherstdecadeofthiscentury,thatsentimentsoundsrighttheubiq-
uitousspreadofwirelessnetworksandmobiledevices,theadventofsocialnetworks,and
theabilitytocommunicateanytimeandfromanywherearenotjustchangingthefaceof
commerceandourabilitytokeepintouchwithfriends,butareinstrumentalinmassive
societalandpoliticalchanges.
Communicationisfundamentaltoourmodernexistence.Itishardtoimaginelifewith-
outtheInternetanditsapplicationsandwithoutsomeformofnetworkedmobiledevice.
Inearly2011,over5billionmobilephoneswereactiveworldwide,overabillionofwhich
had broadband network connectivity. To put this number in perspective, it is larger
thanthenumberofpeopleintheworldwhoin2011hadelectricity,shoes,toothbrushes,
ortoilets!
2
1.1 Objectives
Whatmakesourcommunicationnetworkswork?Thisbookisastartatunderstandingthe
answerstothisquestion.Thisquestionisworthstudyingfortworeasons.First,tounder-
1
TheVertiginousMarchofTechnology,obtainedfromnae.edu.Documentathttp://bit.ly/owMoO6
2
ItisinfactdistressingthataccordingtoarecentsurveyconductedbyTeleNavandwecanttellifthis
is a joke40% of iPhone users say theyd rather give up their toothbrushes for a week than their iPhones!
http://www.telenav.com/about/pr-summer-travel/report-20110803.html
1
2 CHAPTER 1. INTRODUCTION
standthekeydesignprinciplesandbasictechniquesofanalysisusedincommunication
systems. Second,becausethetechnicalideasinvolvedalsoariseinseveralothereldsof
computer science (CS) and electrical engineering (EE), the study of communication sys-
temsprovidesanexcellentcontexttointroduceconceptsthataremorewidelyapplicable.
Beforewediveinanddescribethetechnicaltopics,wetoshareabitofthephilosophy
behind the material and approach used in this book. The material is well-suited for a
one-semestercourseonthetopic;atMIT,suchacourseistaken(mostly)bysophomores
whose background includes some basic programming (for the accompanying labs) and
someexposuretoprobabilityandtheFourierseries.
Traditionally,inbotheducationandinresearch,muchoflow-levelcommunicationhas
been considered an EE topic, covering primarily the issues governing how information
movesacrossasinglecommunicationlink. Inasimilarvein, muchofnetworkinghas
beenconsideredaCStopic,coveringprimarilytheissuesofhowtobuildcommunication
networks composed of multiple links. In particular, many traditional courses on digi-
talcommunicationrarelyconcernthemselveswithhownetworksarebuiltandhowthey
work, while most courses on computer networks treat the intricacies of communication
over physical links as a black box. As a result, a sizable number of people have a deep
understandingofoneortheothertopic,butfewpeopleareexpertineveryaspectofthe
problem.Thisdivisionisonewayofconqueringtheimmensecomplexityofthetopic.Our
goalinthisbookistounderstandtheimportantdetailsofboththeCSandEEaspectsof
digitalcommunications,andalsotounderstandhowvariousabstractionsallowdifferent
partsofthesystemtobedesignedandmodiedwithoutpayingcloseattention(oreven
fullyunderstanding)whatgoesonelsewhereinthesystem.
Onedrawbackofpreservingstrongboundariesbetweendifferentcomponentsofacom-
municationsystemisthatthedetailsofhowthingsworkinanothercomponentmayre-
mainamystery, eventopractisingengineers. Inthecontextofcommunicationsystems,
this mystery usually manifests itself as things that are above my layer or below my
layer. Andsoalthoughwewillappreciatethebenetsofabstractionboundariesinthis
book,animportantgoalforusistostudythemostimportantprinciplesandideasthatgo
intothecompletedesignofacommunicationsystem. Ourgoalistoconveytoyouboth
thebreadthoftheeldaswellasitsdepth.
We cover communication systems all the way from the source, which has some infor-
mationitwishestotransmit,topackets,whichmessagesarebrokenintofortransmission
overanetwork,tobits,eachofwhichisa0ora1,tosignals,whichareanalogwave-
formssentoverphysicalcommunicationlinks(suchaswires,ber-opticcables,radio,or
acoustic waves). We study a range of communication networks, from the simplest dedi-
catedpoint-to-pointlink,tosharedmediacomprisingasetofcommunicatingnodessharing
acommonphysicalcommunicationmedium,tolargermulti-hopnetworksthatthemselves
areconnectedtoothernetworkstoformevenbiggernetworks.
1.2 Themes
Three fundamental challenges lie at the heart of all digital communication systems and
networks: reliability,sharing,andscalability. Wewillspendaconsiderableamountoftime
onthersttwoissuesinthisintroductorycourse,butmuchlesstimeonthethird.
3 SECTION 1.2. THEMES
1.2.1 Reliability
Alargenumberoffactorsconspiretomakecommunicationunreliable,andwewillstudy
numeroustechniquestoimprovereliability. Acommonthemeacrossthesedifferenttech-
niquesisthattheyalluseredundancyincreativeandefcientwaystoprovidereliabilityus-
ingunreliableindividualcomponents,usingthepropertyofindependent(orperhapsweakly
dependent)failuresoftheseunreliablecomponentstoachievereliability.
Theprimarychallengeistoovercomeawiderangeoffaultsanddisturbancesthatone
encountersinpractice,includingGaussiannoiseandinterferencethatdistortorcorruptsig-
nals, leading to possible bit errors that corrupt bits on a link, to packet losses caused by
uncorrectable bit errors, queue overows, or link and software failures in the network. All
theseproblemsdegradecommunicationquality.
Inpractice,weareinterestednotonlyinreliability,butalsoinspeed.Mosttechniquesto
improvecommunicationreliabilityinvolvesomeformofredundancy,whichreducesthe
speedofcommunication. Theessenceofmanycommunicationsystemsishowreliability
andspeedtradeoffagainstoneanother.
Communication speeds have increased rapidly with time. In the early 1980s, people
wouldconnecttotheInternetovertelephonelinksatspeedsofbarelyafewkilobitsper
second,whiletoday100Megabitspersecondoverwirelesslinksonlaptopsand1-10Gi-
gabitspersecondwithwiredlinksarecommonplace.
Wewilldevelopgoodtoolstounderstandwhycommunicationisunreliableandhow
toovercometheproblemsthatarise. Thetechniquesinvolveerror-correctingcodes,han-
dlingdistortionscausedbyinter-symbolinterferenceusingalineartime-invariantchan-
nelmodel,retransmissionprotocolstorecoverfrompacketlossesthatoccurforvariousrea-
sons, anddevelopingfault-tolerantroutingprotocolstondalternatepathsinnetworksto
overcomelinkornodefailures.
1.2.2 Ecient Sharing
Anengineercandoforadimewhatanyfoolcandoforadollar,accordingtofolklore.A
communicationnetworkinwhicheverypairofnodesisconnectedwithadedicatedlink
wouldbeimpossiblyexpensivetobuildforeven moderatelysizednetworks. Sharingis
thereforeinevitableincommunicationnetworksbecausetheresourcesusedtocommuni-
catearentcheap.Wewillstudyhowtoshareapoint-to-pointlink,asharedmedium,and
anentiremulti-hopnetworkamongmultiplecommunications.
Wewilldevelopmethodstoshareacommoncommunicationmediumamongnodes,a
problemcommontowiredmediasuchasbroadcastEthernet,wirelesstechnologiessuch
aswirelesslocal-areanetworks(e.g., 802.11orWiFi), cellulardatanetworks(e.g., 3G),
andsatellitenetworks(seeFigure1-1).
Wewillstudymodulationanddemodulation,whichallowustotransmitsignalsover
different carrier frequencies. In the process, we can ensure that multiple conversations
shareacommunicationmediumbyoperatingatdifferentfrequencies.
We will study medium access control (MAC) protocols, which are rules that determine
hownodesmustbehaveandreactinthenetworkemulateeithertimesharingorfrequency
sharing. Intimesharing,eachnodegetssomedurationoftimetotransmitdata,withno
othernodebeingactive. Infrequencysharing,wedividethecommunicationbandwidth
4 CHAPTER 1. INTRODUCTION
Figure1-1:Examplesofsharedmedia.
Individual images source unknown. All rights reserved. This content is excluded from
our Creative Commons license. For more information, see http://ocw.mit.edu/fairuse.
(i.e., frequency range) amongst the nodes in a way that ensures a dedicated frequency
sub-rangefordifferentcommunications,andthedifferentcommunicationscanthenoccur
concurrentlywithoutinterference.Eachschemehasitssweetspotanduses.
Wewillthenturntomulti-hopnetworks. Inthesenetworks,multipleconcurrentcom-
municationsbetweendisparatenodesoccurbysharingoverthesamelinks. Thatis,one
mighthavecommunicationbetweenmanydifferententitiesallhappeningoverthesame
physicallinks.Thissharingisorchestratedbyspecialcomputerscalledswitches,whichim-
plementcertainoperationsandprotocols. Multi-hopnetworksaregenerallycontrolledin
distributedfashion,withoutanycentralizedcontrolthatdetermineswhateachnodedoes.
Thequestionswewilladdressinclude:
1. Howdomultiplecommunicationsbetweendifferentnodessharethenetwork?
2. Howdomessagesgofromoneplacetoanotherinthenetwork?
3. Howcanwecommunicateinformationreliablyacrossamulti-hopnetwork(asop-
posedtooverjustasinglelinkorsharedmedium)?
Thetechniquesusedtosharethenetworkandachievereliabilityultimatelydetermine
theefciencyofthecommunicationnetwork.Ingeneral,onecanframetheefciencyques-
tion in several ways. One approach is to minimize the capital expenditure (hardware
equipment, software, link costs) and operational expenses (people, rental costs) to build
andrunanetworkcapableofmeetingasetofrequirements(suchasnumberofconnected
devices, levelofperformanceandreliability, etc.). Anotherapproachistomaximizethe
bangforthebuckforagivennetworkbymaximizingtheamountofusefulworkthatcan
bedoneoverthenetwork.Onemightmeasuretheusefulworkbycalculatingtheaggre-
gatethroughput(inbitspersecond,orathigherspeeds,themoreconvenientmegabits
persecond)achievedbythedifferentcommunications,thevariationofthatthroughput
among the set of nodes, and the average delay (often called the latency, measured usu-
allyinmilliseconds)achievedbythedatatransfers. Wewillprimarilybeconcernedwith
throughputandlatencyinthiscourse, andnotspendmuchtimeonthebroader(butno
lessimportant)questionsofcost.
5 SECTION 1.3. OUTLINE AND PLAN
Of late, another aspect of efciency that has become important in many communica-
tionsystemsisenergyconsumption. Thisissueisimportantbothinthecontextofmassive
systemssuchaslargedatacentersandformobilecomputingdevicessuchaslaptopsand
mobilephones.Improvingtheenergyefciencyofthesesystemsisanimportantproblem.
1.2.3 Scalability
Inadditiontoreliabilityandefcientsharing,scalability(i.e.,designingnetworksthatscale
tolargesizes)isanimportantdesignconsiderationforcommunicationnetworks. Wewill
onlytouchonthisissue,leavingmostofittolatercourses(6.033,6.829).
1.3 Outline and Plan
Wehavedividedthecourseintofourparts: thesource,andthethreeimportantabstrac-
tions(bits,signals,andpackets). Forpedagogicreasons,wewillstudythemintheorder
givenbelow.
1.Thesource. Ultimately,allcommunicationisaboutasourcewishingtosendsome
informationintheformofmessagestoareceiver(ortomultiplereceivers). Hence,
itmakessensetounderstandthemathematicalbasisforinformation,tounderstand
howtoencodethematerialtobesent,andforreasonsofefciency,tounderstandhow
besttocompressourmessagessothatwe cansendaslittledataaspossiblebutyet
allowthereceivertodecodeourmessagescorrectly. Chapters2and3describethe
keyideasbehindinformation,entropy(expectationofinformation),andsourcecoding,
whichenablesdatacompression.WewillstudyHuffmancodesandtheLempel-Ziv-
Welchalgorithm,twowidelyusedmethods.
2.Bits. The main issue we will deal with here is overcoming bit errors using error-
correcting codes, specically linear block codes (Chapters 5 and 6) and convolutional
codes(Chapters7and8). Thesecodesuseinterestingandsophisticatedalgorithms
thatcleverlyapplyredundancytoreduceoreliminatebiterrors.
3.Signals.Themainissueswewilldealwithherearehowtomodulatebitsoversignals
anddemodulatesignalstorecoverbits, aswellasunderstandinghowdistortionsof
signalsbycommunicationchannelscanbemodeledusingalineartime-invariant(LTI)
abstraction. Topicsincludegoingbetweentime-domainandfrequency-domainrep-
resentationsofsignals,thefrequencycontentofsignals,andthefrequencyresponse
ofchannelsandlters.Chapters9through14describethesetopics.
4.Packets.ThemainissueswewillstudyarehowtoshareamediumusingaMACpro-
tocol,routinginmulti-hopnetworks,andreliabledatatransportprotocols.Chapters
15through19describethesetopics.
MIT OpenCourseWare
http://ocw.mit.edu
6.02 Introduction to EECS II: Digital Communication Systems
Fall 2012
For information about citing these materials or our Terms of Use, visit: http://ocw.mit.edu/terms.