ETSF15CommunicationSystemsandNetworksStefanHöst

4GmobilecommunicationsystemLTE

Planoflecture

• Somehistoryofmobilenetworks—Pastandfuture• Ideaofnetworkstructure• EPCpacketnetworks• LTEradiochannel• Whatcomesnext?

2

Historyofmobilesystems

1st generation• Analogtransmissionwithnosecurity• Smallregions(countries)• FirstNTT(‘79),secondNMT(’81)• NMT(Nordic),AMPS(NA+Aus),TACS

(UK),TZ80x(NTT),DDI(Jap),Radiocom2000(Fr),RTMI(It)

• NMTshutdown2007

3

Historyofmobilesystems2nd generation• Digitalvoicechannel(10kb/s),Circuitswitched• News:SIMcard,SMS• 3December1992:“MerryChristmas”

• Largerregions(continents)• GSM(Eur),IS-136(N+SAm+Aus),

IS-95(NA+Asia),PDC(Jap)• GSM1991

2.5generation=>Datachannels• HSCSD,GPRS,EDGE,IS-95B

4

Historyofmobilesystems

5

3rd generation• Packetswitchedforvoiceanddata• 144kb/s– 3Mb/s• Global• CDMA2000(2000)• UMTS(2001)

• News:• UTRAN

UniversalTerrestrialRadioAccessNetwork

Historyofmobilesystems

6

4th generation• Packetswitcheddatatraffic

(VoiceoverIPor3G)• 100Mb/s-1Gb/s• Global:LTE• Dec2009(StockholmandOsloby

TeliaSonera)

• News:Smartphonesandapps• iPhone2007

Historyofmobilesystems

7

5th generation• Packetswitcheddatatraffic• 1-100Gb/s• 1Gb/s/user

• ≈2020• News:Smallcells,homecell,IoT (M2M),Car2Car,MassiveMIMO,etc

Mobiletraffic(CISCOVNI)

• Globalmobiledatatrafficgrew69%in2014• 1.5EBinDec’13to2.5EBinDec’14

• Globalmobiledevicesandconnectionsin2014grewto7.4billion(11.5Bin2019)• 26%smartdevicesgenerate80%oftraffic• 4Gdevicesgenerate10timesmoretrafficthannon-4G

8

TotalIPtraffic

9

© 2015 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public. Page 23 of 29

Figure 22. Global IP Traffic, Wired and Wireless

Source: Cisco VNI Global IP Traffic Forecast, 2014–2019

Narrowing the focus to Internet traffic and excluding managed IP traffic yields a more pronounced trend. By 2019, wired devices will account for 19 percent of Internet traffic, and Wi-Fi and mobile devices will account for 81 percent of Internet traffic (Figure 23). In 2014, wired devices accounted for less than half of Internet traffic, at 39 percent.

Figure 23. Global Internet Traffic, Wired and Wireless

Source: Cisco VNI Global IP Traffic Forecast, 2014–2019

59.9

168

TotalMobiletraffic

10

© 2015 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public. Page 10 of 42

Figure 6 shows the impact of the growth of mobile smart devices and connections on global traffic. Globally, smart traffic is going to grow from 88 percent of the total global mobile traffic to 97 percent by 2019. This percentage is significantly higher than the ratio of smart devices and connections (54% by 2019), because on average a smart device generates much higher traffic than a nonsmart device.

Figure 6. Effect of Smart Mobile Devices and Connections Growth on Traffic

Percentages refer to traffic share. Source: Cisco VNI Mobile, 2015

2.5

24.3

Growthofdevices

11

© 2015 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public. Page 8 of 42

Although there is an overall growth in the number of mobile devices and connections, there is also a visible shift in the device mix. This year we see a slight slowdown in the growth of tablets as a new device category, phablets (included in our smartphone category), were introduced. Tablet growth was also affected by the availability of lightweight laptops, which are quite similar to tablets in form factor but have more enhanced capabilities. Throughout the forecast period, we see that the device mix is getting smarter with an increasing number of devices with higher computing resources, and network connection capabilities that create a growing demand for more capable and intelligent networks. We define smart devices and connections as those having advanced computing and multimedia capabilities with a minimum of 3G connectivity. As mentioned previously, 497 million mobile devices and connections were added in 2014, and smartphones accounted for 88 percent of that growth at 439 million net adds. The share of smart devices and connections as a percentage of the total will increase from 26 percent in 2014 to more than half, at 54 percent, by 2019, growing 3.3-fold during the forecast period (Figure 4).

Figure 4. Global Growth of Smart Mobile Devices and Connections

Percentages refer to device and connections share. Source: Cisco VNI Mobile, 2015

This year we are also including Low Power Wide Area (LPWA) connections in our analysis. This wireless network connectivity is meant specifically for M2M modules that require low bandwidth and wide geographic coverage. Because these modules have very low bandwidth requirements and tolerate high latencies, we do not include them in the smart devices and connections category. For some regions, such as North America where the growth of LPWA is expected to be high, their inclusion in the mix would skew the smart devices and connections percentage, so for regional comparison we have taken them out of the mix. Figure 5 provides a comparable global smart-to-nonsmart devices and connections split, excluding LPWA.

7.4

11.5

Growthoftechnology

12

© 2015 Cisco and/or its affiliates. All rights reserved. This document is Cisco Public. Page 11 of 42

Trend 2: Defining Cellular Network Advances (2G, 3G, and 4G) Mobile devices and connections are not only getting smarter in their computing capabilities but are also evolving from lower-generation network connectivity (2G) to higher-generation network connectivity (3G, 3.5G, and 4G or LTE). When device capabilities are combined with faster, higher bandwidth and more intelligent networks, it leads to wide adoption of advanced multimedia applications that contribute to increased mobile and Wi-Fi traffic.

The explosion of mobile applications and phenomenal adoption of mobile connectivity by end users on the one hand and the need for optimized bandwidth management and network monetization on the other hand is fueling the growth of global 4G deployments and adoption. Service providers around the world are busy rolling out 4G networks to help them meet the growing end-user demand for more bandwidth, higher security, and faster connectivity on the move (Appendix B).

Globally, the relative share of 3G- and 3.5G-capable devices and connections will surpass 2G-capable devices and connections by 2017 (45 percent and 38 percent relative share). The other significant cross-over will occur in 2019, when 4G will also surpass 2G connection share. By 2019, 26 percent of all global devices and connections will be 4G-capable (Figure 7). The global mobile 4G connections will grow from 459 million in 2014 to 3 billion by 2019 at a CAGR of 46 percent.

Figure 7. Global Mobile Devices and Connections by 2G, 3G, and 4G

Percentages refer to device and connections share. Source: Cisco VNI Mobile, 2015

4G– LTELongTermEvolution

Standardizedby3GPP(3rd GenerationPartnershipProject)• RadioAccessNetworks(RAN)• Service&SystemsAspects(SA)• CoreNetwork&Terminals(CT)• GSMEDGERadioAccess

Networks(GERAN).

13

EPS(EvolvedPacketSystem)

14

SGW

MME HSS

PGW

eNodeB

Internet

E-UTRAN EPC

UE

EPSchallenges

• Highspeedradiolink• Bandwidthisextremelyexpensive.Squeezeoutallbitsyoucan

• AccesstoInternet• HowtosendIPpackets

• Authenticationandsecurity• Andacounting

• Usermobility• WhentochangeeNB andhowtotransferdatainthenetwork

• Costefficientuseofinfrastructure

15

EPSinterfacesEPC<->eNB• S1interface• SplitinS1-MMEandS1-U• Nocentralised nodeeNB <->eNB• X2interface• CoordinationandpositioningeNB <->UE• LTE-Uu orEUTRAN-Uu

16

EPC

LTE-Uu

E-UTRANEvolvedUniversalTerrestrialRadioAccessNetwork

RANresponsiblefor• Resourcemanagement

• Schedulinganddynamicresouce allocation

• Compression• IPheadcompressionreduces

overhead

• Security• Encryptionofdata

17

• Positioning• UEphysicalposition

• ConnectiontoEPC• Userandcontrolplane

Protocolstack

• S1oftentunneledoverthefixnetworkarchitecture• Atunnelisawaytosendpacketsoverothertypesof

network.E.g.• IPv4overIPv6andviceversa• IPoverIPsec

• GTP:GPRSTunnelingProtocol

18

ExampleoftunnelingIPv4overIPv6usingGRE(GenericRoutingEncapsulation)

19

IPpacketsinEPC(Userplane)

SGW• Collectingcharginginfo• LocalanchortowardseNB• Canchangeduetomobility

20

SGW

MME HSS

PGWeNB

PGW• IPedgeforuser• IPaddressallocationtoUE• QoS filtering• Mobilityanchor• Doesnotchangeduring

session.PreservestheIPaddr

S5/S8S1LTE-UuUE

IPpacketsinEPC(Usermobility)

21

SGW PGW

eNB

eNB

UE

SGW PGWeNB

UE

UE

Protocolstacks(Userplane)

22

GTP:GPRSTunnelingProtocolPDCP:PacketDataConvergenceProtocol

(IP<->Radio,Headercompression,security)RLC:RadioLinkControl

(Segmentation,reordering)

ControlplaneinEPC

HSS(HomeSubscriberServer)• Subscriberdatabase• SIMcardkeyexchange• AAA• Authentication,Authorization

andAccounting23

SGW

MME HSS

PGWeNB

MME(MobilityManagementEntity)• CommunicateswitheNB

andSGW• Managetunnelsand

encryption

S5/S8S1LTE-UuUE

USIMcardUMTSSubscriberIdentityModuleTheUSIMcardisanapplicationonasmartcardandcontains:• IMSI(Internationalmobilesubscriberidentity)15digits• AuthenticationkeyK andsequencenumberSQNAuthenticationprocess:

24

KASME isusedforencryptionofmessages

Theradiochannel

• Licensedfrequencybands(slotsof20MHz)• In0.5-4GHz

• CostinorderofBillions$• Alternative:FDDandTDD• MostcommonFDD

• Efficienttransmission• UElowpower=>long(er)batterylife

=>verylowpoweratreceiver25

Frequencyallocation(3kHz-300GHz)

26

2.4/5.8GHzFree(WiFi)LTE

eNodeBstructure(physicallayer)

27

OFDMError

controlFraming

interleaveDACAFE RF

shift

OFDMOrthogonalFrequencyDivisionMultiplexing

28

P2S

S2P

(a1,...,aN )

ai ∈Z16

16-QAM mapping:

(x1,..., xN ) (y1,..., yN ) = IFFT (

!x)

(yN−L+1,..., yN ,y1,..., yN )

QAM

QAM

QAM

IFFT

cos2π f0t

sin2π f0ts(t)

Re DAC

CP

Im DAC

Can be used with M=4, 16, 64 or 256

Time-frequecymultipleaccess

29

TDMA:

Time

Freq

uenc

y

…

User 1User 2User 3User 4

FDMA:

Time

Freq

uenc

y

…

Time-frequecymultipleaccess

30

OFDMA:

Time

Freq

uenc

y

…

User 1User 2User 3User 4

InLTE• AResourceElement(RE)isonecarrieroveroneOFDM

symbolThisistheleasttime-frequencyresolution• 15kHzX71.4us

Resourceblock(RB)

ApairofRB(subframe)istheleastassignedresource

31

Framestructureintime

32

Resourceallocationinfrequency

BW [MHz] 1.4 3 5 10 15 20

N 128 256 512 1024 1536 2048

#alloc RB (in Freq) 6 12 25 50 75 100

#used carrier 72 144 300 600 900 1200

Oversampling 1.78 1.78 1.71 1.71 1.71 1.71

Max Rate [Mb/s] 6 12 25 50 75 100

33

Max Rate is for 64-QAM, i.e. 6 bit/carrierThen [b/s]Impact of control signals and error control is not included

R = # carrier ⋅6 ⋅7 ⋅2 ⋅10 ⋅100

ExampleOneOFDMsymbolforBW=10MHzand64-QAM

34

ExampleOneOFDMsymbolforBW=10MHzand64-QAM

35

Plotof(all) frequencydomainsamplesasI-Q

4Gmobilenetworks

• PacketNetwork:EPC• Dataplane:PGW&SGW• Controlplane:MME&HSS• Incorenetwork

• AccessNetwork:LTE(E-UTRAN)• Upto20MHzbandwidth(=>100Mbps)• OFDMsignaling

36

Whatcomesnext1BBUhoteling

Basebandunit(BBU)• BinarydatatoBBsamples

• QAMmapping• IFFT• Coding

37

BBU EPCRU

eNodeB

Radiounit(RU)• BBsamplestoRFsignal

• Digital/analogconversion• MixingtoRF• AFE

CPRI

BBU hotel@ Data center

BBU

Whatcomesnext1BBUhotelingToutilizerecoursesbettersplitBBUandRRUinnetwork• UseCPRItotransmitBBsamplesbetweenBBUandRRU

38

BBU EPCRRU

RRU

RRU=RemoteRadioUnit

Problem:Datarateexpansionintransmissionofradiosamplesofapproximatelyafactor10

backhaulfronthaul

Whatcomesnext2Smallcells

• Smallcellsworkin• Shortdistance(lowpower)• Highdatarate(fewusers)

• Inpublicplaces,offices,shoppingmalls,etc

• Evenplansforsmallcellsinhomes• InsteadortogetherwithWiFi?

39

EPC

Internet

Problems• Howtobackhaul/fronhaul• Cantheybepartof

coordination?• Picocell:withcoordination• Femto cell:nocoordination

Whatcomesnext3Convergednetwork• Usersbecomesmoremobile

• Whenpriceandratedifferencesdiminishmostuserwillnotcareaboutchoiceofconnection

• Convergence:• Onenetwork,manyaccesses• Flexiblenetwork• OneAAA(andpricing)• Handoverbetweene.g.LTE-WiFi• Dualconnectionsandtrafficoff-loading• Allunitseverywhere

• Problem:• Partlynewnetworkstructure(e.g.commonIPedge)

40

Whatcomesnext3Convergednetwork

41

Fixedormobilenetwork(separated,SoA)

Fixedandmobilenetwork(convergednetwork)

Whatcomesnext45G

• Roughlyafactorof10inperformance• 10timesfaster,bandwidth,Datarate,etc• 10batterylife(forlowpowerdevices)• 1000timestrafficvolume• 10timesharder(atleast)

• Smallcellseverywhere• M2M;Everythingisconnected

• Car2Car,IoT,Skynet,…

• Expectedtolaunchlatest2020• Problems:

• Backhaul/fronthaul,RANsync,Energyefficiency,etc

42