On accurate simulations of LTE femtocells using an open source simulator

Description

Long-term evolution (LTE) femtocells represent a very promising answer to the ever growing bandwidth demand of mobile applications. They can be easily deployed without requiring a centralized planning, to provide high data rate connectivity with a limited coverage. In this way, the overall capacity of the cellular network can be greatly improved. At the same time, the uncoordinated setup of femtocells poses new issues that require a deep and thorough analysis before spreading this technology worldwide. Unfortunately, to the best of our knowledge, no accurate simulation tools are freely available for enabling this kind of investigation. Thus, we present in this study a simulation tool for LTE femtocells, implemented as a module of the emerging open source LTE-sim framework. It encompasses heterogeneous scenarios with both macro and femtocells, spectrum allocation techniques, user mobility, femtocell access policies, and several other features related to this promising technology. After reviewing the status of the art on LTE femtocells, we detail the description of the module that we propose with a major emphasis on the newly devised loss models for indoor scenarios, the new network topology objects, and the most significant enhancements to the simulator protocol stack. Furthermore, to provide a clear understanding of the practical utility of this new simulator, we investigate two indoor and urban scenarios. A scalability test is also presented to demonstrate the efficiency of the proposed tool in terms of processing requirements. All presented results suggest that this new module can be very interesting for the research community, due to its great flexibility and limited computational cost.

Subjects

• Documents
• Knowledge
• 3GPP
• LTE
• E-UTRA
• Network simulator
• Performance Evaluation

Informations

• Published by: biomed
• Published: 01 January 2012
• Reads: 23
• Language: English
• Document size: 1 MB

Exrait

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

RESEARCH

OpenAccess

OnaccuratesimulationsofLTEfemtocells
usinganopensourcesimulator
FrancescoCapozzi
1*
,GiuseppePiro
2
,LuigiAGrieco
2
,GennaroBoggia
2
andPietroCamarda
2

Abstract
Long-termevolution(LTE)femtocellsrepresentaverypromisinganswertotheevergrowingbandwidthdemandof
mobileapplications.Theycanbeeasilydeployedwithoutrequiringacentralizedplanning,toprovidehighdatarate
connectivitywithalimitedcoverage.Inthisway,theoverallcapacityofthecellularnetworkcanbegreatlyimproved.
Atthesametime,theuncoordinatedsetupoffemtocellsposesnewissuesthatrequireadeepandthoroughanalysis
beforespreadingthistechnologyworldwide.Unfortunately,tothebestofourknowledge,noaccuratesimulation
toolsarefreelyavailableforenablingthiskindofinvestigation.Thus,wepresentinthisstudyasimulationtoolforLTE
femtocells,implementedasamoduleoftheemergingopensourceLTE-simframework.Itencompasses
heterogeneousscenarioswithbothmacroandfemtocells,spectrumallocationtechniques,usermobility,femtocell
accesspolicies,andseveralotherfeaturesrelatedtothispromisingtechnology.Afterreviewingthestatusofthearton
LTEfemtocells,wedetailthedescriptionofthemodulethatweproposewithamajoremphasisonthenewlydevised
lossmodelsforindoorscenarios,thenewnetworktopologyobjects,andthemostsignificantenhancementstothe
simulatorprotocolstack.Furthermore,toprovideaclearunderstandingofthepracticalutilityofthisnewsimulator,
weinvestigatetwoindoorandurbanscenarios.Ascalabilitytestisalsopresentedtodemonstratetheefficiencyofthe
proposedtoolintermsofprocessingrequirements.Allpresentedresultssuggestthatthisnewmodulecanbevery
interestingfortheresearchcommunity,duetoitsgreatflexibilityandlimitedcomputationalcost.
Keywords:
Femtocells,3GPP,LTE,E-UTRA,Networksimulator,Performanceevaluation

1Introduction
theyensurethemaximumachievablephysicalrate.On
Thecapillarydiffusionofsmartphonesandtabletsandtheotherside,theamountofspectrumexpectedtobe
theintegrationofconnectivitycapabilitiesincommonlifelicensedduringnextyearsislessthanwhatrequestedby
objects(suchasTVs,domesticappliances,andvehicles)mobileoperators[4].Accordingtothisvision,thedeploy-
aregeneratingaveryfastgrowthofbandwidthdemandmentofsmallcellswithinthetypicalstructureofcellular
formobileapplications[1].Despiteemergingbroadbandnetworks,representsthesmartestsolutionforboosting
technologies(e.g.,WiMAX[2]andLTE[3])areabletoservicesinthesesystems[5].
enhancetheperformanceofcurrentlyused3Gsystems,Followingthisdirection,the3GPPhasintroduced,
theymightnotbeabletosustaintheexpectedraiseofthewithinthelong-termevolution-advanced(LTE-A)specifi-
trafficvolume.cations,thepossibilitytodeployaheterogeneousnetwork
Asaconsequence,notableimprovementsinbothcapac-(HetNet)composedbymacroandsmall-range(i.e.,micro,
ityandcoverageofmobilecommunicationsystemsarepico,andfemto)cells[6].Microandpicocellscouldbe
required,buttheycannotbefulfilledbyonlyenhancingexploitedforenhancingcoverageandcapacityinsome
the(PHY)layerand/orincreasingtheavailablespectrum.regionsinsidethemacrocell.Whereas,femtocellshave
Fromoneside,infact,latestinnovativePHYtechniquesbeendevisedforofferingbroadbandservicesinindoor
workveryclosetotheShannonlimit;inthismanner,(i.e.,homeandoffices)andoutdoorscenarioswithavery
limitedgeographicalcoverage.Amongthiskindofcells,
*Correspondence:[e-mail]
theroleoffemtocellsbecomefundamental,becauseitis
1ITIA-CNR,v.PLembo,38F-70124,Bari,Italy
Fulllistofauthorinformationisavailableattheendofthearticle

©2012Capozzietal.;licenseeSpringer.ThisisanOpenAccessarticledistributedunderthetermsoftheCreativeCommons
AttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,andreproduction
inanymedium,providedtheoriginalworkisproperlycited.

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

expectedthatinupcomingyearsthemostpartofvoice
callsanddatasessionswilltakeplaceinhomeorbusiness
environments[7].
Afemtocellcanbeeasilysetupwithoutanycentral-
izedcoordination,butsimplyenablingalow-powerand
small-rangeradiobasestation,whichisreferredtoas
homeevolvedNodeB(HeNB).Suchadevicehasplug-
and-playcapabilities,isconnectedtothecorenetwork
throughaDSLline,andoperatesinthespectrumlicensed
forcellularsystems[8].Theuncoordinatednatureof
femtocelldeploymentposesnovelandinterestingchal-
lengesonradioresourcemanagement(RRM).Infact,
classicalapproachesalreadyadoptedin3Gsystemsto
facefrequencyplanning,interferencecoordination,radio
resourcescheduling,andaccesspoliciescouldbenot
usefulanymoreinHetNetscenarios.
Startingfromthispremise,itisevidentthatthedevelop-
mentofnovelnetworkarchitecturesandtheoptimization
ofnextgenerationcellularsystemsbasedonfemtocells
aretopicsworthofinvestigationforbothindustryand
academiccommunities.
Inthisperspective,theavailabilityofaccuratesimu-
lationframeworksappearsoffundamentalimportance.
Severalopen-sourceandcommercialtoolsaretodayavail-
ableforsimulatingsomepartsoftheLTEsystem[9-14],
butnoneofthemprovidesmediaaccesscontrol(MAC)
andPHYmodelsforfemtocellarchitectures,handover
strategiesproperlyconceivedforHetNet,andspecific
propagationlossmodelsforindoorenvironments.Asys-
temlevelsimulatorforLTEfemtocellshasbeenproposed
in[15].However,itscodeisnotyetavailableand,forthis
reason,itcannotadequatelyservetheresearchcommu-
nity.Atthepresent,tothebestofourknowledge,the
onlyvaluabletoolistheonedevelopedforthenetwork
simulator3(NS-3)withintheLENAproject[16].Unfor-
tunately,despiteitprovidesstandardcompliantdefinition
oftheLTEnetworkarchitectureaswellasseveralprop-
agationlossmodelsforindoorscenarios,atthecurrent
statusitcannotbesufficientinhelpingresearchersdur-
ingthestudyoffemtocellrelatedissues.Infact,itdoes
notimplementanyhandoverstrategiesforHetNet,itdoes
notallowtheapplicationofdifferentaccesspoliciesto
theHeNB,itoffersonlyaweaksupportforthequal-
ityofservice(QoS)management[17](e.g.,onlysimple
schedulingstrategies,suchasMaximumThroughputand
ProportionalFair,havebeendevelopedattheMAClayer).
Inaddition,thesetupofarealisticsimulationscenariois
difficulttoaccomplish,especiallyforanon-expertNS-3
.ersuTobridgethisgap,thepresentstudyproposesamodule
thatextendsthefunctionsoftheemergingopensource
LTE-simframework[18],thusallowingalsothesimu-
lationofscenarioswithLTEfemtocells
a
.Itprovidesa
widerangeoffeaturesandhighscalability.Itencompasses

Page2of13

scenarioswithbothmacroandfemtocells,inmulti-cell
andmulti-userenvironments,andimplementsacomplete
LTEprotocolstackalongwithadvancedRRMtechniques,
suchasfrequencyreuseschemes,packetscheduling,and
QoSmanagement.Usermobilityisalsoconsideredand
severalconfigurationsofmobilitypatternsandhandover
mechanismsareconsidered.
Webelievethatthedevelopedtoolrepresentsagood
andusefulresearchtoolfortacklingseveralopenissues
suchas:(i)thedeploymentofheterogeneousLTEnet-
workscomposedbymacroandfemtocells;(ii)thedesign
ofsophisticatedtechniquesforbothradioresourceand
interferencemanagement;(iii)theimplementationof
novelcognitive-basedand/orself-organizedalgorithms,
protocols,andproceduresforthesetupofnextgeneration
broadbandnetworks,andsoon.
Inordertoshedsomelightonthepracticalutilityof
theproposedtool,inthisarticlewereportsomeexamples
aboutitsapplication.Inparticular,somereferenceindoor
andurbanscenariosareevaluated.Itisworthtonote
thatsuchananalysisisonlyintendedtodemonstratethe
generalagreementbetweenwhatisexpectedfromathe-
oreticalpointofviewandtheoutcomesofthesimulation
modulewepropose.
Furthermore,ascalabilityanalysistomeasurethecom-
putationalrequirementsofthedevelopedtool(i.e.,simu-
lationtimeandmemoryusage)hasbeenalsoprovided.
Therestofthearticleisorganizedasfollows:Section2
describesLTEfemtocells,highlightingprosandcons
relatedtothedevelopmentofthisnewtechnologyand
themostimportantopenissuesthatjustifytheneedofa
simulationframework.InSection3,thedevelopedmod-
uleisdescribed,withparticularemphasisonthenewly
implementedpropagationlossmodelsforindoorscenar-
ios,theintroductionofnewnetworktopologyobjects,
andtheenhancementsofsomemodulesoftheLTE-sim
framework.Section4providessomereferenceresultson
significantscenariosinordertoprovideaclearassessment
ofthepracticalutilityofthesimulatorandtodemon-
strateitsscalability.Finally,Section5drawsconclusions
andforecastsfutureworks.
2LTEfemtocells
TheLTEsystemismainlycomposedbytwoparts:theair
interface,i.e.,theevolved-universalterrestrialradioaccess
network(E-UTRAN),andthepacketswitchedcorenet-
work,knownasEvolvedPacketCore.Fromthenetwork
side,theevolvedNodeB(eNB)istheonlynodeofthe
E-UTRANanditisinchargeofprovidingnetworkcon-
nectivitythroughtheairinterfacetoalluserequipments
(UEs)inthecell,accordingtotheclassiccellularnetwork
paradigm.
Atthephysicallayer,theradiointerfacesupportsboth
frequencyandtimedivisionsduplexing.Channelaccess,

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

instead,isbasedonorthogonalfrequencydivisionmul-
tipleaccess(OFDMA),whichprovideshighflexibilityin
termsofschedulingandinterferencemanagement[3].
Accordingto[19],radioresourcesareallocatedina
time/frequencydomain.Inthetimedomain,theyare
distributedeverytransmissiontimeinterval(TTI),each
onelasting1ms.Furthermore,eachTTIiscomposedby
twotimeslotsof0.5ms,correspondingtosevenOFDM
symbolsinthedefaultconfigurationwithshortcyclic
prefix;10consecutiveTTIsformtheLTEFramelasting
10ms.Inthefrequencydomain,instead,thewholeband-
widthisdividedinto180kHzsub-channels,correspond-
ingto12consecutiveandequallyspacedsub-carriers.A
time/frequencyradioresource,spanningoveronetime
slotlasting0.5msinthetimedomainandoveronesub-
channelinthefrequencydomain,iscalledresourceblock
(RB)andcorrespondstothesmallestradioresourcethat
canbeassignedtoanUEfordatatransmission.Notethat,
duetothefixedsub-channelsize,thenumberofsub-
channelsvariesaccordinglytodifferentsystembandwidth
configurations(e.g.,25and50RBsforsystembandwidths
of5and10MHz,respectively).
In[20],the3GPPintroducedanewlow-powerand
small-rangeradiobasedstation,i.e.,theHeNB,forpro-
vidingbroadbandservicesinindoorandoutdoorenviron-
ments.Suchadeviceisconnectedtotheoperatornetwork
throughaDSLlineavailableatconsumers’housesor
offices,likeacommonWi-Fiaccesspoint.Thevery
limitedgeographicalareaitcoversiscalled
femtocell
(Figure1).
Severalbenefitsareexpectedfromfemtocelldeploy-
ment.Firstofall,networkcapacitycanbeincreased.
Accordingto[5],OFDMA-basedtechnologies,suchas
LTE,workveryclosetotheShannonLimit,andtheonly
waytosurelyincreasethecapacityofasinglewireless
linkistoputthetransmitterandthereceivercloserto
eachother,asintherationaleofLTEfemtocells.Inthis

macro cellfemtocell
E-UTRAN interfacecorenetworkE-UTRAN interface
mobile
operator’s
CPEEUBeNUEIP backhualHeNB
LSDFigure1
Networkarchitecturemacroandfemtocells.

Page3of13

way,itisalsopossibletosignificantlyoffloadthemacro-
cellduetotherelevantquotaoftrafficthatfemtocellsare
expectedtoserve.Inaddition,fromaneconomicalpoint
ofview,sinceanoperatorcanaccuratelyidentifythetraffic
generated/receivedineachfemtocell,itcanofferperson-
alizedfeesanddiscountsthatcanbeveryattractivefor
consumers.
2.1Openissues
Aforementionedbenefitscomeatthecostofincreased
complexityinspectrummanagement.Beingunfeasiblea
centralizedfrequencyplanning,infact,theinterference
managementbecomesoneofthemainissuesandnew
solutionsmustbefoundtothisregard.
Thereareseveralinterestingproblemsthatcouldarise
withawidedeploymentofLTEfemtocells,whichhave
tobecarefullyaccountedforinordertoexploitallthe
potentialofthisnewpromisingtechnology.
2.1.1Howtocopewithuncoordinatedfemtocell
deployment?
AnuncoordinatedsetupofHeNBs,carriedoutwithout
takingintoaccountthelocationofmacrocellsandthe
bandwidthallocationplan,bringstoanincrementofthe
interferencelevel.Intheseconditions,amobileoperator
couldnotbeabletoeasilyapplyoptimizedRRMproce-
dures.Wecanidentifytwokindsofinterferences:
co-layer
and
cross-layer
[21].TheformerisproducedbyaHeNB
onusersservedbyotherfemtocells.Thelatter,instead,
definestheinterferencebetweenfemtocellsandmacro-
cellsthatsharethesameportionofthespectrum.Inthis
context,theresearchcommunityisdeeplyinvestigating
innovativesolutionsformitigatingtheinter-cellinter-
ference,furnishingsuchcellswithself-organizationand
cognitivecapabilities[22].
Thebasicideaisthatspectrumassignmentshouldbe
performedonadistributedbasis,byallowingeachHeNB
tomeasuretheradioenvironmentandtoautonomously
identifythebestfrequencybandsthatminimizethe
impactoftheinter-cellinterference.

2.1.2Femtocellsaccesspolicies:openor/andrestricted
accessmode?
Sincefemtocellsareexpectedtobedeployedinseveral
environments(e.g.,home,business,andoutdoorscenar-
ios)bybothmobileoperatorsandconsumers,newissues
regardingthenetworkaccesspolicyhavetobeconsidered.
AtypicalHeNBdeployedbyamobileoperatorcanbe
setupforworkinginopenaccessmode(i.e.,anyconnec-
tionrequestcanbeaccepted).Inthisway,anyuser,located
initscoveragearea,cantakeadvantagefromthehigher
capacityofferedbythefemtocell.
Fromanotherside,afemtocell,installedbyaconsumer
fordomesticuses,hastooperateinrestrictedaccess

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

mode:theaccessisallowedonlytoalimitednumber
ofdevices.Inthiscase,unauthorizeduserscouldsuffer
theimpactoftheinterferencegeneratedbythis“private”
HeNB[23].
Agoodtradeoffbetweentheaforementionedapproa-
chesisrepresentedbythe“hybrid”accessmethod:the
HeNBisinchargeofguaranteeingahighperformance
connectiononlytoalimitednumberofregisteredusers.
Atthesametime,itcouldassignacertainamountofradio
resourcestootherusers,offeringthemaminimumservice
level[24].Apracticalexampleofthispolicyistheworld
widespreadFONmodel,whichhasbeenfirstlyconceived
fortheWiFitechnology[25].
2.1.3Radioresourcemanagement
Differentlyfrom3Gnetworks,LTEsystemsarebasedon
aflattenedarchitecture,withthemainconsequencethat
alltheRRMprocedureshavetobehandledbytheeNB.
Thisaspectbecomesverycriticalforfemtocellsduetothe
limitedcomputationalcapabilitiesofHeNBs:forinstance,
therearenoguaranteesthatsophisticatedscheduling
strategies,suchasthoseproposedin[26-28],couldbe
straightlyadoptedinLTEfemtocells.Asaconsequence,
theresearchoflightweightRRMtechniquesremainsfully
opentonovelapproachesandmethodologies.
2.1.4Howtohandleusermobility?
Anothertypicalproblemoffemtocelldeploymentis
strictlyconnectedtousermobility:increasingthenum-
berofcellswithinthesamearealeadstomoreandmore
handoveroperations.Thisaspectcanbecomeverycrit-
icalbecause,despitebasestationsareconnectedamong
themusingadedicatedprotocolsuite(i.e.,throughthe
X2interface),LTEalwaysrequireshardhandover,which
is,ingeneral,aresource-demandingprocedure.Alsoin
thiscase,theresearchofnovelprocedurestocounteract
theimpactofhandoverswouldbehighlybeneficialforthe
deploymentofLTEfemtocells[29].Withinthiscontext,
somestudyhasbeencarriedout(see,forinstance,[30]),
butwithoutanyexplicitreferencetofemtocells.
2.1.5Lackingofamodelingframework
Alltheaforementionedopenissuesareconnectedtothe
lackofanunifiedmodelingapproachtoLTEfemtocells.
Mainlyduetotheabsenceoffinalspecificationdocuments
from3GPP,thismeansthatacommonprocedureand
consequentlyacommonmodelingframeworktoletthe
researchersproperlycomparetheirresultsaremissing.
Therefore,weclaimthataflexiblesimulationplatform
willsupportwellalltheactivitiesofresearchers,indus-
tries,andstandardizationbodiesthatintendtodesign
newnetworkarchitecture,protocols,andalgorithmsfor
nextgenerationbroadbandsystems.

Page4of13

3Theproposedmodule
Themoduleweproposehasbeenimplementedwithinthe
emergingopensourceframeworkLTE-sim[18].Inwhat
follows,wewillbrieflysummarizeLTE-simkeyaspects
andwewilldescribethedetailsofthenewlydeveloped
features.
3.1BasicbackgroundandrequiredupgradesonLTE-sim
LTE-simisanevent-drivensimulatorwritteninC++using
thewell-knownobject-orientedparadigm.Itencompasses
severalaspectsofLTEnetworks,includingthemodelsof
boththeE-UTRANandtheevolvedpacketsystem,down-
linkanduplinktransmissions,singleandmulti-cellenvi-
ronments,QoSmanagement,multiusersenvironment,
usermobility,handoverprocedures,andfrequencyreuse
techniques.Furthermore,theentireLTEprotocolstack
isframedfromtheapplicationtothePHYlayer,includ-
ingradioresourcecontrol(RRC),radiolinkcontrol,and
MACentities.Thisisdoneforthethreekindsofnet-
worknodes,thatis,UE,eNB,andMobilityManagement
Entity/Gateway.Inaddition,thesoftwaresupportsalso
well-knownschedulingstrategies(suchasProportional
Fair,ModifiedLargestWeightedDelayFirst,andExpo-
nentialProportionalFair,LogandExprules),Adaptive
modulationandcodingschemescheme,channelqual-
ityindicator(CQI)feedback,andseveralotheraspects
relatedtotheLTEtechnology.
ThehighflexibilityandmodularityofLTE-simallowed
ustodeviseacompletesystemforsimulatingLTEfemto-
cells,builtontopofexistingfeaturessuchasapplication
objects,tracing,interactionamongmacrocells,interfer-
encecomputation,mobility,handoverprocedures,and
oos.nThenewmodulerequiredtheupgradeandtheaddition
ofclassesandfunctionssuchas:(i)theHeNBnetwork
device;(ii)novelRRCandMACentities;(iii)dedicated
networktopologyelements(e.g.,femtocell,buildings,and
streets);and(iv)newchannelmodels.
InlinewithLTE-simdesigncriteria,theproposed
extensionhasbeenfreelyreleasedundertheGPLv3
license[31].Infact,webelievethattheopennatureofthis
softwarewouldfacilitatethecross-validationandtheinte-
grationofdifferentapproachesproposedbyresearchers
andpractitionersworkinginthisfield,thusacceler-
atingtheprogressoftheknowledgeinthisscientific
domain.
3.2Networkdevices
TheHeNBdevicehasbeencreatedformodelingthebase
stationofafemtocell.SimilarlytotheeNB,itisidenti-
fiedbyanuniqueIDanditspositionisdefinedintoa
Cartesiansystem.Itkeepstrackofinformationrelatedto
allregisteredUEs,liketheUEidentifier,CQIfeedbacks,
uplinkchannelquality,anduplinkschedulingrequest

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

••Page5of13

(i.e.,theamountofdatathatauserneedstotransmit,mode,havetheconsidereduserinthesubscribergroup
reportedtothetargetbasestationthroughanuplinklist.
controlmessage).
AHeNBcontainsalltheentitiesoftheE-UTRANpro-
3.4Newtopologyobjects
tocolstack(i.e.,RRC,MAC,andPHY)whichimplemen-Threenewnetworktopologyobjectshavebeenintro-
tationsdifferwithrespecttotheeNB.Inparticular,atRRCduced:
andMAClayers,enhancedfeatureshavebeenadded,e.g.,
anewhandoverproceduremoresuitableforheteroge-
Femtocell:ithasonlygeometricpropertiesanditis
neousenvironmentsandthepossibilitytoconfigurethe
usedfordetailingthepositionandtheIDofa
HeNBinbothclosedandrestrictedaccessmodes(see
femtocell.
Section3.3fordetails).
Building:itiscomposedbyanumberofapartments,
3GPPspecificationsdonotexplicitlydefinethephysi-
eachonedelimitingtheareaofagivenfemtocell.As
calpowertransmissionofaHeNB.However,wesetthe
definedin[34],twodifferenttypesofbuildinghave
defaultvalueofthemaximumtransmissionpowerequal
beendeveloped:(a)DualStripeblocksand(b)
5
×
5
to20dBm,accordingtosuggestionsreportedin[32].
apartmentgrid(seeFigure2).Theformerconsistsin
Obviously,thisisnotastandardizedconfiguration:the
twobuildingscomposedoftworowsof10
powertransmissionofthebasestationcouldbesettaking
apartmentseach.Thelatter,instead,isabuilding
intoaccountconditionsofbothsurroundingenvironment
composedof25apartmentslocatedovera
5
×
5
grid.
andnetwork.Tocopewiththisrequirement,LTE-sim
EachbuildingisidentifiedbyanuniqueIDandits
allowstheusertofreelychangeinastaticordynamical
positionisdefinedintoaCartesiansystem.Forboth
fashionthepowertransmissionoftheHeNB,accordingto
itsneeds.
Aslightenhancementhasbeenappliedalsototheclass
modelingtheUE.WithrespecttotheoldLTE-simimple-
mentation,weintroducedabooleanflag,updatedevery
timetheuserpositionchanges.Suchaflagisusedforrec-
ognizingwhetheraUEisinsideoroutsideabuildingata
certaintimeinstant;notethatthisinformationisfunda-
mentaltocatchpropagationlossesduetowallsbetween
thetransmitterandthereceiver.
3.3Newhandovermanagement
Whileausermovesalongacertainpath,the
handover
manager
hasthedutyofperformingcellre-selectionand
hardhandoverprocedures.Itisimplementedintothe
RRCentityasproposedin[33].
Initspreliminarilyversion,LTE-simsupportedonly
a
position-based
handoverprocedure,whichwasused
topickasservingnodetheclosesteNB.Inhomoge-
neousscenarios,wherealltheeNBstransmitatthe
samepowerandaspacefreepathlossmodelcanbe
considered,thiskindoftechniquewellapproximatesa
genericpolicybasedonthereceivedpower.Obviously,
thisbecomesunfeasibleinheterogeneousscenarioslike
theonesincludingbuildingsandfemtocells.
Forthisreason,weconceivedanewstrategythatfor
eachnodeselects,astargetbasestation,theoneproviding
thestrongestreceivedsignal.
Furthermore,inordertocopewithseveralaccesspoli-
cies,aspecialstructureiscreatedfortheMAClayerofthe
HeNB.Itcontainstheidentifiersoftheauthorizedusers
(i.e.,thesocalledsubscribergroup[20]).Accordingly,the
handoveralgorithmallowsagivenusertojoinonlythose
cellsthateitherworkinopenmodeor,beinginrestricted

(a)

(b)Figure2
Availablebuildingconfigurations:
(a)
DualStripe
blocksand
(b)
5
×
5
apartmentgrid.

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

•theaforementionedbuildingtypes,itispossibleto
definethenumberoffloors.Eachapartmenthasa
squaredformandanareaof100m
2
.
Ingeneral,eachapartmentcontainsuptooneactive
femtocell,i.e.,anactiveHeNBisworkinginthe
femtocell.Thismeansthat,forinstance,a
5
×
5
grid
buildingcancontainupto25femtocells.The
presenceofanactivefemtocellinasingleapartment
canberandomlydecidedthroughthedefinitionofan
activityratio[34],thatistheprobabilitythatanactive
homebasestationispresentinanapartment.
Street:itmodelstworowsofbuildingslocatedalong
awideroad.Itsgeometricparameters(e.g.,thestreet
width,thedistancebetweentwoadjacentbuildings,
andthenumberofbuildingalongthestreet)canbe
easilycustomizedbasedonthesimulationneeds.
Asanexample,Figure3showsatypicalurbancross
composedoffourdifferentstreetsand5
×
5gridbuildings
objects.
Inaddition,wewouldliketoremarkthatthedevel-
opedsimulatorisflexibleand,asaconsequence,itcanbe
extendedforintegratingrealstreetmaps.Thiswouldbe
veryusefulforinvestigatingmorerealisticenvironments.

Figure3
Exampleofacrosscomposedoffourdifferentstreets
and
5
×
5
gridbuildings.

Page6of13

3.5Channelmodels
The
Channel
moduleofthesimulatorhandlespacket
transmissionsandmodelsthepropagationlossbymeans
fourdifferentphenomenaassuggestedin[35]:(i)thepath
loss,(ii)thepenetrationloss,(iii)theshadowing,and(iv)
thefastfadingduetothesignalmultipath.
Forauserlocatedinabuildingandservedbythebase
stationofamacrocell,thepathlosscalculationtakesin
considerationalsoanadditionalattenuationfactordue
tothepresenceofanexternalwall(defaultvalueofthe
externalwallattenuationis20dB[34]).
Moreover,inordertocopewiththepeculiarfeatures
offemtocells,twonewpathlossmodelshavebeenintro-
duced.Thefirstoneistheindoorpropagationmodel
definedin[34]foradenseurbandeploymentoffem-
tocells.Itevaluatesthepathloss,
P
L
,consideringonly
thedistance,
R
,betweenthetransmitterandthereceiver
expressedinmeters:
P
L
[dB]
=
127
+
30
·
log
10
(
R
/
1000
)
.(1)
Thesecondonehasbeendevelopedwithinthe
WinnerII
projectforindoorresidentialenvironments[36].Itoffers
ahighaccuracyatthecostofanincreasedcomputational
complexity:
fcP
L
[dB]
=
A
·
log
10
(
R
)
+
B
+
C
·
log
10
+
X
;(2)
5where
R
isexpressedinmeters;thecentralfrequency
f
c
isexpressedingigahertz;thevaluesofotherparameters
A
,
B
,and
C
dependonthenumberofwallsandfloors
betweenthetransmitterandthereceiver.
Itisimportanttoremarkthatthequalityofthesig-
nalreceivedbyusersinsideabuildingdependsonhow
thebandwidthissharedamongfemtocells.Toinvestigate
thisconcept,wecomparedthedistributionofsignal-
to-interferenceplusnoiseratio(SINR)providedbythe
WinnerII
channelmodelinsidea5
×
5apartmentgrid,
supposingtoassignthespectrumbymeansofthereuse-1
(i.e.,allfemtocellssharethesamebandwidthof20MHz)
andthereuse-1/2scheme(i.e.,twoportionsofthespec-
trum,of10MHzeach,aredistributedamongHeNBs
imposingthattwoadjacentfemtocellscannotworkonthe
sameoperativebandwidth).FromFigure4,itisevident
thattheadoptionofafrequencyreuseschemeimproves
thechannelqualityperceivedbyusers.Infact,bydis-
tributingthewholespectrumamongHeNBs,itispossible
toreducetheinterferencelevel.However,thisadvantage
couldbereachedatthecostofthenetworkthroughput
(e.g.,thehigherthenumberofportionsofthespectrum
consideredduringthefrequencyplanningstrategy,the
loweristhebandwidthavailableforeachfemtocell).Such
aspectswillbebetterdiscussedinSection4.1.

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

0402020

0-20y

4035035202510120100-20-10
x)(a403836343230282624220-1020
20
10-20-100
18
y-20x(b)Figure4
SINRmeasuredinsideabuildingwith:(a)nofrequencyreuseand(b)
reuse-1/2
scheme.

4002002

10

Page7of13

4Workingwiththeproposedmodule
downlinktrafficmixcomposedbyonevideoflow,one
Inthissection,weprovideanoverviewaboutsomeoftheVoIPflow,andonebesteffortflow.
possiblestudiesthatcanbecarriedoutusingoursimula-Forthevideoapplication,weusedatraffictrace
tionframework.Tothisend,theimpactofco-layerinter-obtainedfromatestsequence(i.e.,“foreman.yuv”)avail-
ferenceinindoorenvironmentsgeneratedamongHeNBableat[37].Theoriginalsequence(at25frame/s,
aswellastheperformancegainscomingfromfemtocellCIFresolution352
×
288,andYUVformat)hasbeen
deploymentinanurbanscenariohavebeenanalyzed.Atfirstlyrepeatedforthewholesimulationtime.Then,the
last,ascalabilitystudyaboutcomputationalrequirementsobtainedvideohasbeencompressedusingH.264stan-
(i.e.,simulationtimeandmemoryusage)ofthesimulatordardcompressionattheaveragecodingrateof440kbps.
isalsoprovided.SimulationparametersaresummarizedInstead,forG.729voiceflows,weadoptedanON/OFF
inTable1.Markovmodel,wheretheONperiodisexponentiallydis-
tributedwithmeanvalue3s,andtheOFFperiodhasa
4.1Studyingtheimpactofco-layerinterferenceinan
truncatedexponentialpdfwithanupperlimitof6.9sand
indoorenvironment
anaveragevalueof3s[38].DuringtheONperiod,the
Theimpactthattheco-layerinterferenceduetothecom-sourcesends20byteslongpacketsevery20ms(i.e.,the
municationbetweenhomebasestationanduserhasbeensourcedatarateis8kbps),whileduringtheOFFperiod
evaluatedinanindoorscenariocomposedbyasingle5
×
5therateiszerobecauseweassumethepresenceofavoice
apartmentgrid.activitydetector.Finally,forthebesteffortflowswehave
Theanalysishasbeencarriedoutbyconsideringthreeconsideredinfinitebuffersources.
differentfrequencyreuseschemes,e.g.,thereuse-1,theForalltheaforementionedspectrumconfigurations,we
reuse-1/2,andthereuse-1/4
b
,andbyimposingtohavecomparedthebehaviorofwell-knownpacketschedulers:
ineachapartment8UEsreceivingatthesametimeatheproportionalfair(PF)[3],thelogarithmic(LOG)rule

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

Table1Simulationparameters
Parameter
Totalbandwidth
eNBpowertransmission
HeNBpowertransmission
CQIApartmentsize
#apartmentsinabuilding
Radiusofthemacrocell
#buildings
FrequencyReuseScheme
#users
Scheduler
fficraT

Page8of13

IndoorenvironmentUrbanscenarioScalabilitytest
Hz0M243dBm,equallydistributedamongsub-channels
20dBm,equallydistributedamongsub-channels
Fullbandwidthandperiodicreportingscheme.Measuredperiod:2ms.
100m
2
100m
2
100m
2
252525
N.A.500m500m
156from1to6
reuse-1,reuse-1/2,andreuse-1/4reuse-1reuse-1
8perHeNBfrom40to120from0to300
PF,LOGrule,andFLSPFPF
Video,VoIP,andbest-effortbest-effortbest-effort

[39],andtheframelevelscheduler(FLS)[40,41]).Inpar-Togainafurtherinsight,weevaluatedtheJainfairness
ticular,networkperformancehasbeenevaluatedwithindex[42],findingthatitishigherthan0.8inallconsid-
referencetothethroughputachievedbybesteffortflowseredconditions.Thismeansthatneitherthefrequency
andthepacketlossrate(PLR)ofmultimediaones(i.e.,reusestrategynortheschedulingalgorithmaffectsthe
videoandVoIP).fairnessinbandwidthsharingofbesteffortflows.
FromFigure5itispossibletonotethatthethroughputFigure6reportsthePLRofvideoflows,showingthat,
ofbesteffortflowscanbemaximizedusingthereuse-thereuse-1/2isabletoprovidethebestQoS.
1/2scheme,whichisabletoprovidethebesttrade-offMoreover,inlinewithresultspresentedin[40],the
betweenthebandwidthavailableineachfemtocellandtheFLSapproach,whichhasbeendesignedforguaranteeing
interferencelevelproducedbyHeNBsinsidethebuilding.boundeddelaystomultimediaflows,providesthelowest
Theadoptionofthereuse-1/4schemeproducesalwaysPLR.Thus,itensuresthehighestvideoqualitytomobile
theworstnetworkperformance.Infact,despiteenor-users.Thisperformancegaincomesattheexpenseofper-
mousbenefitsthatthereuse-1/4schemeoffersintermsformanceexperiencedbybesteffortflows,whichasmaller
ofSINR(duetotheverylimitedleveloftheco-layeramountofbandwidthisassignedtowhenFLSisuse(see
interference),itimposesthateachHeNBusesasmallFigure5).
bandwidth,thusdecreasingtheoverallachievablesystemWithreferencetoVoIPflows,weobservedthatthey
throughput.achievedsmallerPLRvaluesthantheonesrelatedto

50PF45LOG RULE
FLS4035305202510150

1/11/21/4
Frequency Reuse Scheme
Figure5
Aggregatecellthroughputforbesteffortflowswithdifferentfrequencyreuseschemes.

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

210

101

010

-1101/11/2
Frequency Reuse Scheme
Figure6
PLRofvideoflowswithdifferentfrequencyreuseschemes.

14/

PFLOG RULE
SLF

Page9of13

videoswithoutanynoticeabledifferencesamongallcon-Simulationswerecarriedoutvaryingthenumberofusers,
sideredschedulingapproaches(thePLRislessthat5%eachonereceivingadownlinkflowmodeledwithan
whenthereuse-1schemeisused;0%otherwise).Therea-infinitebuffersource.
sonisthatsuchapplicationsgetahigherpriorityfromtheFigure8showstheaverageuserthroughputofoutdoor
schedulerbecausetheyhavealowersourcebitrate.andindoorusers.
WecanconcludethatthetypicalproblemconnectedInthetypicalurbanscenariowithoutfemtocells,indoor
tointerferencemanagement,asalreadyexplainedinuserswillbeanywayservedbythemacrocellsufferingfor
Section2,liesintheunpredictabilityofthedeployedsce-lowchannelqualityduetothestronginfluenceofwall
nario.Infact,inadenseurbanenvironmentthenetworkattenuation.Introducingfemtocellsinsidethebuildings,
operatorcouldnothaveenoughinformationabouttheindooruserswillexperiencesignificantincreaseinterms
femtocelldensityandtheHeNBspositionsineachbuild-ofthroughput.Performanceofoutdoorusers,onthe
ing.Thus,thepossibilitytoapplyanykindoffrequencyotherhand,shouldnottakeanyadvantageintermsof
planningstrategyvanishes.Therefore,theproposedsce-experiencedSINR.Infact,whenhomebasestationsand
narioisusuallyconsideredforinvestigatingtheimpactofeNBssharethesamebandwidth,outdooruserslocated
differentdynamicspectrumallocationpoliciesforinter-nearbythebuildingwillsensehighlevelofinterfer-
ferenceavoidanceinfemtocellenvironments,andthepro-encewithconsequentperformancelosses.Nevertheless,
posedanalysisrepresentsavalidstartingpointforstudies
inthisdirection.
4.2Capacityenhancementinurbanenvironment
Now,weinvestigatetheimpactofthefemtocelldeploy-
mentinurbanenvironments.Tothisend,wedesigned
ascenarioconsistingofonemacrocelland56buildings
locatedasinatypicalurbancross(seeFigure7).
TheeNBislocatedatthecenterofthemacrocellandit
transmitsusinganomni-directionalantennaina20MHz
bandwidth.Twodifferentcasesareobjectofthestudy:
•
traditionalurbanenvironmentwithoutfemtocells
whereonlyonemacrocellandbuildingsareusedas
referencecase;
•
urbanenvironmentwithfemtocellswhereone
macro cell
femtocellperapartmentisassumedtobeactiveand
Figure7
Simulatedscenariowithacrossdesignedwith4
workingonthesameoperativebandwidthof
differentstreetsand
5
×
5
gridbuildings.
themacrocell.

street1
eBN

building,
5x5 apartmentgrid
street3

Capozzi
etal.EURASIPJournalonWirelessCommunicationsandNetworking
2012,
2012
:328
http://jwcn.eurasipjournals.com/content/2012/1/328

210110010

-110

-210

201101010

1-10

40 UEs
80 UEs
120 UEs

no femtocells

40 UEs
80 UEs
120 UEs

Scenario
(a)

-210no femtocells
Scenario
)b(Figure8
Averagethroughputfor:(a)outdoorand(b)indoorusers.

with femtocells

with femtocells

Page10of13

Figure8demonstratesthatperformanceofoutdoorusersTheinvestigatedLTEnetworkiscomposedofone
improveaswell.Thisapparentlycounterintuitiveresultismacrocellandanumber,
N
,ofbuildings,with
N
rang-
mainlyduetothemacrocellunloadingeffectcontributedingfrom1to6.Buildingsareuniformlydistributedinside
bytheintroductionoffemtocells(seeSection2).Inotherthemacrocelland,ineachofthem,25femtocellsare
words,acertainnumberofusersisservedbyHeNBsandenabled.Simulationshavebeencarriedoutconsideringa
theamountofresourcesavailableforremainingoutdoorfixednumberofUEsequalto30thatbelongtothemacro-
usersconsequentlyincreases.cellandvaryingthenumberofUEsineachbuildingin
Finally,Figure9showsthattheoverallsystemcapac-therange[10–50].Oneactivedownlinkflowwithinfinite
itycanbegreatlyimprovedthankstotheadoptionofbuffercharacteristicsisconsideredpereachUE.More-
femtocells.over,asimulationtimeof30sisconsidered.Withthese
Wehighlight,onceagain,thattheproposedstudiesaimconditions,the“lightest”scenariowillbecomposedof26
atshowinghowourmodulecanbeeasilyusedtoinves-cellsand40users,whereasthe“heaviest”onewillembrace
tigatethemainissuesrelatedtothedeploymentoffem-151cellsand330users.
tocells.Wethinkthat,startingfromthisbasicscenario,Thechoicetouseinfinitebufferapplications(i.e.,there
severalnoveltechniquescanbetestedandcompared.isalwaysapacketstotransmit)isusefultoevaluatethe
simulatorbehaviorunderstrenuousconditions.
4.3Scalabilitytest
Figure10showstimeandmemoryrequiredintheafore-
Asfinalaspect,weproposeascalabilityanalysisofthementionedscenarios.Programexecutiontimelinearly
proposedmoduleintermsofbothsimulationtimeandincreaseswithboththenumberofbuildingsandthe
memoryusageonaLinuxmachinewitha2.6GHzCPUnumberofusers.Nevertheless,itremainslimitedalsoin
and4GBytesofRAM.scenarioswiththehighestcomputationalload.