location aware son whitepaper tl

13
The Critical Importance of Subscriber-centric Location Data for SON Use Cases

Upload: sulissetiawati

Post on 17-Dec-2015

17 views

Category:

Documents


0 download

DESCRIPTION

Location Aware Son Whitepaper Tl

TRANSCRIPT

  • The Critical Importance of Subscriber-centric Location Data for SON Use Cases

  • Page2

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    Version 1.0Issued 27December2012Theinformationcontainedinthisdocumentandanydocumentationreferredtohereinorattachedhereto,isofaconfidentialnatureandissuppliedforthepurposeofdiscussiononlyandfornootherpurpose.Thisinformationshouldonlybedisclosedtothoseindividualsdirectlyinvolvedwithconsiderationandevaluationofanyproposals,allofwhoshallbemadeawareofthisrequirementforconfidentiality.Alltrademarksareherebyacknowledged.

  • Page3

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    AriesoSolutionsJDSUacquiredAriesoinMarch2013,addingtheworld'sleadingintelligent,locationawaresolutionsformobilenetworkoperatorstoitsCommunicationsTestportfolio.Ariesosolutionslocate,storeandanalysedatafrombillionsofmobileconnectionevents,givingoperatorsarichsourceofintelligencetohelpboostnetworkperformanceandenrichuserexperience.Thisintelligencetransformstheeffectivenessofnetworkperformanceengineering;enablescustomercentricselfoptimisingnetworks;createstrueunderstandingofcustomerexperienceandenablesmonetizationofuniqueinsights.TheprovenAriesocarriergradesolutionsareresilientandhighlyscalable.Operatingonfivecontinents,clientsincludemobileoperatorgroupssuchasAmricaMvil,AT&T,MTN,TelefnicaandVodafone,andleadingequipmentvendorsincludingAlcatelLucentandNSN.JDSU(NASDAQ:JDSU;andTSX:JDU)innovatesandcollaborateswithcustomerstobuildandoperatethehighestperformingandhighestvaluenetworksintheworld.Ourdiversetechnologyportfolioalsofightscounterfeitingandenableshighpoweredcommerciallasersforarangeofapplications.LearnmoreaboutJDSUatwww.jdsu.comandfollowusonJDSUPerspectives,Twitter,FacebookandYouTube.MoreinformationonAriesocanbefoundatwww.arieso.com.

  • Page4

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    ContentsARIESOSOLUTIONS.........................................................................................................................3EXECUTIVESUMMARY.....................................................................................................................5AUTOMATICNEIGHBOURRELATIONS..............................................................................................6COVERAGEANDCAPACITYOPTIMISATION......................................................................................8ENERGYSAVINGS............................................................................................................................9MOBILITYLOADBALANCING.........................................................................................................10MOBILITYROBUSTNESSOPTIMIZATION.........................................................................................11SUMMARY....................................................................................................................................12

  • Page5

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    ExecutiveSummarySelfOptimizingNetworksofferconsiderablegainsinoperationalexpenditureefficienciesasmanynetworkimprovementtasksformerlydonemanuallybyengineerscannowbecarriedoutbyautomatedmechanisms.WhilethisisaprimarygoalofSON,therearealsoexpectationsthatSONwilloffergainsinotherareas,includingcapitalexpenditureandperformance.Inmanyregards,SONisbestconsideredastheautomationofnetworkimprovementactivitieswhichwereformerlydoneinamanualfashion.Assuch,considerationofpreviouslymanualtasksprovidesvaluableinsightintotheimportantingredientsforsuccessfulSONsolutions.Asaresultoftheintrinsicallyspatialnatureofwirelessengineering,aconsiderationoflocationinformationplaysakeyroleinnearlyallmanualoptimizationtasks.Classicexamplesincludecoverageplotsinradioplanningtools,spiderplotsinneighbourlistanalysis,andsignalstrengthplotsindrivetestpostprocessingsoftware.Therearealsomajortrendsinthewirelessindustryregardingtheuseofsubscribercentriclocationdata.ThecriticaldependencyonsubscribercentriclocationinformationcontinuesinSON,especiallyinthefollowingpopularSONusecases: AutomaticNeighbourRelations CoverageandCapacityOptimization EnergySavings MobilityLoadBalancing MobilityRobustnessOptimization

  • Page6

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    AutomaticNeighbourRelationsAutomaticNeighbourRelations(ANR)istheearliestandmostextensivelydeployedSONusecasetodatearoundtheworld.ANRreferstotheabilitytogenerateusefulneighbourlistsforeachsectorinthenetwork(anexampleforonesectorisshownbelowinFigure1).Subscribercentricdataiscriticallyimportantforthegenerationofneighbourlistdetailsforinterfrequency,intrafrequencyandinterradioaccesstechnology(IRAT)applications.Thishelpstoensurethatthosesectorsthataremostoftenreportedbymobileswillbehighlyprioritizedintheensuingneighbourlistfortheservingsector.

    Figure1:AutomaticNeighbourRelationsexample

    Interfrequencyscenariosplaceanemphasisontheneedforlocationdataduetoexpectedvariationsinpropagationdistancesasafunctionofcarrierfrequency.ItcanalsoaccountforpowervariationsduetotheuseofdifferentRFmodulesandcabling.Subscribercentriclocationdataexplicitlyanswersthequestionofhowfarindividualcarriersextendandwhatoverlaycarriersareappropriatehandovertargetsindifferentportionsofthenetworkunderstudy.Interradioaccesstechnologyscenariosplaceanadditionalemphasisontheneedforlocationdata.Thefootprintsofeachoftheradioaccesstechnologiesunderstudymustbecarefullytakenintoconsideration(inatrafficweightedmanner)inordertoensurethatthebestIRATneighboursareidentifiedandemployed.ThistrafficweightingiskeysinceiteffectivelyallowsforapopularvotebytheactualsubscribersastowhicharethemostappropriateIRATneighbours.

  • Page7

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    ExecutionofANRfunctionalityintheabsenceofsubscribercentriclocationdataposesseveralrisks.Theensuingtrial&errorsearchinawideparametricspace(whichmoreoftenthannotresemblesarandomwalk)resultsinsuboptimalnetworkoperation.Whileitis,intheory,possibletofindanoptimalcollectionofparameters,itismoreoftenthecasethatlocalmaxima(whicharesuboptimalbydefinition)willprohibitprogressalongatrajectorythatleadstotheglobalmaximum.Thisisarecurringthemeintheuseofgeographicallyblindoptimizationstrategies,aswillbenotedintheremainderofthispaper.AnotherriskassociatedwithnotusingsubscribercentriclocationdatainvolvestheimpairedabilitytoassessanddebugthesolutionsfoundbytheANRprocess.WhileengineeringinvolvementisnotexplicitlyrequiredintheheartofclosedloopSONoperations,itisstillthecasethatanySONsolutionmustbesubjecttoscrutinyanddiagnosticevaluation.Theabsenceoflocationdataprohibitstheengineerfromassessingthesituationsofsubscriberswhoarerecommendingtheaddition,deletionormaintenanceofparticularneighbours(thusprovidingacustomercentricviewofthenetwork).Forexample,locationinformationisimportantforidentificationofovershootingneighbours;itisoftenbettertoincreasetiltand/orreducetransmitpowertoeliminatetheexcessiveinterferenceoftheovershooterthantokeeptheovershootingneighbouronaneighbourlist.

  • Page8

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    CoverageandCapacityOptimisationTheCoverageandCapacityOptimization(CCO)usecaseadjustsnetworkparameters(etilt,powerlevels,etc.)inordertomaximallysatisfycoverageandcapacityobjectives.InallpracticalCCOscenarios,oneofthekeyinputconstraintsisthelocationofsubscribertraffic.Thistrafficconstraintiseitherexplicitlyknowntothealgorithmcomputingtherequiredchange,orimplicit,inthatwhilstnotknowninadvance,itwillinfluencetheimpactofthechangesoncemade.UseofcustomercentriclocationdataasanexplicitinputtoCCOiscriticallyimportantbecauseitallowsfordirectconsiderationoftheconsequencesofeveryplannednetworkchangebeforeitismadeinthelivenetwork.ThisexplicitconsiderationallowstheCCOprocesstomoreeasilyarriveatoptimalnetworksolutionssuchasthoseshowninFigure2below.TheseCCOtrialresultsshowthatinthecoreofthenetwork,areaswithRSCPlevelsbelow95dBm(shownasred)havealmostdisappeared.

    Figure2:CoverageandCapacityOptimizationtrialresult

    AsnotedintheANRusecase,theavailabilityofsubscribercentriclocationdataallowsformoreeffectivediagnosticanalysisofCCOsolutions.Theabsenceofthisinformationmakesitnecessarytoappealtootherinferiordatasourcesinordertointerprettheconditionofthenetwork.Theseinferiordatasourcesincludedrivetestdata(whichonlyaddressroadlevelconditions)andswitchstatistics(whichonlyprovidecoarse,sectorlevelspatialresolution).

  • Page9

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    EnergySavingsTheexplosionindatademandoverthepastseveralyearshasresultedintheadditionofmanycellsitesinordertosatisfyincreasingcapacityobjectivesduringhoursofpeakdemand.Thisisespeciallytrueinsituationswheretheamountoflocallyavailablespectrumisparticularlylimited.Itshouldbenotedthatthelocationswherethecapacityrelatedcellswereaddedalreadyhadadequatecoverage(asservedbythepreexistingnetworkinfrastructure).Assuch,thesenewcapacityobjectivesstandinstarkcontrasttotheoldercoverageobjectivesthatdominatedtheearlierdecadesofwirelessnetworkbuildouts.However,manyoftheserecentlyaddedsitesdonotsatisfymissioncriticalcapacityobjectivesduringoffpeakhours.Giventhattheadditionofthesenewsitesdidnotservetoimprovethelocalcoverage,itisreasonabletoexpectthatthesesites(orothermacrositesnearby)canbepowereddownduringoffpeakhoursinordertoaccomplishpowersavings.Underthesecircumstances,itisnecessarytoensurethattheoverallcoverageobjectivesarestillsatisfiedandthattheresultingnetworkstillhastherequisitecapacityasrequiredbytheactualdemandsofnearbysubscribers.Subscribercentriclocationdataiscriticallyimportanttothisusecaseinordertoensurethattheoptimalselectionofsitestobepoweredoffcanbedetermined.Suchdataalsoallowspredictionsoftheconsequencesthatwillbeseenoncethechangesarecutin.TheresultsofanEnergySavingsanalysisareshownbelowinFigure3.

    Figure3:EnergySavingsanalysis(before&after)

    Thisparticularexampleshowsthatthereareconsiderableopportunitiestopowerdownpartsofthenetworkduringoffpeakhourswhileensuringthatcustomersenjoythesameorbettercoveragethattheyexperienceduringpeaktrafficconditions.BasedonaTier1marketanalysis,ithasbeenfoundthatthesavingsopportunitiesvaryoveranumberofoperationalscenariosfromaconservative28%toanaggressive77%.Forasinglenetworkoperatorwith10,000sites,ithasbeenconservativelyestimatedtheproposedpowerdownstrategieswouldresultinanannualsavingsof$4.3million.Equivalentanalysesperformedwithouttheuseofcustomercentricdatainvariablyresultinaprohibitivelycomplextrial&errorwalkthroughanexponentiallylargespace.Forasfewasthirtycellsites,thereareoverabillionpossibleon/offcombinations.AssessmentoftheconsequencesofanyparticularcombinationissimilarlyconstrainedintheabsenceofcustomercentriclocationdataasnotedinearlierSONusecases.

  • Page10

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    MobilityLoadBalancingTheMobilityLoadBalancing(MLB)usecaseinvolvessettingbothidleandactiveparametersinordertoensurethattrafficissuitablyspreadacrossmultipleradioaccesstechnologies.AnexampleofMLBisshowninFigure4belowwheretheredarrowsdenoteIRAThandoverbetweenLTE/UMTS,LTE/GSMandUMTS/GSMlayers.Subscribercentriclocationdataprovideskeyinputintothesettingoftheseparametersinamannerthatoptimizesthespreadoftrafficacrossthelocallyrelevantlayers,subjecttothespatialvariationsinbothsubscriberdemandaswellassubscriberdevicetype(includingwhetherdevicescanaccommodatedifferentairinterfacetechnologies).OptimalMLBstrategieswillalsotakeintoaccountthelocalspatialsupportoftheairinterfacetechnologies(similartotheANRusecasediscussionnotedearlier).

    Figure4:MobilityLoadBalancingexample

    LoadBalancingexercisesperformedwithouttheuseofsubscribercentriclocationdatacanoftenresultinsuboptimal,poorlydifferentiated(nearlyonesizefitsall)parametersettingsintheoutputsofMLBprocesses.ThisiscloselyrelatedtothemannerinwhichtrafficloadbalancingisaccomplishedinnonSONsystems.Atbest,thediscoveryofatrulyoptimalsolutionisgreatlydelayedbytrial&errorsearchesthroughacomplexparametricspace.DiagnosticassessmentofanyMLBoutcome(includingthedefault,onesizefitsallsetting)willalsosuffersinceother,suboptimaldatasourceswillneedtobeconsidered(drivetest,switchstatistics,etc.)Diagnosticassessmentsincludetheincreasinglyimportantchallengeofunderstandingwhycertain4Gand3Gdevicesarestrandedonlowerperformanceairinterfaces.Thesestrandedscenariosareparticularlyimportantduetothestrainthatisplacedonthecustomerexperienceandthegreatlyincreasedprobabilityofchurn.

  • Page11

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    MobilityRobustnessOptimizationTheMobilityRobustnessOptimization(MRO)usecaseinvolvesoptimizationofhandoverexperiencesthroughchangesofavarietyofnetworkparameters.Suboptimalhandoverconditionsinvolveoneormoreofthefollowingsymptoms:

    Tooearlyhandover Toolatehandover Unnecessaryhandover Pingponghandover Handovertothewrongcell

    TheexamplebelowinFigure5showsanMROsituationwherehandoversintheredovalareoccurringinanunnecessarymanner(i.e.,wherehandoversfromthefirstcellarefollowedbyaverybriefconnectiononthesecondcell,followedbyhandoverseitherbacktothefirstcellortosomeotherthirdcell).ChangestoMROnetworkparametersresultedintheeliminationofmanyofthesehandovers(asseenintheafterpictureontheright)withoutnegativelyimpactingotheraspectsofnetworkoperation.Velocitydata(whichcanliterallybederivedfromlocationdata)canalsobeofusetooptimallydeterminetimeconstantsassociatedwithMRO.Thisisofparticularinterestinthisexamplegiventhemajoreastwestroadwayrunningthroughthemiddleoftheredoval.

    Figure5:MobilityRobustnessOptimizationexample(before&after)

  • Page12

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    AsnotedinearlierSONusecases,theomissionofsubscribercentriclocationdataresultsinprocessesthatrequirelengthysearchesthroughcomplexparametricspacesand/orsuboptimal,onesizefitsallparametersettings.InterpretationofanyMROsetting(includingthedefault,onesizefitsallsetting)willalsosuffersinceother,suboptimaldatasourceswillneedtobeconsidered(drivetest,switchstatistics,etc.)

    SummarySelfOptimizingNetworksofferconsiderablegainsinoperationalexpenditureefficiencies,capitalexpenditureefficienciesandnetworkperformance.Theintrinsicallyspatialnatureofwirelessengineeringhasinthepastmadeconsiderationoflocationinformationthekeytosuccessinmanualoptimizationactivities.AsSONbecomesincreasinglyanprominentpartofnetworkfunction,sotheuseoflocationinformationbecomesevermorefundamental.Useofsubscribercentricsourcestoobtainlocationinformationprovidesadoublebenefit:1)Itprovidesready/relevant/inexpensiveaccesstokeydata(incontrasttodrivetesting&switchstatistics)and2)Itensuresthatthefocusremainsonthecustomer.Indeed,theneedforthislocationinformationtobederivedfromtheexperiencesofactualsubscribersisfoundtobeanaturalextensionofthesubscribercentrictrendsbeingembracedthroughoutthewirelessindustry.

  • Page13

    AriesoCommercialinConfidence Copyright2013AriesoLtd

    AriesoLtdAstorHouseNewburyBusinessParkLondonRoadNewburyBerkshireRG142PZUnitedKingdomTel: +44(0)1635232470Fax: +44(0)1635232471

    AriesoInc3495PiedmontRdBldg11Suite550AtlantaGA30305USATel: +16789042424Fax: +16789042429

    Email: [email protected]: www.arieso.com