POBox680,HoodRiver,OR97031 http://www.rysavy.

com,15413867475

October13,2011

Commentson CitiWirelessSupplyandDemand
1 Overview
On September 22, Citi Investment Research & Analysis released a report titled Wireless Supply and Demand. This report calculates the capacity of wireless networks based on certain values of spectral efficiencyofcertaintechnologies.Relativetomyownanalysis,thespectralefficiencyvaluesusedinthe Citi report are incorrect to a significant degree, and hence much of the technical analysis and related conclusionsarealsoincorrect.Inparticular,thebottomlineconclusionthatCitiarrivesatWedonot believe the US faces a spectrum shortage is invalid because it is predicated on the incorrect analysis that4Gtechnologieshavesixtimesmorecapacitythantheyreallydo.Thisincorrectcapacityanalysis means Citis estimations of the amount of spectrum needed to support deployment of 4G LTE technologies are also incorrect. Citis analysis of spectrum demand for widespread 4G deployment wouldonlyanticipateonesixththeamountofspectrumthanisactuallyneededforsuchdeployment.

2 Basis of Rysavy Research Analysis

Since 2002, Rysavy Research has published twelve publicallyavailable reports that include spectral efficiency analysis. The most recent report, published by 4G Americas in September, 2011, is titled Mobile Broadband Explosion 3GPP Broadband Evolution to IMTAdvanced.1 Spectral efficiency is coveredinpages53through61.Thespectralefficiencyvaluespresentedinthispaperaretheoutcome of a detailed analysis done jointly and on a consensus basis by Rysavy Research working with representatives of 4G Americas boardofgovernor companies2 who are 3GPP technical experts. The representatives were from both leading wireless operators and infrastructure vendors. Due to the experienceandexpertiseofthisworkgroup,thespectralefficiencyvaluespresentedarehighlycredible. Thesevaluesarealsoconsistentwithotherindustrygroupsasdetailedfurtherbelow.

Availableathttp://www.rysavy.com/Articles/2011_09_08_Mobile_Broadband_Explosion.pdf.

4GAmericasboardofgovernorcompaniesarelistedat http://www.4gamericas.org/index.cfm?fuseaction=page&sectionid=160.

3 Analysis
SpectralefficiencyisthevaluebywhichamountsofspectrummeasuredinHertz(Hz)canbeequatedto bits per second (bps) of data throughput. Given exploding mobilebroadband data growth, understanding the real capacity of cellular networks is of huge importance. In 2010, Rysavy Research published a report on this topic titled, Mobile Broadband Capacity Constraints and the Need for Optimization.3Itisimportanttorealizethatthereisadifferencebetweenaveragespectralefficiency and peak spectral efficiency. Under ideal conditions of low network loading and good signal quality, todays adaptive wireless networks can achieve extremely high throughput rates, and so the peak spectral efficiency value can be very high. For networkcapacity analysis, however, one must use the average spectral efficiency value, a value that is much lower than the peak value, since this is what allowsforthecalculationoftheaggregatecapacityofacellsector. Onpage7, theCitireportpresentsHSPA+downlinkspectralefficiencyas2.1bps/Hzand LTEspectral efficiency as 8.6 bps/Hz. In contrast, my analysis, as depicted in Figure 22 of Mobile Broadband Explosion and shown below in Figure 1 below presents HSPA+ spectral efficiency for current deployment configurations as 1.05 bps/Hz and LTE as 1.4 bps/Hz. These are huge and dramatic differences.Relativetomyanalysis,theCitireportoverstatesHSPA+spectralefficiencyby100%andLTE spectralefficiencyby614%.ItisunusualthattheCitireportdoesnotprovideanyexternalreferencesfor itsspectralefficiencyvalues.

Availableathttp://www.rysavy.com/Articles/2010_02_Rysavy_Mobile_Broadband_Capacity_Constraints.pdf.


Figure1:RysavyResearchDownlinkSpectralEfficiencyComparison

MuchoftheCitireportdependsonthesevalues.Giventheseerrors,muchofthedataandconclusions intheCitireportsectionsaremisleadingorincorrect. Page 1: Bottom Lines We do not believe the US faces a spectrum shortage. This conclusion is invalid because it is predicated on the incorrect analysis that 4G technologies have six times more capacitythantheyreallydo.ThisincorrectcapacityanalysismeansCitisestimationsoftheamountof spectrumneededtosupport4Gtechnologiesarealsoincorrect.Sincecapacityisdirectproportionate totheamountofavailablespectrum,Citisanalysisonspectrumdemandforwidespread4Gdeployment wouldonlyanticipateonesixththeamountofspectrumthanisactuallyneeded. Page 1: Full 4G Can Deliver 5Mbps 100% conversion of 538 MHz allows carriers to offer 5Mbps with10%simultaneoususageduringpeakbusyhour.Thisisincorrectbecauseagainitassumes4Gto have over 6 times the capacity than it really does. Assuming the Citi model is correct (which I cannot confirm based on the details provided), with 6 times less capacity, either the 5 Mbps value must be reducedbyafactorofsixto.83Mbps,or6timestheamountofspectrumisactuallyneeded,3.2GHz,to meetthis5Mbpsobjective. Page3:First,UScarriershavelicensedroughly538MHzofspectrumtoday.And,accordingtotheFCC NationalBroadbandPlan,upto300MHzofadditionalspectrummaybecomeavailableovertime.As

 such,thereislotsofspectrum.538MHzmaysoundlikealotofspectrum,butwhetherthisamount of spectrum is sufficient can only be assessed by modeling demand and calculating the amount of spectrum needed to satisfy that demand using available technologies. The Rysavy Research demand model predicts, consistent with FCC projections, that at least 500 MHz of additional spectrum will be requiredoverthenextdecade.TheCiticonclusionoftheindustrycurrentlyhavinglotsofspectrumis notcrediblebecauseagainitassumestheindustrycanextractfarmorecapacitythanitreallycanfrom currentallocationsusingavailabletechnologies.Thelimitsofavailablespectrumandtheconsequences of insufficient spectrum are detailed in my March 2011 Rysavy Research report, The Spectrum Imperative: Mobile Broadband Spectrum and its Impacts for U.S. Consumers and the Economy An EngineeringAnalysis.4 Page4:Therealimpedimenttodayisthatincumbentwirelessnetworkshavenotconvertedenough oftheircurrentspectrumto4Gservicesanddonothaveenoughexcessspectrumwithintheirpipeline toacceleratethe4Gconversion.Thisisanincorrectconclusionbecauseitisbasedontheassumption that4Gtechnologiesat8.6bps/Hzarefourtimemoreefficientthanenhanced3Gtechnologiesat2.1 bps/Hz. Referring to Figure 1 above (which is consistent with other industry values), 4G technologies suchasLTEintheconfigurationsbeingdeployed5areatmostonly50%moreefficientthanenhanced3G technologies, not 400% more efficient. There are many reasons that operators are deploying LTE, including higher peak speeds, lower network costs, lower latency, qualityofservice control, but huge capacitygainsisnotoneoftheimmediatereasons. Page7,Figure6:ThreeWirelessTechnologyStandards.ThisiswhereCitistatesthatHSPA+has2.1 bps/Hz/sector and 8.6 bps/Hz/sector, values that are completely inconsistent with commonly used industryvalues,asdiscussedabove.Basedontheseerroneousspectralefficiencyvalues,Citicalculates that with 10 MHz for the downlink, a threesector LTE cell would have an aggregate capacity of 258 Mbps6.Intherealworld,however,itwouldhaveadramaticallylowercapacityof42Mbps.7ForHSPA+, Citicalculatesaggregatecapacityof63Mbps8,butintherealworldHSPA+wouldonlyhavecapacityof 31.5Mbps.9

Availableathttp://www.rysavy.com/Articles/2011_03_Spectrum_Effects.pdf. Forexample,LTEisbeingdeployedtodayina2X2MultipleInputMultipleOutput(MIMO)configuration. Calculatedat3sectorsX10MHzX8.6bps/Hz. Calculatedat3sectorsX10MHzX1.4bps/Hz. Calculatedat3sectorsX10MHzX2.1bps/Hz. Calculatedat3sectorsX10MHzX1.05bps/Hz.

 Page13topage14,Figures13to17.CarrierLevelandTargetSpeedScenarios.Citidoesnotdisclose themethodologyoftheirmodelsinthesefigures,butgiventhatthefiguresassumeerroneouscapacity valuesfor3Gand4Gtechnologiesasdiscussedabove,noneofthequantitativeresultsarevalid. Page 16: Question #6: What Can Dish Offer with Its Spectrum. We've used our wireless model to assesswhatsortofspeedsDishNetworkmightofferconsumersifitbuildsa4Gnetworkusing47MHz of spectrum. The data suggests Dish could offer 10% of the population 5Mbps service. This conclusion, assuming Citis model is correct (which I cannot confirm based on the details provided) exceptforspectralefficiencyvalues,overstatesthecapacityofwhatDishcouldofferbyafactorofsix. ThusDishcouldonlyoffer5Mbpstoonesixthof10%(1.7%)ofthepopulationoronesixthof5Mbps (.83Mbps)to10%ofthepopulation.

4 Other Sources of Spectral Efficiency Information

Withoutdoinganexhaustivelistingofothersourcesofspectralefficiencyinformation,thissectionlists several instances that confirm the spectralefficiency values used in my analysis above are consistent with common industry values. It is important to note that spectralefficiency values depend on a considerablenumberofassumptionsaboutnetworkconfiguration,deviceconfiguration,aswellasthe mobilityofusers.Hence,itisnotunusualforvaluesbetweendifferentsourcestovarysomewhat.Such variability,however,isintensofpercentatmost,neverinthehundredsofpercentassuggestedbythe Citireport. 3GPPTechnicalReport25.912V10.0.0(201103),FeasibilitystudyforevolvedUniversalTerrestrial RadioAccess(UTRAandUniversalTerrestrialRadioAccessNetwork(UTRAN),availableat http://www.3gpp.org/ftp/Specs/archive/25_series/25.912/25912a00.zip,page5,Table13.4j,Case3,E UTRA(LTE)2X2MIMO,1.56bps/Hz.3GPPistheorganizationthatdevelopedtheLTEspecificationand soshouldknowsomethingaboutthecapabilitiesofthetechnology.Thistablepresentsthe3GPP assessmentofLTEperformance.Thevalueisslightlyhigherthanthe1.4bps/Hzinmy4GAmericas paperduetodifferencesinunderlyingassumptions.The4GAmericasvaluesaremeanttobemore representativeofrealworldconditionsandrealisticallyavailabledevices.10Adjustedforassumptions, the4GAmericasvaluesareconsistentwith3GPPvalues.Asfor4X2and4X4MIMOconfigurations,these mayeventuallybedeployedbutrepresentsignificantaddedexpenseinsubscriberunits,the infrastructure,orboth,andatthistimealldeploymentsare2X2MIMO. Qualcomm,LTERelease8andBeyond, availableathttp://www.qualcomm.com/common/documents/articles/LTE_Benefits_090409.pdf,page
Notethatmy4GAmericaspaperin2010usedavalueforLTEof1.5bps/HzbutIdecreaseditto1.4bps/Hzin 2011. My explanation in this years report was, LTE spectralefficiency values are slightly lower than last years versionofthispaper,becauseofmorerefinedassumptionsthatbettermatchrealisticavailabledevices.
10

 25,15.1Mbpsin10MHzor1.51bps/Hzforthedownlink.HSPA+islistedas12.5Mbpsin10MHzor 1.25bps/Hz. FederalCommunicationsCommission,MobileBroadband:TheBenefitsofAdditionalSpectrum, availableathttp://transition.fcc.gov/Daily_Releases/Daily_Business/2010/db1021/DOC302324A1.pdf, page14,HSPA+1.08to1.29bps/Hz,andLTE1.36to1.5bps/Hz. WiMAXForum,ComparingMobileWiMAXwithHSPA+,LTE,andMeetingtheGoalsofIMTAdvanced, availableathttp://www.wimaxforum.org/files/wimax_lte/wimax_and_lte_feb2009.pdf,page13,LTE downlinkspectralefficiencyof1.5bps/Hz. InternationalTelecommunicationsUnion,REPORTITURM.2134Requirementsrelatedtotechnical performanceforIMTAdvancedradio,availableathttp://www.itu.int/pub/RREPM.21342008/en, page4,refertobasecoverageurbandownlink,2.2bps/Hz.Thesearetherequirementsthatfuture technologiessuchasLTEAdvancedmustmeet.Thisrequirementisconsistentwiththeupperportionof theLTEbarinFigure1abovewhichshowsLTEachieving2.25bps/Hzforthedownlinkusingfuture enhancementssuchas4X4MIMOwithSuccessiveInterferenceCancellation(SIC),or4X2MIMOwith CoordinatedMultipointTransmission(CoMP),or8X2MIMOwithSingleUser/MultiUser(SU/MU) MIMOswitching.11Inotherwords,evenwitheverycurrentlyplannedimprovementappliedtoLTE, spectralefficiencyonlyreaches2.25bps/Hz.

5 About Rysavy Research

RysavyResearchLLCisaconsultingfirmthathasspecializedinwirelesstechnologysince1993.Projects have included reports on the evolution of wireless technology, test reports, spectrum analysis for broadband services, evaluation of wirelesstechnology capabilities, strategic consultations, system design,articles,coursesandwebcasts,andnetworkperformancemeasurement. Peter Rysavy, president of Rysavy Research, specializes in the capabilities and evolution of wireless technology.Hehaswrittenmorethanahundredandtwentyarticles,reportsandwhitepapers,andhas taughtfortypublicwirelesscoursesandwebcasts.Hehasalsoperformedtechnicalevaluationsofmany wireless technologies including municipal/mesh WiFi networks, WiFi hotspot networks, mobile browsertechnologies,cellulardatanetworks,andwirelessemailsystems. From 1988 to 1993, Peter Rysavy was vicepresident of engineering and technology at LapLink where projects included LapLink, LapLink Wireless, and connectivity solutions for a wide variety of mobile platforms.Priortothat,hespentsevenyearsatFlukeCorporationwhereheworkedontouchscreen anddataacquisitionproducts.
RefertoMobileBroadbandExplosion,at http://www.rysavy.com/Articles/2011_09_08_Mobile_Broadband_Explosion.pdfforanexplanationofthese differentLTEconfigurations.
11

 PeterRysavyisalsotheexecutivedirectorofthePortableComputerandCommunicationsAssociation (PCCA, http://www.pcca.org), a group that evaluates wireless technologies, investigates mobile communications architectures and promotes wirelessdata interoperability. Peter Rysavy graduated with BSEE and MSEE degrees from Stanford University in 1979. More information is available at http://www.rysavy.com.

6 Disclaimer
RysavyResearchprovidesthisdocumentandtheinformationcontainedhereintoyouforinformational purposes only. Rysavy Research provides this information solely on the basis that you will take responsibility for making your own assessments of the information. Although Rysavy Research has exercisedreasonablecareinprovidingthisinformationtoyou,RysavyResearchdoesnotwarrantthat theinformationiserrorfree.RysavyResearchdisclaimsandinnoeventshallbeliableforanylossesor damagesofanykind,whetherdirect,indirect,incidental,consequential,orpunitivearisingoutoforin anywayrelatedtotheuseoftheinformation.