San Francisco Municipal Fiber Network: CTC Report - 10.11.17

ThePotentialforUbiquitous,OpenFiberto
thePremisesinSanFrancisco

PreparedfortheBroadbandforSanFranciscoInitiative
October2017

CTCTechnology&Energy|IMGRebel|BroadbandforSanFrancisco|October2017
Contents
1 ExecutiveSummary.................................................................................................................1
1.1 FibertothePremisesDesignandCostEstimate............................................................2
1.2 DeliveryMechanismAnalysis...........................................................................................5
1.3 PotentialP3Models:DualP3s(DarkFiberandLitServices)oraSingleDarkFiberP3...9
1.4 ProcurementStrategyandRecommendations..............................................................15
2 FTTPDesignandCostEstimate.............................................................................................17
2.1 SummaryofCostEstimates...........................................................................................17
2.2 ApproachtoTechnicalAnalysis......................................................................................19
2.3 FTTPNetworkDesign.....................................................................................................25
2.4 FTTPCostEstimate.........................................................................................................35
3 ThePurelyPublicBroadbandOptionInvolvesHighRisk......................................................44
3.1 Background.....................................................................................................................44
3.2 Benefits...........................................................................................................................46
3.3 Challenges......................................................................................................................46
3.4 FinancialAnalysis............................................................................................................47
4 ThePurelyPrivateOptionWillNotMeettheCitysGoalsofOpen,UbiquitousFibertoAll
HomesandBusinesses..................................................................................................................49
4.1 AT&T...............................................................................................................................50
4.2 Comcast..........................................................................................................................51
4.3 EthericNetworks............................................................................................................52
4.4 Monkeybrains.................................................................................................................52
4.5 Wave...............................................................................................................................53
4.6 Webpass.........................................................................................................................55
4.7 Sonic...............................................................................................................................56
4.8 FactorsThatCanImpactPrivateDeploymentPatterns.................................................58
4.9 BenefitsandChallengesofthePrivateInvestmentModel...........................................59
4.10 RiskofConsolidationandReductioninCompetition.................................................60
5 P3StructureCouldEnabletheCitytoAchievetheGoalofOpen,UbiquitousFiber...........61
5.1 P3Background................................................................................................................61
ii
5.2 TheCoreStructureoftheFormalP3.............................................................................62
5.3 ApplyingtheFormalP3ModeltoBroadbandCreatesSomeRisks...............................63
6 P3ModelAnalysis.................................................................................................................64
6.1 FrameworkGoalsUsedtoDevelopPotentialProjectDeliveryModels........................64
6.2 ClassicP3ModelIllustratesLimitationsofBroadbandP3Approaches.........................65
7 AlternativeP3Approach:DualP3sforDarkandLitFiber....................................................68
7.1 OverviewoftheDualP3Model.....................................................................................68
7.2 HowtheDualP3ModelAddressestheProjectPrinciples............................................72
7.3 FinancialAnalysisandModelsofDualP3......................................................................73
7.4 StructuringtheLitFiberP3toAchieveCityGoals.........................................................88
8 SingleDarkFiberP3Approach..............................................................................................93
8.1 Benefits...........................................................................................................................93
8.2 Challenges......................................................................................................................94
8.3 FinancialAnalysis............................................................................................................95
9 RecommendationsforImplementationandProcurementStrategy..................................101
9.1 APreProcurementCapacityLeasingProcessShouldPrecedetheP3Procurementsoas
toImproveP3BiddingOutcomesbyAccountingforSomeNewRevenueOpportunity.......103
9.2 The Procurement Timeline Should Be Developed in Light of the Complexities and
NewnessoftheStrategy.........................................................................................................105
9.3 TheP3ProcurementProcessShouldBeginasEarlyasPossibleinOrdertoGainInputs
fromBiddersPriortoPuttingTogethertheProposedFundingAuthorization......................107
AppendixA:Glossary..................................................................................................................108
AppendixB:ValueforMoneyAssessmentandValueforMoneyDrivers................................110
AppendixC:RiskMatrix..............................................................................................................116
AppendixD:DirectEconomicBenefitsofFiberConstructioninSanFrancisco.........................125
AppendixE:TechnicalPerformanceMetricsandStandards......................................................127
AppendixF:ListofInformationtoBeAvailabletoBidders........................................................134
AppendixG:PotentiallyUsableCityFacilities............................................................................151
AppendixH:DigitalInclusionandBroadbandAffordability.......................................................153
iii
Appendix I: Potential Technical Framework for the Lit Fiber, MultiProvider, OpenAccess
Network......................................................................................................................................172
AppendixJ:FinancialAnalysisofPurelyPublicModel...............................................................180
Figures
Figure1:DemarcationBetweenCityandPartnerNetworkElementsinLitandDarkModels....19
Figure2:CrowdedUtilityPoleLineinBayview............................................................................21
Figure3:UtilityPoleRequiringMakeReadyinExcelsior.............................................................22
Figure4:TreeTrimmingRequiredonRichmondPoleLine...........................................................23
Figure5:CleanPoleLineinSilverTerrace....................................................................................23
Figure6:HighLevelFTTPArchitecture.........................................................................................26
Figure7:FTTPServiceAreas.........................................................................................................29
Figure8:SampleFTTPAccessLayerDesign..................................................................................30
Figure9:TotalEstimatedCostversusTakeRate(LowCostEstimate)........................................36
Figure10:WaveServiceinSanFrancisco.....................................................................................55
Figure11:SonicServiceinSanFrancisco.....................................................................................58
Figure12:ClassicP3Model..........................................................................................................65
Figure13:DualP3Model..............................................................................................................68
Figure14:DualP3Model..............................................................................................................74
Figure15:FTTPSubscribersCanAccessCivicNetworkDirectlyorviaanRSP'sNetwork...........92
Figure16:P3ValueDrivers.........................................................................................................111
Figure17:CityFiberFacilities.....................................................................................................151
Figure18:PropertyOwnedbyCityandCounty.........................................................................152
Figure19:PercentofHouseholdswithMinimum200KbpsService..........................................155
Figure20:PercentofHouseholdswithMinimum10Mbps/1MbpsService.............................156
Figure21:YearlyHouseholdIncomebyCensusTract................................................................157
Figure22:ResponsibilitiesinaLayeredServiceModel..............................................................173
Figure23:MultipleServicesProvidedovertheSameFTTPNetwork........................................174
Figure24:ServicefromRSPAIsProvisionedtotheCustomerONTviaanEVC........................174
Figure25:APointtoPointTopology.........................................................................................175
Figure26:AMultipoint,orLAN,Topology.................................................................................175
Figure27:ATreeTopology.........................................................................................................175
Figure28:LayeredOperationalResponsibilitiesfortheOpenAccessFTTPNetwork...............176
Figure29:TroubleReportWorkflow..........................................................................................177
Figure30:ServiceProvisioningWorkflow..................................................................................179
iv
Tables
Table1:EstimatedLitFTTPCost.....................................................................................................4
Table2:EstimatedDarkFTTPCost.................................................................................................5
Table3:EstimatedConnectionFees.............................................................................................11
Table4:EstimatedDarkFiberP3MilestonePaymentandAvailabilityPayment........................11
Table5:EstimatedLitFiberP3MilestonePaymentandAvailabilityPayment............................12
Table6:EstimatedLitFTTPCost...................................................................................................18
Table7:EstimatedDarkFTTPCost...............................................................................................19
Table8:FieldSurveyFindingsandMakeReadyAssumptions.....................................................20
Table9:EstimatedLitFTTPCost...................................................................................................35
Table10:EstimatedOutsidePlantCosts(DirectionalBoringCostEstimate)..............................38
Table11:EstimatedOutsidePlantCosts(SidewalkRestorationCostEstimate).........................39
Table12:EstimatedCentralNetworkElectronicsCosts...............................................................40
Table13:PerSubscriberActivationCostEstimates.....................................................................42
Table14:EstimatedDropandLateralCosts(DirectionalBoringEstimate).................................43
Table15:EstimatedDropandLateralCosts(SidewalkReplacementEstimate)..........................43
Table16:BaseCaseDarkFiberP3CapitalExpenditure...............................................................76
Table17:BaseCaseLitFiberP3CapitalExpenditure...................................................................78
Table18:BaseCaseBusinessPassingFeeAssumptions..............................................................81
Table19:EstimatedDarkFiberP3MilestonePaymentandAvailabilityPayment......................83
Table20:EstimatedLitFiberP3MilestonePaymentandAvailabilityPayment..........................85
Table21:EstimatedConnectionFees(BaseCaseandBestCase)................................................86
Table22:SensitivityAnalysisofDualP3ResidentialConnectionFees........................................87
Table23:SensitivityAnalysisofSingleDarkFiberP3ResidentialConnectionFees....................97
Table24:SensitivityAnalysisofSingleDarkFiberP3ResidentialConnectionFeeswithAdditional
Fees...............................................................................................................................................99
Table26:EstimateofJobCreationDuringNetworkConstruction............................................126
Table27:AvailableInformationonExistingInfrastructure........................................................134
Table28:AvailableInformationonROWandPublicProperty..................................................137
Table29:AvailableInformationonPotentialUsers...................................................................139
Table30:AvailableInformationonConstructionStandardsandRestrictions..........................140
Table31:NumberofCensusTractsbySubscriptionRate..........................................................154
Table32:MunicipalRetailModelFinancialSummary...............................................................181
Table33:IncomeStatement.......................................................................................................183
Table34:CashFlowStatement..................................................................................................184
Table35:CapitalAdditions.........................................................................................................186
Table36:LaborExpenses............................................................................................................188
Table37:OperatingExpensesandP&IPayments......................................................................190
v
Table38:BaseCaseScenario......................................................................................................191
Table39:TakeRateDropsby10Percent...................................................................................192
Table40:VoiceandVideoRevenueShareDropsby50Percent...............................................192
Table41:ResidentialInternetFeeDecreasesby$5perMonth................................................193
Table42:ResidentialInternetFeeIncreasesby$5perMonth..................................................193
Table43:ReduceOperatingandStaffingExpensesby25Percent(ContentFeesRemaintheSame)
.....................................................................................................................................................194
Table 44: Increase Operating and Staffing Expenses by 25 Percent (Content Fees Remain the
Same)..........................................................................................................................................194
Table45:UseFTTPHighCostEstimate......................................................................................195
vi
1 Executive Summary
Broadband networks rank among the most important infrastructure assets of our timefor
purposesofeconomicdevelopmentandcompetitiveness,innovation,workforcepreparedness,
healthcare, education, democratic discourse, and environmental sustainability. The City and
County of San Francisco (City) recognizes the need for essential highspeed, affordable
broadbandservicesintheCityandisconsideringhowtofacilitatedevelopmentofsuchservices,
includingtothosewithnoorlimitedaccessorabilitytoaffordhighqualityservices.
The City commissioned this report in early 2017 from a Project Team that includes CTC
Technology & Energy, a public sector broadband consultancy, 1 and IMG Rebel, a financial
advisoryandP3structuringfirm.2
TheCitydirectedtheProjectTeamtodevelopadeliverystrategyfordeliveryofservicesover
ubiquitous fibertothepremises (fiber connectivity to every home and business in San
Francisco) that would be open, offering equal potential access to the network by multiple
entitiessoastoenableandstimulatecompetition.TheCitydirectedtheProjectTeamtoconsider
thepotentialfor(1)fullypublic,(2)fullyprivate,or(3)publicprivatepartnershiparrangements
forfinancing,designing,constructing,maintaining,andoperatingthenetwork.Inadditiontothe
ubiquityandopennessparametersforthefiber,theCityarticulatedthefollowingkeyprinciples
toguidetheeffort:
Equity: Every resident and business in San Francisco should have access to fast and
affordablebroadbandconnectivitynecessarytoparticipateandthriveinthe21stcentury.
With13.2percentoftheCitysresidentslivingbelowthepovertylevel,affordabilityof
broadbandservicesisasignificantfactorintheCitysdigitalinclusionefforts.The City
intendstoprioritizeprovidingservicetotraditionallyunderservedhouseholds.
1
CTCTechnology&Energyisa30yearoldbroadbandconsultancy,focusedonpublicsectorandnonprofitefforts
toimprovegovernmentandutilitycommunications,expandbroadbandinternetservice,anddevelopnew
strategiesfordeliveryofnetworkcapabilities,includingbroadbandpublicprivatepartnershipsandinnovative
collaborations.CTCoffersstrategy,engineering,businessplanning,projectmanagement,andgrantplanning
services.CTCsclientsincludethecitiesofSanFrancisco,Seattle,Portland(Ore.),Washington,D.C.,Boston,
Atlanta,SanAntonio,Vancouver,B.C.,andhundredsofothercitiesandcountiesofallsizes.CTCistheleading
broadbandconsultancyworkingwithapplicationofP3conceptstobroadbandinsuchprojectsasKentuckyWired
andtheWestminster,Maryland,fiberP3.CTCalsohelpeddefinetheframeworkforbroadbandP3sintheBenton
FoundationspaperonTheEmergingWorldofBroadbandPublicPrivatePartnerships
(https://www.benton.org/sites/default/files/partnerships.pdf).
2
IMGRebelisamongtheworldsleadingP3advisors.IMGRebelhasadvisedonmorethan50P3transactions
acrosssectorsasleadtransactionadvisor,includingmanyfirstofitskinddeals.IMGRebelhassupportedmostof
theleadingU.S.federal,state,andlocalP3practitioners,includingthestatesofFlorida,Texas,andVirginiaandthe
FederalHighwayAdministrations(FHWA)OfficeofInnovativeProjectDeliveryasafinancialand/orP3advisor.
1
Jobs,Innovation,Growth:Investmentinnewconnectivitywillresultinincreasedlocal
employmentandprovidenumerouseconomicdevelopmentadvantagesfortheentire
Citytocontinuetobetheinnovationcapitaloftheworld.
LocalAuthority:ASanFranciscoownedfibernetworkwillprotecttheCitysstated
valuesofequity,privacy,netneutrality,andopenaccess.
ToidentifyasustainableapproachtofacilitatingboththeCitystechnicalsolution(ubiquitous,
open access FTTP) and policy goals (with a focus on equity), the Project Team undertook the
followinganalyses:
1. Developedatechnicaldesignandcostestimatefordeploymentofafibernetworkthat
wouldconnecteveryhomeandbusinessinSanFrancisco
2. Consideredthreedeliverymodelsforexecutionoftheubiquitous,openFTTPgoal:purely
public,purelyprivate,andpublicprivatepartnership(P3)
3. Developed two innovative P3 delivery strategies and financial models for financing,
designing, constructing, maintaining, and operating a network that meet the Citys
requiredparameters:
a. A Dual P3 Model in which the City would develop two P3s, one for dark
fiber/outside plant only, and one to deliver lit services to every home and
business,whilealsoprovidingwholesaleservicestocompetingserviceproviders
b. ASingleDarkFiberP3ModelinwhichtheCitywoulddeveloponeP3fordarkfiber,
and then lease dark fiber to competing service providers who would light and
operatethefiber
4. DevelopedaprocurementstrategyfortheP3thatappliesP3bestpracticestothenew
assetclassofbroadbandwhileseekingtoappropriatelyimprovethebroadbandmarket
andleveragethepotentialdemand(onthepartofbothconsumersandinternetservice
providers)fornewcompetitionopportunities
1.1 FibertothePremises Design and Cost Estimate

CTCs engineers prepared a network design and cost estimate for an open fiber optic
infrastructure to connect to all homes and businesses in San Francisco. 3 The recommended
architectureisahierarchicaldatanetworkthatprovidescriticalscalabilityandflexibility,bothin
initialnetworkdeploymentandincapacitytoaccommodatetheincreaseddemandsoffuture
applicationsandtechnologies.Itincludesthefollowingcharacteristics:
3
ThisanalysisisdescribedindetailinSection2below.
2
Capacityabilitytoprovideefficienttransportforsubscriberdata,evenatpeaklevels
Availabilityhighlevelsofredundancy,reliability,andresiliency;abilitytoquicklydetect
faultsandreroutetraffic
Failsafeoperationphysicalpathdiversitytominimizeoperationalimpactresultingfrom
fiberorequipmentfailure
Efficiencynotrafficbottlenecks;efficientuseofresources
Scalabilityabilitytogrowintermsofphysicalserviceareaandincreaseddatacapacity,
andtointegratenewertechnologies
Manageabilitysimplifiedprovisioningandmanagementofsubscribersandservices
Flexibilityabilitytoprovidedifferentlevelsandclassesofservicetodifferentcustomer
environments; can support an openaccess network or a singleprovider network; can
provideseparationbetweenserviceprovidersonthephysicallayer(separatefibers)or
logicallayer(separateVLANorVPNprovidingnetworkswithinthenetwork)
Securitycontrolledphysicalaccesstoallequipmentandfacilities,plusnetworkaccess
controltodevices
WithinputfromCitypublicworksandengineeringstaffregardingbothcostsandCityprocesses,
the Project Team then developed cost estimates, 4 assuming City cost structures, 5 for the
following:
4
ThenetworkdesignandcostestimatesarebasedondataandinsightgatheredbyCTCengineersindiscussions
withCitystakeholdersandanextensivefieldanddesksurveyofcandidatefiberroutes.Thedesignandcost
estimationmethodologyarediscussedindetailbelow.
5
AllofCTCscostestimatesforthisreportarebasedonourexperiencewithmunicipalcoststructures,ratherthan
thoseoftheprivatesector.Asaresult,someoftheseestimateswillvaryfromthelikelycoststhatwouldbe
incurredbyaP3concessionaire.Wechosethismethodologyfortworeasons:First,whiletheprivatesectorcan
realizesomekindsofefficienciesthatpublicentitiescannot,thisisunlikelytoimpactthemostsignificantofthe
costareas,labor.WepresumebasedonCityguidancethatanyP3awardeewillberequiredtopaysalariesand
benefitsincompliancewithCitypolicy,whichmaybehigherthanprivatemarketrates.WenotethattheP3
concessionairemaybeabletoimproveonourcostestimatesthroughthebenefitsofscale,particularlyinthe
areasofequipmentpurchases,customerservice,andotheroperatingcosts,butgiventheCityslivingwage
commitments,significantlyreducedcostsforconstructionlaborareunlikely.Second,giventhatthisproject
representsafirstofitskindevaluation,theredoesnotexistinthebroadbandfieldasetofdataregardingprivate
costsandbiddinglevelsforFTTPP3s.ThisisunlikeP3planninginothersectors,wheremultipleP3shavebeen
executedandpublicsponsorsareabletodrawondatadevelopedinearlierprojectsandbids.Asaresult,we
deliberatelychosetheconservativemethodologyofestimatingbasedonmunicipalcoststructures.Weanticipate
thatP3bidderswillbeabletogreatlyimproveonourconservativeestimatesintheareasofequipment
3
AlitFTTPnetwork,includingthecoststodeployFTTPoutsideplantinfrastructure,as
well as all electronics, consumer drops, and customer premises equipment (CPE). This
estimateshowsthetotalcapitalcosttobuildanFTTPnetworktosupportubiquitous1
Gbps(gigabitpersecond)dataservice
AdarkFTTPnetwork,includingthecosttodeployonlytheoutsideplantinfrastructure.
ThisisthetotalcapitalcostfortheCitytobuildadarkFTTPnetworkforusebyoneor
moreentities
TheProjectTeamestimatesthatthelitFTTPnetworkdeploymentwillcost$1.9billion,6inclusive
of outside plant construction labor, materials, engineering, permitting, pole attachment
licensing, network electronics, drop installation, CPE, and testing. 7 We estimate the average
constructioncostperpassing8willbe$2,050.9ThecostsaresummarizedinSection2.4.10
Table1:EstimatedLitFTTPCost
CostComponent LowCostScenario HighCostScenario

OutsidePlant $700million $760million
CentralNetworkElectronics 112million 112million
FTTPServiceDropandLateral
680million 740million
Installations
CustomerPremisesEquipment 290million 290million
TotalEstimatedCost: $1.8billion $1.9billion
WeestimatethatadarkFTTPnetworkdeploymentwillcost$1.5billion,11inclusiveofoutside
plant construction labor, materials, engineering, permitting, pole attachment licensing, and
procurement,equipmentmaintenance,customerservice,andothercostsassociatedwithoperationsofthelit
fibernetwork.
6
ThisistheHighCostScenarioasdescribedbelow.
7
BecausetheCitydirectedthattheanalysisassumethenetworkwillbedeployedinautilitymodeltoallhomes
andbusinesseswithinSanFrancisco,theestimatedtotalcostbreakdownassumesthat100percentofresidents
andbusinesseswillbeconnectedtotheservice.Inadeliverymodelinwhichthenetworkisbuiltonlytothose
consumerswhochoosetopurchaseservice,thetotalcosttoconnectandassociatednetworkdeploymentcosts
wouldbelower.
8
Thepassingcountincludesindividualsingleunitbuildingsandunitsinmultidwellingandmultibusiness
buildingsassinglepassings.Inlargermultidwellingunits(MDU),thecostestimateassumesthatexistingwiring
willbeusedtoservetheunits.WeassumethateachunitinlargeMDUswillcost$500.
9
10
Weapplyanuncertaintyof+/20percentbasedonourexperienceandtypicalindustrypractice.
11
4
lateraldropandmaterials.Thisestimate,whichdoesnotincludeanyelectronicsorsubscriber
equipment,issummarizedinSection2.1.2.12
Table2:EstimatedDarkFTTPCost

FTTPServiceDropandLateralInstallations 680million 740million
TotalEstimatedCost $1.4billion $1.5billion
Thecostestimatesarebasedonacombinationofaerialandundergroundconstruction,anda
range of potential construction scenarios. The lowcost scenario assumes directional boring
throughout the City using test pitting to minimize outages to existing utilities. The highcost
scenariorequireshighlevelsofsurfacerestoration(i.e.,completesidewalkreplacement).
The dark fiber estimate assumes that the City or its partner constructs and owns the FTTP
infrastructureuptoademarcationpoint(anetworkinterfacedevice,orNID)ateachresidence
and business, and that the dark fiber backbone, distribution, and drop fiber are then lit by a
separateentity.Thelitfiberentitywouldberesponsibleforallnetworkelectronicsandcustomer
premisesequipment(CPE).Thenetworkarchitecturecouldsupportmultipleentitiesusingfiber
assetsonaleasedbasisorotherbasis.
TheoutsideplantconstructionlandscapeintheCityhaschangedsinceCTClastestimatedthe
costoffiberconstructionforSanFranciscoin2009(andevensincetheCostQueststudytwoyears
ago).Moreprovidershaveenteredthemarketorexpandedtheirservicefootprints,whichmeans
more crowding on utility poles and higher pole makeready costs. Further, new surface
restorationstandardswillincreasethecostofundergroundconstructionandtheamountoftime
neededtoconstructthenetwork.Utilitypolereplacementcostswillincrease,too,becausetaller
poleswillbeneededinmanylocationstosupportthenewfiberattachment.Andexistingpoles
continuetoageincreasingthenumberofreplacementsneeded.
1.2 Delivery Mechanism Analysis

TheProjectTeamwentthroughastructuringexercisetodeveloparecommendedapproach.The
focuswastoidentifyadeliverymodelthatwould:
Meetthegoalsspecifiedforthisinitiative(open,ubiquitousfiberopticstoeveryhome
andbusiness,withopportunityforcompetitionoverthenetwork)
12
Weapplyanuncertaintyof+/20percentbasedonourexperienceandtypicalindustrypractice.
5
Createabroadbandmarketthatprovideslongtermbestvaluefortaxpayers,users,and
theCity
Createafeasiblebusinesscasefortheproject
Createamarketablestructurethatcanbefinanced
TheProjectTeamconsideredthreedeliverymodels:purelypublic,purelyprivate,andpublic
privatepartnership(P3).
1.2.1 Fully Private or Public Delivery Models Do Not Satisfy the Citys Goals
TheProjectTeamfirstevaluatedthetwopolaroptionsthathaveexistedasbroadbanddelivery
modelsinthepast:eitherfullyprivateorfullypublicnetworks.13
Private broadband investment is resulting in new broadband competition in unevenly

distributedwaysacrossSanFrancisco.AT&Tisupgradingitsnetworktofiberincertainareas,but
not, to our knowledge, on a ubiquitous basis. Comcast has recently upgraded its residential
service footprint with DOCSIS 3.1 electronics that will allow speeds of up to a gigabit in the
downstreamdirection,butthisnetworkstillreliesoncoaxialcablefordistributionandthegigabit
serviceispricedinexcessof$150permonth.14
Onthecompetitivesideoftheindustry,anewclassofcompetitorsincludingSonic,Webpass,
andMonkeybrainsismakingimportantinvestmentsinfiberandothertechnologiesinsomeSan
Francisco neighborhoods, but these new networks are available only in certain areas and to
certainbuildingsorconsumers.
Both incumbents and competitors deploy new or upgraded infrastructure on a limited basis,
focusedongeographicareaswherecostsarelowestandrevenuesandreturnoninvestmentare
likelytobehighest.Asaresult,mostnetworkinvestmentsaregenerallyunevenacrosstheCity,
withsomeSanFranciscoresidentsofferedfarbetteroptionsthanothers,andmanyunableto
affordthehighercapacityoptions.Inaddition,noneoftheincumbentorcompetitiveproviders
in San Francisco operates its network on an open basis, allowing its competitors to reach
consumersoveritsinfrastructure.
13
ThepublicandprivatedeliverymodelsaresummarizedanddiscussedinmoredetailinSections3and4below.
14
ComcastrepresentativestoldCTCandCitystaffinaSeptember2017meetingthattheresidentialgigabit
product(with35Mbpsupstream)isofferedattheeverydaypriceof$160permonth,plusamodempurchaseof
$200or$10permonth.Introductory,promotionalpricingforthesameproductis$70permonth,plusthecostof
equipment,basedona24monthcontract.Gigabitservicewillbeavailabletobusinessesbytheendoftheyear,
butathigherpricing.
6
WethereforeconcludethattheprivatesectorisnotcurrentlypoisedtodelivertheCitysstated
goalsofubiquitous,open,fiberbasedservicethroughoutSanFrancisco.
Apurelypublicnetworkcouldmeetgoalsforubiquityandopenness,buttheCityhasraiseda
numberofareasofconcern,includinginternalcapacitytobuildandoperateanewutility;public
finance challenges in light of the many important project vying for public financing, and the
substantialrisksofproceedingonapurelypublicbasis.
Indeed,inapurelypublicmodel,theCitycouldmeetitspolicygoalsbutwouldassumeallproject
risk, including construction, operations, performance, customer service, demand and market
considerations,andtechnologychange.Aparallelsetofconcernsrelatestopublicsectorcapacity
and interest in building, maintaining, and operating a communications utility, including the
particular challenge of operating a communications network in an area of the country where
hiringprofessionalsinthatfieldismostcompetitive.
AndthemarketriskassumedbytheCityinapurelypublicmodelisconsiderable.Toputthe
initiativeintocontextofothercityledeffortsandthelargerbroadbandmarket,wedevelopeda
financialmodelforatraditionalmunicipalnetworkinwhichSanFranciscowouldbuild,own,and
operate the network, and provide retail services to residents and businesses. 15 This model
suggests that the network would require a take rate of 45 percent (assuming the lower cost
estimate) to 53 percent (assuming the higher cost estimate) to break even, both of which
representveryaggressivegoals.16Fewmunicipalnetworkshavemanagedtoreachthislevelof
penetration.
1.2.2 Application of P3 Delivery to Broadband Involves Novel Questions and

Opportunity
P3 delivery represents a logical alternative opportunity for deployment, maintenance, and
operationsofubiquitousfiber.ItenablestheCitytoownthecoreinfrastructureanduseitto
achievepolicygoalssuchascompetitionandequityatthesametimeasshiftingsomeriskto
theprivatesectorandleveragingprivatesectordeliverycapabilities.17
InaformalP3structure,aselectedprivatepartnertakesresponsibilityforsomecombinationof
design, construction, financing, operations, and maintenance, 18 at least partially funded or
guaranteedbythepublicpartnerovertheperiodoftheconcession.
15
ThemodelisdiscussedindetailinSection3below.
16
Takerateisoneofthetraditionalmetricsforevaluatingtheperformanceofbroadbandnetworks,bothpublic
andprivate.
17
TheP3deliverymodelissummarizedanddiscussedinmoredetailinSections58below.
18
FinancialStructuringofPublicPrivatePartnerships(P3s),U.S.DepartmentofTransportation,2013
http://goo.gl/gCJIZK(accessedFebruary2017).
7
Formal P3s have emerged in the broadband sector in other parts of the world where
infrastructureismorefrequentlydeliveredthroughaP3thanintheUnitedStates.Inthepast
fewyears,arangeofcreativebroadbandP3structureshasstartedtoevolveintheUnitedStates.
ThenatureoftheseP3srangefrompublicsectorfacilitationofprivatebroadbandinvestment
(through process, regulation, and economic development incentives) to public deployment of
infrastructureforusebytheprivatesector.
WhatisnovelabouttheBroadbandforSanFranciscoInitiativeisthatitcontemplatesapplication
oftheclassicP3structuretoabroadbandnetworkthatwouldreach100percentofthehomes
andbusinessesinthecommunity.
TheapplicationofP3structuresandprinciplestobroadbandisnotstraightforward.FormalP3s
developedasaneffectivetoolforgovernmentstoshiftappropriateriskstotheprivatesectorin
infrastructureareasthathavetraditionallybeenentirelytheresponsibilityofgovernment,such
astransportation,publictransit,prisons,andwaterandsewerutilities.
In contrast, in the broadband environment, application of a P3 involves an opposite set of

relationships.Inanareathattraditionally,intheUnitedStates,hasbeenprimarilytheprovince
ofprivatecompanies,broadbandP3scontemplateallocationofsomeriskandresponsibilityto
thepublicsector.
While government has long had a critical and insufficiently acknowledged role in shaping the
broadbandmarket,thatrolehasoftenbeenlimitedtoregulation,deregulation,preferredaccess
to assets, and so onas well as certain kinds of public subsidy for private networks such as
throughtheFederalCommunicationsCommissionsHighCostandConnectAmericaFunds.The
P3structurecontemplatesassumptionofthetypeoffinancialriskthattraditionallyhasbeenthe
domainofprivateinvestorsinasomewhatcompetitivemarketplacewithconstantlyshifting
servicesandconsumerpreferences.
1.2.3 The Classic P3 Model Applied to Broadband Creates Some Challenges

To understand how a formal P3 can be structured for fibertothepremises, which has never
before been done in the United States, we first considered the classic model and then
developedalternativemodelsthatwebelievewillbettersuitthisnewclassofprojects.
In the Classic P3 Model, the City would negotiate a longterm contract with a P3 broadband
concessionairewhowoulddesign,construct,finance,operate,andmaintainthenetwork.The
Citywouldsponsorandowntheinfrastructure,19buttheconcessionairewouldcontrolaccessto
19
ThetermsponsorcomesfromthetermProjectSponsorwhichistheentitywiththeoverallaccountabilityfor
theproject.Inthiscase,theCityofSanFranciscoisthesponsoroftheproject.TheCityhasultimateresponsibility,
andisaccountabletothepublic/taxpayers.
8
it.TheCitywouldachievesomeofitsprimarygoalsunderthismodel.TheCityscontractwith
the concessionaire would also shield the City from some longterm project risks. The
concessionairewouldholdthelongtermperformancerisksassociatedwiththeproject,aswell
assomeportionoftheconstructionandmarketrisk.
However,intermsoftheprinciplesthatguideourfinancialandstrategicanalysis,theClassicP3
Modelraisesconcerns.First,asingleconcessionairewithalongtermconcession(asinatollroad
scenario)wouldnotdelivertheopen,competitiveenvironmenttheCityseeks.Second,uncertain
revenuesincreaserisktotheprojectbusinesscase.Potentialcommercialrevenuesunderthis
modelarehighlyuncertain,duebothtotheunpredictablenatureofbroadbandequipmentand
networktechnologiesandthematurityoftheSanFranciscomarket.Transferringthisrevenue
risktotheconcessionairewouldrequiretheCitytopayaveryhighcostoffinancing.Third,risk
mitigationisrequiredtoattractinvestorsandfinanciers.Duetothehighlyuncertainrevenue
streamsinthismodel,itisunlikelythatfinanciersandinvestorswillbewillingtoinvestunless
they would have a quasimonopoly, including ratesetting powers, and/or some type of
contribution from the City (e.g., a guarantee, risksharing mechanism, or partial payment) to
mitigaterisk.
1.3 Potential P3 Models: Dual P3s (Dark Fiber and Lit Services) or a Single
Dark Fiber P3
We then developed two alternative P3 approaches to meet the Citys goals, one that
contemplatesP3deliveryofbothdarkfiberandlitservices(aDualP3Model)andonethat
contemplatesadarkfibermodelonly(SingleDarkFiberP3Model).20
1.3.1 Potential Dual P3 Model: Separate the Dark Fiber and Lit Service P3s
TherecommendedDualP3ModelrecognizesthatthemostcriticalchallengeintheClassicP3
Modelisthedisparitybetweentheriskprofilesassociatedwiththedarkfiberinvestmentand
theprovisioningoflitservices.CombiningthesetwodistinctriskprofileswithinoneP3makes
theClassicP3modelhighlychallengingfromtheperspectiveofdevelopingasustainablebusiness
model.
Ourapproachistocreateasmuchmarketcompetitionaspossibleatboththedarkfiberandlit
serviceslayers.Inthisapproach,theCitysconcessionaireswouldnotentertheinternetbusiness,
butratherwouldenablerobustprivatesectorcompetition.
Assuch,therecommendedDualP3ModelcomprisestwoseparateP3contracts:
20
ThisanalysisisdescribedindetailinSections7and8below.
9
1. DarkFiberP3:Includesdesign,build,partialfinance,andoperationsandmaintenance
(O&M)oftheoutsideplantinfrastructure,includingthefiberitself.ThisisalongtermP3
contract(likely30years)fordarkfiberinvestmentsthatincludespartialpublicfinancing.
2. Lit Fiber P3: Includes equipment and O&M investments, and provision of baselevel
service.ThisisashortertermP3contract(likelysevento10years,timedtomatchthe
usefullifeofmostofthenetworkequipment).
Fromapracticalstandpoint,theDarkFiberP3investmentscanbeseparatedfromtheLitFiber
P3 because interrelationships between the two are limited. This approach results in a more
explicit distinction between the dark fiber network provider and the service provider on that
network,withtheintenttoenablemoretechnologicalflexibilityandcompetitioninlitservice
delivery.
1.3.2 Financial Analysis and Modeling of Dual P3 Model

The Project Team developed a financial model for the Dual P3 structure to understand the
required Connection Fee per residence, required Connection Fee per business, and expected
payments to the P3 concessionaires based on cost estimates for full FTTP deployment. The
analysisalsosoughttounderstandhowcostandrevenuevariationsinfluencethesefees.21
Table3showsthata$51permonthresidentialConnectionFeeanda$73permonthaverage
businessConnectionFeewouldcoveralloftheprojectcosts(includingthesubsidy)andmake
the project financially feasible under the base case assumptions. Because these results are
uncertain(whichistypicalforthisstageofprojectdevelopment),thesensitivityanalysisshows
thelikelyrangeofrequiredpaymentsbasedondifferingassumptionsforsixmainvariablesthat
canaffecttheConnectionFees:capitalcosts,operationscosts,contracttermoftheLitFiberP3,
levelofsubsidy,additionalrevenues,andfinancingconditions.
Uncertainties related to capital costs are the most significant, followed by the financing
uncertaintiesandalternativestrategiesregardingthesubsidyforlowincomeresidents.Asthe
projectprogressesandadditionalanalysisandinformationisadded,theseuncertaintieswillbe
reduced. 22 The sensitivity analyses found that the worstcase scenario garnered a residential
21
Thisanalysisisbasedonpreliminaryinformationandassumptionsasavailablefortheprojectatthistime,andare
uncertainwhichiscommonatthisstageofprojectdevelopment.Weexpectthatasmoreanalysisisconductedon
theproject,andtheprojectprogresses,wewillbeabletorefineouranalysisbasedonmoresubstantiveinformation
regardingexpectedupfrontrevenues,commitmentofpublicresources,andsoon.Additionally,duetothelevelof
uncertainty, most results of the financial analysis are represented as ranges rather than exact outcomes, as
appropriate.
22
AmongthequestionstheProjectTeamconsideredwashowtostructuretheLitFiberP3andcraftanRFPforaLit
FiberConcessionairethatwouldservetheCitysdefinedprinciplesandgoals.Amongothermatters,theLitFiber
approachcontemplatesdeliveryofabaselevelofbroadbandservicetoeverySanFranciscohouseholdandbusiness;
10
ConnectionFeeof$67permonth,whereasthebestcasescenarioresidentialConnectionFeeis
$26permonth.
Table3:EstimatedConnectionFees
Output BaseCase Range

ConnectionFeeresidential $51/passing/month $26$67/passing/month
ConnectionFeebusiness(average) $73/passing/month $38$97/passing/month
As Table 4 illustrates, the Dark Fiber P3 Concessionaire is expected to require a $1.36 billion
milestonepaymentatsubstantialcompletion(i.e.,paidinyears2and3),incombinationwith
annualavailabilitypaymentsof$36millionfor30yearstofinancetheprojectandcoverallofthe
darkfiberprojectcosts(includingthesubsidy)underthebasecaseassumptions.Underthemost
conservative, worstcase scenario, a milestone payment of $1.63 billion in combination with
annualavailabilitypaymentsof$45millionwouldcoveralloftheprojectcosts.
Table4:EstimatedDarkFiberP3MilestonePaymentandAvailabilityPayment

$1.087billion
DarkFiberMilestonePayment(Total) $1.358billion
$1.629billion
$27million/year
DarkFiberAvailabilityPayment(Annual) $36million/year
$45million/year
Thetwopartpaymentapproach(i.e.,amilestonepaymentincombinationwithanavailability
payment) is intended to optimize the projects financing structure. If the City would prefer a
payment structure without a milestone payment, the financing costs would be significantly
higherandanannualavailabilitypaymentof$129millionover30yearswouldberequiredunder
thebasecasescenario.Underthemostconservative,worstcasescenario,therequiredannual
availabilitypaymentwouldbe$173million.
As Table 5 illustrates, the Lit Fiber P3 Concessionaire is expected to require a $147 million
milestonepaymentatsubstantialcompletion,incombinationwithannualavailabilitypayments
of $62 million, to finance the project and cover all of the lit fiber project costs (including the
subsidy)underthebasecaseassumptions.
definingwhatthatbaselevelofserviceshouldbeisthekeychallengeofstructuringtheLitFiberP3.Insummary,
theProjectTeamdeterminedthatprovisionofalitfiberservicebytheCitysconcessionairewouldcreatesignificant
challenges,eitherthroughreductioninthepotentialcommercialmarket(iftheConcessionaireofferedahighspeed
productaspartofthebaselevelservice)orthroughinefficiency(iftheConcessionaireofferedalowbandwidth
productthatwouldwastethecapacityofthenetworkandentailsignificantcostrelativetobenefit).Thisanalysisis
describedindetailinSection7below.
11
Table5:EstimatedLitFiberP3MilestonePaymentandAvailabilityPayment

$123million
LitFiberMilestonePayment(Total) $147million
$176million
$42million/year
LitFiberAvailabilityPayment(Annual) $62million/year
$76million/year
1.3.3 Potential P3 Model: Single Dark Fiber P3 Only

TheProjectTeamalsodevelopedasecondpotentialbusinessstructuretomeettheCitysgoals
forthisinitiativeinwhichtheCitywouldnotbuildlitservicesintotheP3planbutwoulddevelop
a single dark fiber P3. The dark fiber structure would be designed to maximize competition
through leasing of dark fiber to competing ISPs rather than through Cityfunded provision of
competitionatthelitserviceslayers.
1.3.4 Benefits
Solongasamarketemergesfordarkfiberleasesthatwouldenablenewcompetition,aSingle
Dark Fiber P3 Model offers greater simplicity, lower cost, and potentially lower risk and
reduced opposition from incumbent providers that oppose the Citys efforts to increase
broadband competition. With a Single Dark Fiber P3, the City would spend dramatically less
because it would not take on the financial commitment necessary for a Lit Fiber Concession.
DevelopingaSingleDarkFiberP3alsoreducescomplexityandadministrativeburdenfortheCity.
Unlikelitservicesandequipment,darkfiberisastable,predictableasset.
ThemodelalsoremovesfromtheCityspurviewnetworkoperations.Ratherthandetermining
parametersandtermsfortheseareasofoperationthroughtheLitFiberPG,theCitysrolewould
belimitedtothesimplerdarkfiberlayerofthenetwork.
TheSingleDarkFiberP3ModelalsopotentiallyreducesthechallengesinherentintheCityswork
to promote competition and better broadband. Regardless of merits, City efforts to expand
broadbandandbroadbandcompetitionhavebeenaggressivelyopposedbyincumbentphone
andcablecompaniesnationwide.Inadarkfibermodel,theCity(anditsConcessionaire)playno
roleinprovidingservicestothepublic,butratherprovideadarkfiberplatformtoenablethe
private sector to enter the service business. Frankly, incumbents are likely to oppose this
initiativeregardlessofhowlimitedtheCitysroleis,butthecustomaryincumbentclaimofunfair
competitionfromthepublicsectorisundercutbythemodel.
Further, by removing the City from direct roles in shaping services, pricing, or performance
standards,theSingleDarkFiberP3preservesthefullretailmarketforcompetingISPsrather
thanusingaP3todeliversomelevelofservicestothepublic.
12
1.3.5 Challenges
Though simple and efficient, the single dark fiber P3 model reduces options with regard to
fundingandtocompetition.Withoutaservicedeliveredtoeveryhomeandbusiness,aproperty
relatedfeecannotbeusedfortheConnectionFeetofundtheproject.23
Thecompetitionconsiderationsarealsocomplexwithrespecttocomparisonbetweenthesingle
dark fiber P3 and the dual P3 models. While the single dark fiber P3 creates unprecedented
opportunityatthedarkfiberlayer,italsodoesnotdelivercompetitionaseffectivelyasaLitFiber
3potentiallycan.Themarket,dependingonhowthefiberispriced,maysupportonlyoneortwo
dark fiber lessees for the foreseeable future. In contrast, a lit network can provide very
inexpensiveopenaccesstomanyserviceproviders,includingverysmalloperationswithmodest
resources,atlowincrementalcost.Theunknownsregardingthedarkfibermarketalsoraisethe
questionofwhetherthemodelwillensuresufficientcompetitionemergetosupporttheCitys
affordabilityanddigitalequitygoals.
1.3.6 Financial Analysis and Modeling of Single Dark Fiber P3

TounderstandthepotentialConnectionFeenecessarytosupporttheSingleDarkFiberP3Model,
the Project Team developed a version of the financial model for dark fiber only. The model
suggeststhatthenecessaryConnectionFee,assumingnodarkfiberrevenues,willbeintherange
of$15to$37perresidentialpassingpermonth,withabasecasevalueof$28.
TheProjectTeamnextmodeledarangeofscenariostounderstandhowdarkfiberrevenuesmay
reducetheConnectionFeenecessary.Giventhatadarkfibermarketofthistype(connecting
everyhomeandbusiness)hasneverbeenattemptedinacityofthesizeofSanFrancisco,we
analyzeddatafromtwootherprojectsinsmallercitieswherethismodelhasbeenpioneeredto
understandthefinancialimplicationsforSanFranciscooftherevenuesgeneratedinthosecities.
The City ofWestminster, Maryland, for example, has a 20year contract with Ting Internet in
which the city is obligated to finance and build all of the outside plant (including drops to
customerspremises).Tingisleasingfiberwithatwotieredleasepayment.Onemonthlyfeeis
basedonthenumberofpremisesthefiberpasses($6permonthperpassing);thesecondfeeis
basedonthenumberofsubscribersTingenrolls($17permonthpersubscriber).24Inanother
23
Arangeofpotentialfundingsourcesisanalyzedintheappendicestothisreport.
24
BecauseTingwillpayWestminsteramonthlyfeeof$6foreveryhomeandbusinesspassed,Tingisfinancially
obligatedtothecityfromdayone,evenifithasnocustomers.ThisstructuregivesthecityconfidencethatTingis
highlyincentedtosellservicestocovercosts.Tingwillalsopaythecity$17permonthforeachcustomeritserves.
Inlateryears,whenTingsrevenuehitscertainthresholds,Tingwillpaythecityasmallfractionofitsrevenueper
user.Thatmechanismisdesignedtoallowthecitytoshareinsomeoftheupsideofthenetworkssuccess.Inother
words, the city will receive a bit of entrepreneurial reward based on the entrepreneurial risk the city is taking.
Perhapsmostsignificantly,thereisalsoamechanismbuiltintothecontractthatensuresthatthetwopartiesare
13
variationonthismodel,Huntsville,Alabamasmunicipalelectricutilityisbuildingafibernetwork
andGoogleFiberisleasingfiberthroughoutthecitytoserveresidencesandsmallbusinesses.25
GoogleFibersleaseisfor20yearsbasedonaratesheetthatprovidesforvariouslevelsofpricing
basedonamountsandvolume.ThepricingstructureiscomplexbutweestimateGoogleFibers
monthlycosttobeapproximately$6to$7permonthperpassing.26
The financial model suggests that, applying the Westminster and Huntsville pricing to San
Francisco at a take rate of 50 percent for residential customers and 30 percent for business
customerswouldbringtheexpectedConnectionFeedownfrom$28permonthwithoutdark
fiberrevenuesto$16(Westminsterpricing)to$22(Huntsvillepricing)permonth.Ourbasecase
foradditionalrevenuesassumesamonthlyperpassingfeeof$6andamonthlypersubscriber
fee of $6 at a take rate of 50 percent for residential customers and 30 percent for business
customers,resultinginaresidentialConnectionFeeintheSingleDarkFiberP3Modelof$20per
month.
1.3.7 Upfront securing revenues

Thekeyuncertaintyinthefinancialanalysisofthisinitiativeistherevenuepotential.Whereas
theanalysisofpassingfeeandcustomerfeestructuresprovidearoughindication,therevenue
potentialoftheinitiativeisinherentlyuncertain.Thisiswhyapreleaseofaccesstothenetwork
inadvanceoftheprocurementandconstructionisimportant.Suchanapproachwouldnotonly
reducetheConnectionFeethatisrequiredtomaketheinitiativefinanciallyfeasible,butwould
alsocreatemorerobustcashflowandtherebyenhancetheabilitytofinancetheinitiative.Similar
preleaseorpresaleagreementsarecommoninothersectors,suchasrealestateandpower.
The Project Team prepared a simple sensitivity analysis that shows that an annual revenue
streamof$25millionwouldreducetherequiredConnectionFeefrom$51to$44permonth
undertheDualP3Modelandfrom$28to$22permonthundertheSingleDarkFiberP3Model.
(Notethatthisisanillustrationonly;the$25millionfigureisnotbasedonanyresearchandis
highlyuncertain.)Theseresultsfurtheremphasizethesignificanceoftheupfrontsecuritization
ofrevenues.
trulysharingriskaroundthefinancingoftheoutsideplantinfrastructure.InanyquarterinwhichTingsfinancial
obligationstothecityareinsufficienttomeetthecitysdebtservice,Tingwillpaythecity50percentoftheshortfall.
Insubsequentquarters,ifTingsfeestothecityexceedthedebtservicerequirements,Tingwillbereimbursedan
equivalent amount. This element of the financial relationship made the deal much more attractive to the city
becauseitisacleardemonstrationofthefactthatitsprivatepartnerisinvestedwithit.
25
Huntsvilleisnotfinancingorbuildingthedropstothehome;itsobligationtoGoogleistoprovidefiberthat
passesthepremisesonly.
26
Incontrast,TingsobligationstoWestminsterarebasedinpartonhowmuchfiberitusesandinpartonhow
manycustomersitsecuresandrevenuesitgenerates.Asaresult,Westminsterwillhavelesspredictabilityand
certaintyaboutitsrevenuesfromTing,buthasthepotentialtoshareinupsideintheeventthatTingisvery
successfulinthatmarket.
14
1.3.8 Innovative, LowerCost Financing Structure

BothP3modelsassumethattheDarkFiberP3willbefinancedinaninnovative,efficientwayto
reducethetotalfinancingcoststotheCity.WeanticipatethattheCitycanlargelyconvertits
darkfiberdebtfromP3capital(whichiscustomarilyquiteexpensiverelativetopublicfinance)
to revenue bonds. The revenue bonds would be secured with the Connection Fee revenues,
whicharelikelytobeasufficientlyrobuststreamofrevenuestosupportalowercostrevenue
bond.WerecommendthattherevenuebondnotentirelyreplaceP3capital,butratherthatthe
City include some P3 capital to ensure the concessionaires commitment and continuous
incentivetoperform.Inthisway,theCitymaybeabletosignificantlyreduceitsfinancecosts,
whilestillenjoyingthebenefitsofP3financingrelatedtoeffectiverisktransferandperformance
incentives. Any longterm dark fiber revenues that the City is able to secure in advance can
furthersupportthisfinancingstructure.27
1.3.9 Subsidy of Connection Fee for Lower Income San Franciscans

AP3thatinvolvesassessmentofafeeortaxtoresidencesandbusinesseswouldimposesome
burden on lower income San Franciscans. In consultation with City staff, CTC analyzed the
potential P3 structures with consideration of subsidizing the Connection Fee for some San
Franciscohouseholds.Underthebasecaseassumptionsasubsidywouldbeofferedto15percent
ofhouseholds.
1.4 Procurement Strategy and Recommendations

BothofthealternativebroadbandP3deliverymodelsdiscussedaboverequireaprocurement
strategythatisgroundedinthetraditionalP3processbutcustomizedandexpandedtoaccount
forboththeuniqueaspectsofthiseffortandtheP3investmentcommunityslackoffamiliarity
withthebroadbandsector.28Andtheuniquenatureofbroadbandasanassetclassalsorequires
addingnewfeaturestotheprocurementprocess,includingsecuringofftakeagreementsorpre
leaseofaccesstothenetworkinadvanceoftheprocurement.Werecommendthefollowing
elementsforaP3procurementprocess:
1. Industrysounding.Apreprocurementindustrysoundingwillclarifyandmarkettheproject
andproposedP3structuregeneratingmarketappetiteandsolicitinginputstofurtherrefine
thestructure.Theindustrysoundingcouldincludeindustrydayinteractionsoroneonone
interactionswithinterestedparties.
2. Preprocurementsaleofdarkfiberleases.Apreprocurementsalefordarkfiberandother
network capacity in advance of the P3 RFP will quantify and secure revenues that would
otherwise be speculative; this will enable (1) bids with lower requirements for City
27
ArecommendationtosecuredarkfiberrevenuesinadvanceofnetworkconstructionisdescribedinSection9.1.
28
DetaileddiscussionofaprocurementandfiberleasingstrategyisinSection9below.
15
funding/guarantees, and (2) a more competitive bidding process in which all bidders can
accountfortheserevenuestreams.Thissaleofcapacityalsohasthepotentialtoconcretize
theopportunity,forinvestors,serviceproviders,andthepublic.
3. TwoP3procurements,iftheDualP3structureisadopted.SeparateP3procurements(one
fordarkfiberandotheroutsideplantinfrastructure,anotherfortheequipment,operations,
and services) should be run partially in parallel to enable adjustments as necessary and
efficientlyproceed.
4. IntegrationofRFPdevelopmentandindustryconsultation/negotiations.AlthoughtheCity
hasbeencontemplatinganRFPdevelopmentprocessfollowedbynegotiationswithoneor
more selected bidders, we recommend integrating these phases and building multistep
consultationwithselectedbiddersintoamoreextensiveprocessofRFPdevelopmentand
vetting.ThisprocesswhichshouldbemoreextensivethaninotherP3sectors,giventhe
noveltyofthiseffortwillenablethenecessarylevelofinteractionwithpotentialbidders,
givingtheCityopportunitytorefinethesecomplexprocurements,whileenablingthebidders
tofurtherunderstandandcraftbidsforthisnew,unfamiliarassetclass.
Withinthisframework,wealsorecommendthattheprocurementprocessshould:
Beginasearlyaspossibleinordertogatherindustryfeedback.Akeyriskforthisprojectis
thattherewillbeamismatchbetweentheleveloffundingpreapproved,andwhatbidders
willproposeduringtheprocurementprocess.Inanidealsituation,bidderswouldbeselected
inadvanceoffinalfundingapprovals,enablingafirmunderstandingoftheleveloffinancial
commitment needed. Starting engagement with potential bidders through an industry
sounding,andinitiatingtheRFQandRFPprocessasearlyaspossibleinordertoobtainsome
levelofinteractionandinformationfrombidders,willhelpprovideasmuchinformationasis
feasibleatthatearlystageoftheprocess.
Build in sufficient time and flexibility. A procurement timeline that accounts for the
pioneeringandsingularnatureofthiseffortwillbuildinsufficienttimeandflexibilitytovet
all elements of the bid requirements. In the event the City elects to pursue the dual P3
structure,anextendedtimelinewilllikelyalsoberequired.
ApplyP3procurementbestpractices.Eventhoughthisprocurementwillbethefirstofits
kind,severalP3procurementbestpracticescanbeapplied.Theseincludeshortlistingthree
or four qualified bidders in the RFQ phase, and communicating with bidders prior to bid
submissioninthedraftRFPphase.
16
2 FTTP Design and Cost Estimate

CTCsengineerspreparedanetworkdesignandcostestimatesfordeployingagigabitfiberto
thepremises (FTTP) network to all homes and businesses in the City. We developed cost
estimates for two FTTP deployment approaches to provide data relevant to assessing the
financialviabilityofnetworkdeployment,andtodevelopingabusinessmodelfortheBroadband
forSanFranciscoInitiative:
AlitFTTPnetwork,includingthecoststodeployFTTPoutsideplantinfrastructure,as
well as all electronics, consumer drops, and customer premises equipment (CPE); this
estimateshowsthetotalcapitalcosts(bytheCityortheCityandpartners)tobuildan
FTTPnetworktosupportaubiquitous1Gbps(gigabitpersecond)dataservice
AdarkFTTPnetwork,includingthecosttodeployjusttheFTTPoutsideplant;thisis
thetotalcapitalcostfortheCitytobuildadarkFTTPnetworkforusebyoneormore
entities,includingprivateinternetserviceproviders
ThenetworkdesignandcostestimatesarebasedondataandinsightgatheredbyCTCengineers
indiscussionswithCitystakeholdersandanextensivefieldanddesksurveyofcandidatefiber
routes.
2.1 Summary of Cost Estimates

Thecostestimatesarebasedonacombinationofaerialandundergroundconstruction,anda
range of potential construction scenarios. The lowcost scenario assumes directional boring
throughout the City using test pitting to minimize outages to existing utilities. The highcost
scenariorequireshighlevelsofsurfacerestoration(i.e.,completesidewalkreplacement).
2.1.1 Lit FTTP Cost Estimate

WeestimatethelitFTTPnetworkdeploymentwillcost$1.9billion,29inclusiveofoutsideplant
construction labor, materials, engineering, permitting, pole attachment licensing, network
electronics,dropinstallation,CPE,andtesting.30Weestimatetheaverageconstructioncostper
passing31is$2,050.ThesecostsaresummarizedinTable6.32
29
30
BecausetheCitydirectedthattheanalysisassumethenetworkwillbedeployedinautilitymodel,the
estimatedtotalcostbreakdownassumesthat100percentofresidentsandbusinesseswillsubscribetotheservice.
Inadeliverymodelinwhichresidentsdecidewhetherornottosubscribetoservice,thismarketpenetrationrate
ortakeratewouldbelessthan100percent,andsomenetworkdeploymentcostswouldbelessasaresult.
31
Thepassingcountincludesindividualsingleunitbuildingsandunitsinmultidwellingandmultibusiness
buildingsassinglepassings.InlargerMDUs,thecostestimateassumesthatexistingwiringwillbeusedtoserve
theunits.WeassumethateachunitinlargeMDUswillcost$500.
32
17
Table6:EstimatedLitFTTPCost33

Installations
Actualcostsmayvaryduetofactorsthatcannotbepreciselyknownuntilthedetaileddesignis
completed, or until construction commences. These factors include: 1) costs of private
easements,2)utilitypolereplacementandmakereadycosts,3)variationsinlaborandmaterial
costs, 4) surface restoration, and 5) the Citys operational and business model. We have
incorporatedsuitableassumptionstoaddressthesefactorsbasedonourexperiencesinsimilar
markets.
The outside plant construction landscape in the City has changed since we last estimated
constructioncostsforSanFranciscoFTTPin2009.Sincethenmoreprovidershaveenteredthe
market,whichmeansmoreattachersontheutilitypolesandmoreconduitintheground.New
providersmeanspacethatwasoriginallyavailablefortheCitytoinstallfiberisnolongerthere,
whilecostsformakereadyandundergroundconduitinstallationhaveincreased.Further,new
surfacerestorationstandardswillincreasethecostofundergroundconstructionandtheamount
of time needed to construct the network. Utility pole replacement costs will increase, too,
becausetallerpoleswillbeneededinmanylocationstosupporttheexistingattachersandnew
Cityfiber.Andexistingpolescontinuetoageincreasingthenumberofreplacementsneeded.
Allofthesechangesincreasethecostofoutsideplantconstruction.
2.1.2 Dark FTTP Cost Estimate

WeestimatethatadarkFTTPnetworkdeploymentwillcost$1.5billion,34inclusiveofoutside
plant construction labor, materials, engineering, permitting, pole attachment licensing, and
lateral drop and materials. This estimate does not include any electronics or subscriber
equipment.CostcomponentsaredescribedinTable7.
33
Weapplyanuncertaintyof+/20percentbasedonourexperienceandtypicalindustrypractices.
34
18
Table7:EstimatedDarkFTTPCost35

Installations
The dark fiber estimate assumes that the City or its partner constructs and owns the FTTP
infrastructure up to a demarcation point (a network interface device) at each residence and
business,andthatthedarkfiberbackbone,distribution,anddropfiberarethenlitbyaseparate
entity.Thelitfiberentitywouldberesponsibleforallnetworkelectronicsandcustomerpremises
equipment(CPE)aswellasnetworksales,marketing,andoperations.Thenetworkarchitecture
couldsupportmultipleentitiesusingfiberassetsonaleasedbasisorotherbasis.
Figure1:DemarcationBetweenCityandPartnerNetworkElementsinLitandDarkModels
2.2 Approach to Technical Analysis

CTC engineers performed a range of analytical tasks to gather data and insight for the FTTP
design.
2.2.1 Field Survey

A CTC outside plant engineer performed a preliminary survey of San Francisco onsite and via
Google Earth Street View to develop estimates of permile cost for aerial fiber in the
communicationsspace,andpermilecostsforundergroundfiber(wherepolesarenotavailable).
35
19
The engineer reviewed available green space, necessary makeready on poles, and pole
replacementallofwhichhavebeenfactoredintothedesignandcostestimate.36
Table8summarizestheconditionsdeterminedthroughourfieldanddesksurvey,withrespect
tothreepopulationdensitiesweusedinthecostestimationmodelhigh,medium,andlow.
MakereadyandaerialconstructioncostswillbehigherintheCitythanothercitiesbutwillstill
belessexpensivethanundergroundconstruction.
Table8:FieldSurveyFindingsandMakeReadyAssumptions
Medium
HighDensity LowDensity
Density
AerialConstruction 20% 50% 60%
PolesperMile 50 50 50
MovesperPole 4 4 3
PolesRequiring
60% 60% 60%
MakeReady
PolesRequiring
10% 10% 10%
Replacement
CostPerMove $450 $450 $450
CostPerPole
$35,000 $35,000 $35,000
Replacement
CTCsoutsideplantengineernotedthatthequalityofthepolesandpoleattachmentsinSan
Francisco varied, as they do in many citiesbut overall, many poles would not support an
additionalcommunicationsattachmentwithoutsignificantmakeready.Whilethepolelinesmay
not be ideal for aerial construction, the cost of makeready is still less than the cost of
undergroundconstructionintheCity.Assuch,werecommendaerialconstructionwherepole
linesexist.
The following figures illustrate existing City utility conditions observed by CTCs outside plant
engineer.
36
Greenspaceisrightofwaytypicallycoveredingrassorsidewalkthatwouldbelessexpensivetoconstructin
thanthestreet.Makereadyiftheworkneededtobringautilitypoleintocomplianceandclearspaceforan
additionalattachment.Polereplacementisthereplacementofautilitypoletypicallywithonethatistallerand/or
ahigherclassthatwillsupportadditionalattachments.
20
Figure 2 shows a typical utility pole line along a major roadway in Bayview. The poles are
crowded,whichwillnecessitatesignificantmakeready.
Figure2:CrowdedUtilityPoleLineinBayview
21
Figure3showsanothercrowdedpolelineinExcelsior.
Figure3:UtilityPoleRequiringMakeReadyinExcelsior
22
Figure4showsapolelineinRichmondthatwillrequireextensivetreetrimming.
Figure4:TreeTrimmingRequiredonRichmondPoleLine

Figure5showsacleanpolelinewithrelativelylittlerequiredmakereadyinSilverTerrace.
Figure5:CleanPoleLineinSilverTerrace
23
2.2.2 Subsurface Investigations

CTCengineersmetwiththeCityandvariousserviceproviderstodiscusstheirexperienceswith
undergroundconstruction.TheCitysdensityandagemakeitextremelydifficultandexpensive
toconstructundergroundfacilities,especiallyinolderneighborhoods.
2.2.2.1 Directional Boring

Directionalboringinvolvesthediggingofsmalltestpits,generallyatlocationswherepullboxes
willbeinstalled,anddrillinghundredsoffeetbetweenthesepits.Theboringequipmentthen
pullstheconduitbackthroughtheborepathasitisremoved.
To minimize damaging underground utilities, the City is developing new regulations that will
requirecompaniesthatusedirectionalboringtotakeextraprecautionspriortoconstruction.
ThenewrequirementsincludereviewingCityprovidedutilitymaps,performingpreinspection
oftheexistingutilities,andtestpittingthecrossingsofallknownandassumedutilities.37These
impendingrequirementsareinresponsetoarecentundergroundingprojectinwhichCityutilities
werehitbyacompanythathaduseddirectionalboringastheconstructionmethod.
Evenwiththenewregulations,however,thereisstillriskinvolvedwiththistypeofconstruction.
The City has identified cases where undocumented sewer service lines have been damaged
becausetheyenterhomesandbusinessesatshallowerdepthsthantheCityscurrentstandard.
2.2.2.2 Microtrenching
TheCityiscontemplatingamicrotrenchingpolicy.Microtrenchingcutsanarrowtrench(i.e.,one
to two inches wide) into roadway or sidewalk pavement. The contractor then installs narrow
conduits,intowhichspecializedfiberisblown.Becausemicrotrenchingminimizestheamountof
surface restoration required, it can be a costeffective alternative to traditional underground
construction. This approach is widely used in New York City, Boston, and Vancouver, British
Columbia,especiallybywirelessserviceproviders.
However,becausemicrotrenchingrequiresatrench12to18inches38deepintothesidewalk,a
contractorcouldstillhitexistingservicelinestohomesandbusinesses.Also,undertheCitys
proposedrules,microtrenchingisonlyallowedinthesidewalk,andrequirestotalreplacement
ofanysidewalkslabsthatarecut.Theserequirementswillincreasethecostofmicrotrenching
construction,comparedtothemorecommonpracticeofonlyrequiringthecuttobefilled,or
comparedtomicrotrenchingalongthecurb.
37
Ourlowcostestimateassumesthatdirectionalboringisusedtoreducecosts,withtestpittingperformedevery
25feetonaverage.
38
Thedepthofthemicrotrenchingisbasedonthesawblade.Deeperdepthssuchas18inchesaremore
expensivethanshallowerdepths.
24
Inaddition,inSanFrancisco,therehavebeeninstancesofsidewalkreplacementcausingcutsto
existing telecommunications providers service drops. This has occurred in places where the
dropswereinstalledatextremelyshallowdepthsbecauseotherutilitieswerelocateddeeper,
belowthetelecommunicationsdrops.Weanticipatethatamicrotrenchingapproachthatfollows
theCitysnewruleswouldresultinconstructioncostsbetweenourlowandhighestimates.
2.2.2.3 Street Restoration

AnyblockofastreetinSanFranciscothatistrenchedmorethan25feet(longitudinally)requires
therepavingoftheentiretrenchedlaneofthatblock.Thisrequirement,whileprotectingthe
roadways,will also increase thecost of underground construction. It is partly because of this
requirementthatmanyoftheCitysundergroundareaswillrequireconstructiononbothsides
oftheroad,insteadofthemorecommontechniqueofbuildingfiberononesideoftherightof
way,thenboringunderthestreettoreachtheresidencesandbusinessesontheotherside.
2.2.2.4 Sidewalk Restoration

IftheCityoradarkfiberpartnerweretobuildtheFTTPnetworkinadherencetotheCityscurrent
sidewalkrestorationstandards,whichrequirereplacementofeachdisturbedconcreteslab,the
projectwouldresultinthefullrebuildofmostoftheCityssidewalksinundergroundconstruction
areas.Thislargescalerebuildshouldbeconsideredoneofthemainoutcomesoftheproject,if
not the primary one. The sheer amount of underground construction required to build the
network,combinedwiththesidewalkrestorationrequirements,wouldresultinanamountof
sidewalkreplacementequivalenttothatofalargescalecapitalimprovementproject.
TheCitycouldviewthesidewalkreplacementcostsasaseparatecapitalimprovementproject
implementedconcurrentlywiththefiberbuild.Atanestimatedcostof$150perconcreteslab,
sidewalk restoration would account for approximately $100 million of the estimated outside
plantanddropcosts.Additionalcapitalimprovementmoneycouldbeusedatthesametimeto
furtherimprovethestreetscapebyaddingrampsatcorners,repairingstreetcurbs,installingtree
boxes,replacingstreetsigns,andsoon.
2.3 FTTP Network Design

Thearchitectureofthephysicalplantdeterminesthenetworksscalabilityforfutureusesand
howtheplantwillbeoperatedandmaintained;thearchitectureisalsothemaindeterminantof
thetotalcostofthedeployment,becauseoutsideplantisboththemostexpensivepartofthe
networkandthelongestlasting.Figure6(below)showsalogicalrepresentationofthehighlevel
FTTPnetworkarchitecturewerecommendinSanFrancisco.Thisdesignisopentoavarietyof
architecture options. The drawing illustrates the primary functional components in the FTTP
network,theirrelativepositiontooneanother,andtheflexibilityofthearchitecturetosupport
multiplesubscribermodelsandclassesofservice.
25
Figure6:HighLevelFTTPArchitecture
26
Therecommendedarchitectureisahierarchicaldatanetworkthatprovidescriticalscalabilityand
flexibility, both in terms of initial network deployment and its ability to accommodate the
increased demands of future applications and technologies. The characteristics of this
hierarchicalFTTPdatanetworkare:
Capacityabilitytoprovideefficienttransportforsubscriberdata,evenatpeaklevels
Availabilityhighlevelsofredundancy,reliability,andresiliency;abilitytoquicklydetect
faultsandreroutetraffic
Failsafeoperationphysicalpathdiversitytominimizeoperationalimpactresultingfrom
fiberorequipmentfailure
Efficiencynotrafficbottlenecks;efficientuseofresources
Scalabilityabilitytogrowintermsofphysicalserviceareaandincreaseddatacapacity,
andtointegratenewertechnologies
Manageabilitysimplifiedprovisioningandmanagementofsubscribersandservices
Flexibilityabilitytoprovidedifferentlevelsandclassesofservicetodifferentcustomer
environments; can support an openaccess network or a singleprovider network; can
provideseparationbetweenserviceprovidersonthephysicallayer(separatefibers)or
logicallayer(separateVLANorVPNprovidingnetworkswithinthenetwork)
Securitycontrolledphysicalaccesstoallequipmentandfacilities,plusnetworkaccess
controltodevices
Thisarchitectureoffersthescalabilitynecessarytomeetlongtermneeds.Itisconsistentwith
bestpracticesforanopenaccessnetworkmodelthatmightpotentiallyberequiredtosupport
multiple network operators, or at least multiple Retail Service Providers requiring dedicated
connectionstocertaincustomers.ThisdesignwouldsupportacombinationofGigabitPassive
OpticalNetwork(GPON)anddirectActiveEthernetservices(withtheadditionofelectronicsat
thefiberdistributioncabinets),whichwouldenablethenetworktoscalebymigratingtodirect
connectionstoeachcustomer,orreducingsplitterratios,onanasneededbasis.
Thedesignassumesplacementofmanufacturerterminatedfibertapenclosureswithintheright
ofwayoreasements,providingwatertightfiberconnectorsforcustomerservicedropcablesand
eliminating the need for service installers to perform splices in the field. This is an industry
standardapproachtoreducingbothcustomeractivationtimesandthepotentialfordamageto
distributioncablesandsplices.Themodelalsoassumestheterminationofstandardlateralfiber
connectionswithinlargermultitenantbusinesslocationsandmultidwellingunits(MDU).
27
2.3.1 Network Design

ThenetworkdesignandcostestimatesassumetheCityoritsdarkfiberpartnerwill:
Identifyandprocurespaceattwocorefacilitiestohousenetworkelectronicsand
providebackhaultotheinternet
UtilizeexistingCitylandtolocate10distributionhubfacilitieswithadequate
environmentalandbackuppowersystemstohousenetworkelectronics
Constructfiberopticstoconnectcoresitestodistributionhubsinabackbonering
Constructadditionalbackbonefibertoconnectthedistributionhubstofiber
distributioncabinets(FDC)
ConstructfiberopticsfromtheFDCstoeachresidenceandbusiness(i.e.,from
terminationpanelsintheFDCtotaplocationsintherightofwayoronCityeasements)
Constructfiberlateralsintolarge,multitenantbusinessfacilitiesandMDUs
LeveragingtheCitysconduitandfiberresourcescoulddecreasethecostsassociatedwithboth
constructingabackboneandidentifyinglocationstohouseelectronicsthatareneartheCitys
existingresources.GiventhehighcostofconstructioninmanyareasoftheCity,usingexisting
conduitmayprovideanoticeableimpactonthecostofconstruction.However,foraprojectof
thismagnitude,thecostsavingsarelikelywithintherangeofourprojectedcontingency.
TheFTTPnetworkandserviceareas,depictedinFigure7(below),weredefinedbasedonthe
followingcriteria:
Designtargets512passingsperFDC
Serviceareasweredefinedbypassingdensityandexistingutilitiesandarebrokeninto
thecategoriesofhigh,medium,andlowdensities
MultipleFDCsserveeachservicearea
FDCsaresuitabletosupporthardenednetworkelectronics,providingbackuppower
andanactiveheatexchange39
Distributionplantavoidscrossingmajorroadways,bridges,andrailways
39
ThesehardenedFDCsreflectanassumptionthattheCitysoperationalandbusinessmodelwillrequirethe
installationofproviderelectronicsintheFDCsthatarecapableofsupportingopenaccessamongmultiple
providers.WenotethattheoverallFTTPcostestimatewoulddecreaseifthehardenedFDCswerereplacedwith
passivefiberdistributioncabinets(whichwouldhouseonlyopticalsplitters)andtheproviderselectronicswere
housedonlyathublocations.IncludescostsassociatedwiththeSurfaceMountedFacilitiesOrdinance.
28
Coupled with an appropriate network electronics configuration, this design serves to greatly
increase the reliability of fiber services provided to customers compared to that of more
traditionalcableandtelephonenetworks.Thebackbonedesignminimizestheaveragelengthof
nondiverse distribution plant between the network electronics and each customer, thereby
reducingtheprobabilityofserviceoutagescausedbyafiberbreak.
Figure7:FTTPServiceAreas
29
Theaccesslayerofthenetwork,encompassingthefiberplantfromtheFDCstothecustomers,
dedicatesasinglefiberstrandfromtheFDCtoeachpassing(potentialcustomeraddress).This
traditional FTTP design allows either network electronics or optical splitters in the FDCs. See
Figure8belowforasampledesign.
Figure8:SampleFTTPAccessLayerDesign
30
2.3.2 Network Core and Hub Sites

ThenetworkcoresiteslinktheFTTPnetworktothepublicinternetanddeliverallservicestoend
users. The proposed network design includes two core locations, based on the networks
projectedcapacityrequirementsandtheneedforgeographicalredundancy(i.e.,ifonecoresite
weretofail,thesecondcoresitewouldcontinuetooperatethenetwork).
Thelocationofcorenetworkfacilitiesalsoprovidesphysicalpathdiversityforsubscribersandall
upstreamserviceandcontentproviders.Forthedesignandcostestimates,weassumethatthe
Cityscoresiteswillbehousedinsecurelocationswithdiverseconnectivitytotheinternetand
theCitysexistingfiberopticnetwork.
ThecorelocationsinthisplanwillhouseprovidersOperationalSupportSystems(OSS)suchas
provisioningplatforms,faultandperformancemanagementsystems,remoteaccess,andother
operationalsupportsystemsforFTTPoperations.Thecorelocationsarealsowhereanybusiness
partnerorcontent/serviceproviderswillgainaccesstothesubscribernetworkwiththeirown
pointofpresence.Thismaybeviaremoteconnection,butcollocationisrecommended.
ThecorelocationsaretypicallyruninaHighAvailability(HA)configuration,withfullymeshed
andredundantuplinkstothepublicinternetand/orallothercontentandserviceproviders.Itis
imperativethatcorenetworklocationsarephysicallysecureandallowunencumberedaccess
24x7x365toauthorizedengineeringandoperationalstaff.
Theoperationalenvironmentofthenetworkcoreandhublocationsissimilartothatofadata
center.Thisincludescleanpowersources,UPSbatteries,anddieselpowergenerationforsurvival
through sustained commercial outages. The facility must provide strong physical security,
limited/controlled access, and environmental controls for humidity and temperature. Fire
suppressionishighlyrecommended.
Equipmentistobemountedsecurelyinracksandcabinets,incompliancewithnational,state,
andlocalcodes.Equipmentpowerrequirementsandspecificationmayinclude48voltDCand/or
120/240voltsAC.Allequipmentistobeconnectedtoconditioned/protectedcleanpowerwith
uninterruptedcutovertobatteryandgenerator.
Forthecostestimate,weassumedthatthecorefacilitieswouldbelocatedinleaseddatacenters
anddistributionhubswillbelocatedatexistingCityfacilitieswithlandforatelecommunications
shelter.Thesquarefootageneededdependsonthebusinessmodel,howeverweestimated100
squarefeetneededpercoresiteand200squarefeetneededperhubsitetobeabletosupport
multipleserviceoperators.Anetworksupportingmultipleserviceprovidersrequiresmorespace
than a single service provider, but less space at the core and hubs may limit future network
scalability.
31
2.3.3 Distribution and Access Network Design

Thedistributionnetworkisthelayerbetweenthehubsandthefiberdistributioncabinets(FDCs,
whichprovidetheaccesslinkstothetaps).Thedistributionnetworkaggregatestrafficfromthe
FDCstothecore.Fibercutsandequipmentfailureshaveprogressivelygreateroperationalimpact
astheyhappenclosertothenetworkcore,soitiscriticaltobuildinredundanciesandphysical
pathdiversitiesinthedistributionnetwork,andtoseamlesslyreroutetrafficwhennecessary.
Thedistributionandaccessnetworkdesignproposedinthisreportisflexibleandscalableenough
tosupporttwodifferentarchitectures:
1. Housingboththedistributionandaccessnetworkelectronicsatthehubs,andusingonly
passivedevices(opticalsplittersandpatches)attheFDCs,or
2. Housingthedistributionnetworkelectronicsatthehubsandpushingtheaccessnetwork
electronicsfurtherintothenetworkbyhousingthemattheFDCs
Byhousingallelectronicsatthehubs,thenetworkwillnotrequirepowerattheFDCs.Choosing
a network design that only supports this architecture may reduce costs by allowing smaller,
passiveFDCsinthefield.However,thisarchitecturewilllimittheredundancycapabilityfromthe
FDCstothehubs.
Bypushingthenetworkelectronicsfurtherintothefield,thenetworkgainsaddedredundancy
byallowingtheaccesselectronicstoconnecttothehubsites.Intheeventonehubhasanoutage,
thesubscribersconnectedtotheFDCwouldstillhavenetworkaccess.Choosinganetworkdesign
thatonlysupportsthisarchitecturemayreducecostsbyreducingthesizeofthehubs.
Selecting a design that supports both models would allow the City to accommodate many
differentserviceoperators,theirnetworkdesigns,andvariousnetworktechnologies.Thisdesign
wouldalsoallowserviceproviderstostartwithasmalldeployment(i.e.,placingelectronicsonly
atthehubsites)andtogrowbypushingelectronicsclosertotheirsubscribers.
2.3.3.1 Access Network Technologies

FDCscansitonacurb,bemountedonapole,orresideinabuilding.Thismodelrecommends
installingsufficientFDCstosupporthigherthananticipatedlevelsofsubscriberpenetration.This
approach will accommodate future subscriber growth with minimal reengineering. Passive
optical splitters are modular and can be added to an existing FDC as required to support
subscribergrowth,ortoaccommodateunanticipatedchangestothefiberdistributionnetwork
withpotentialfuturetechnologies.
TheFTTPdesignalsoincludestheplacementofindoorFDCsandsplitterstosupportMDUs.This
wouldrequireobtainingtherighttoaccesstheequipmentforrepairsandinstallationinwhatever
32
timeframe is required by the service agreements with the customers. Lack of access would
potentially limit the ability to perform repairs after normal business hours, which could be
problematicforbothcommercialandresidentialservices.
Inthismodel,weassumetheuseofGPONelectronicsforthemajorityofsubscribersandActive
Ethernet for a small percentage of subscribers (typically business customers) that request a
premiumserviceorrequiregreaterbandwidth.GPONisthemostcommonlyprovisionedFTTP
serviceused, for example, by AT&T Fiber, Verizon (in its FiOS systems), Google Fiber, and
Sonic.net.
Furthermore,providersofgigabitservicestypicallyprovidetheseservicesonGPONplatforms.
EventhoughtheGPONplatformislimitedto1.2Gbpsupstreamand2.4Gbpsdownstreamfor
thesubscribersconnectedtoasinglePON,operatorshavefoundthatthevariationsinactual
subscriberusagegenerallymeansthatallsubscriberscan obtain1Gbpsondemand(without
provisioned ratelimiting), even if the capacity is aggregated at the PON. Furthermore, many
GPONmanufacturersaredevelopingtechnologytosupportupto10Gbpsandfasterspeedsas
userdemandincreases.
GPON supports highspeed broadband data, and is easily leveraged by tripleplay carriers for
voice,video,anddataservices.TheGPONOLTusessinglefiber(bidirectional)SFPmodulesto
supportmultiple(mostcommonlylessthan32)subscribers.
GPONusespassiveopticalsplitting,whichisperformedinsidefiberdistributioncabinets(FDC),
toconnectfiberfromtheOLTstothecustomerpremises.WithGPONservice,theFDCshouse
multiple optical splitters, each of which splits the fiber link to the OLT between 16 to 32
customers.
ActiveEthernet(AE)providesasymmetrical(up/down)servicethatiscommonlyreferredtoas
Symmetrical Gigabit Ethernet. AE can be provisioned to run at subgigabit speeds, and like
GPON, easily supports legacy voice, voice over IP (VoIP), and video content. AE is typically
deployedforcustomerswhorequirespecificservicelevelagreementsthatareeasiertomanage
andmaintainonadedicatedservice.
For subscribers receiving Active Ethernet service, a single dedicated fiber goes directly to the
subscriberpremiseswithnosplitting.BecauseAErequiresdedicatedfiber(homerunfiber)40
fromtheOLTtotheCPE,andbecauseeachsubscriberusesadedicatedSFPontheOLT,thereis
asignificantcostdifferenceinprovisioninganAEsubscriberversusaGPONsubscriber.
40
Homerunfiberisafiberopticarchitecturewhereindividualfiberstrandsareextendedfromthedistribution
sitestothepremises.Homerunfiberdoesnotuseanyintermediaryaggregationpointsinthefield.
33
ThefiberplantisdesignedtoprovideActiveEthernetserviceorPONservicetoallpassings.The
networkoperatorselectselectronicsbasedonthemixofservicesitplanstoofferandcanmodify
orupgradeelectronicstochangethemixofservices.
2.3.3.2 Expanding the Access Network Bandwidth

GPONiscurrentlythemostcommonlyprovisionedFTTPtechnology,duetoinherenteconomies
whencomparedwithtechnologiesdeliveredoverhomerunfibersuchasActiveEthernet.The
costdifferencebetweenconstructinganentirenetworkusingGPONandActiveEthernetis40
percentto50percent.41GPONisusedtoprovideservicesupto1Gbpspersubscriberandispart
ofanevolutionpathtohigherspeedtechnologiesthatusehigherspeedopticsandwavedivision
multiplexing.
This model provides many options for scaling capacity, which can be done separately or in
parallel:
1. ReducingthenumberofpremisesinaPONsegmentbymodifyingthesplitterassignment
and adding optics. For example, by reducing the split from 16:1 to 4:1, the peruser
capacityintheaccessportionofthenetworkisquadrupled.
2. AddinghigherspeedPONprotocolscanbeaccomplishedbyaddingelectronicsattheFDC
or hub locations. Since these use different frequencies than the GPON electronics, no
otherCPEwouldneedtobereplaced.
3. AddingWDMPONelectronicsastheybecomewidelyavailable.Thiswillenableeachuser
tohavethesamecapacityasanentirePON.Again,theseusedifferentfrequenciesthan
GPONandarenotexpectedtorequirereplacementoflegacyCPEequipment.
4. Option 1 could be taken to the maximum, and PON replaced by a 1:1 connection to
electronicsanActiveEthernetconfiguration.
These upgrades would all require complementary upgrades in the backbone and distribution
Ethernet electronics, as well as in the upstream internet connections and peeringbut they
wouldnotrequireincreasedfiberconstruction.
2.3.3.3 Customer Premises Equipment (CPE) and Subscriber Services

In the final segment of the FTTP network, fiber runs from the FDC to customers homes,
apartments,andofficebuildings,whereitterminatesatthesubscribertapafiberoptichousing
locatedintherightofwayclosesttothepremises.Theserviceinstallerusesapreconnectorized
dropcabletoconnectthetaptothesubscriberpremiseswithouttheneedforfiberopticsplicing.
41
EnhancedCommunicationsinSanFrancisco:PhaseIIFeasibilityStudy,CTCreport,October2009,atp.205.
34
The drop cable extends from the subscriber tap (either on the pole or underground) to the
building,entersthebuilding,andconnectstocustomerpremisesequipment(CPE).
2.4 FTTP Cost Estimate

ThissectionprovidesasummaryofcostestimatesforconstructionofalitFTTPnetworktoall
Cityresidencesandbusinesses.Thisdeploymentwillcostanestimated$1.9billion,42inclusiveof
outsideplantconstructionlabor,materials,engineering,permitting,poleattachmentlicensing,
networkelectronics,dropinstallation,customerpremisesequipment(CPE),andtesting.
Table9summarizesthesecosts.43
Table9:EstimatedLitFTTPCost

Installations
ToillustratethepotentialcostiftheCityweretochooseamodelotherthantheutilitymodel
(i.e.,ifresidentsweretochoosewhetherornottobuyservices),Figure9showsthetotallow
estimatedcostwithrespecttopotentialtakerates.Theprojectedcostisroughlylinearbytake
ratebecausethecostofaddingadditionalsubscribersisafixedcost.
42
43
35
Figure9:TotalEstimatedCostversusTakeRate(LowCostEstimate)

2.4.1 Outside Plant Cost Estimation Methodology
Aswithanyutility,thedesignandassociatedcostsforconstructionvarywiththeuniquephysical
layoutoftheserviceareanotwostreetsarelikelytohavetheexactsameconfigurationoffiber
opticcables,communicationsconduit,undergroundvaults,andutilitypoleattachments.Further,
soilconditions,suchastheprevalenceofsubsurfacehardrock;theconditionofutilitypolesand
feasibilityofaerialconstructioninvolvingtheattachmentoffiberinfrastructuretoutilitypoles;
andcrossingsofbridges,railways,andhighwaysalsoaffectcosts.
Inourestimates,weextrapolatedthecostsforstrategicallyselectedsampledesignsbasedon
streetmileageandpassings.Specifically,wedevelopedsampleFTTPdesignstogeneratecosts
perpassingforthreetypesofpopulationdensitieshighmediumandlow. TheCitywasdivided
intoseparateareasaccordingtoCensustracks.Thedensityofeachareawasdeterminedbythe
totalnumberofidentifiedpassingsperthesquarefootageandstreetmileageofthegivencensus
track.Thenumberofpassingsforeachcensustrackwasderivedfromaddressdataprovidedby
theCity.
Our observations found it most prudent to estimate a range of potential costs for FTTP
construction.Ourlowcostmodelassumesdirectionalboringwithtestpittingevery25feetand
36
minimal surface restoration. 44 The highcost model assumes that sidewalk replacement is
requiredforsurfacerestorationandthatconduitwouldbeinstalledinundergroundareasaspart
ofthesidewalkreplacement.
Theassumptions,sampledesigns,andcostestimateswereusedtoextrapolateacostperpassing
fortheoutsideplant.Thisnumberwasthenmultipliedbythenumberofpassingsineacharea
basedontheCitysGISdata.TheactualcosttoconstructFTTPtoeverypremisesintheCitycould
differfromtheestimateduetochangesintheassumptionsunderlyingthemodel.
Forexample,ifaccesstotheutilitypolesisnotgrantedormakereadyandpolereplacement
costs are too high, more of the network would have to be constructed undergroundwhich
couldsignificantlyincreasethecostofconstruction.Alternatively,iftheCitycouldpartnerwith
alocaltelecommunicationsproviderandoverlashtoexistingpoleattachments,thecostofthe
build could be significantly lower. Further and more extensive analysis would be required to
developamoreaccuratecostestimateacrosstheentireCity.
2.4.2 Components of Construction Cost Estimate

Thecostcomponentsforoutsideplantconstructionincludethefollowingtasks:
Engineeringincludessystemlevelarchitectureplanning,preliminarydesignsandfield
walkouts to determine candidate fiber routing; development of detailed engineering
printsandpreparationofpermitapplications;andpostconstructionasbuiltrevisions
toengineeringdesignmaterials.
QualityControl/QualityAssuranceincludesexpertqualityassurancefieldreviewof
finalconstructionforacceptance.
General Outside Plant Construction consists of all labor and materials related to
typicalundergroundoraerialoutsideplantconstruction,includingconduitplacement,
utility pole makeready construction, aerial strand installation, fiber installation, and
surface restoration; includes all work area protection and traffic control measures
inherenttoallroadwayconstructionactivities.
Special Crossings consists of specialized engineering, permitting, and incremental

construction(materialandlabor)costsassociatedwithcrossingsofrailroads,bridges,and
interstate/controlledaccesshighways.
44
Bycomparison,thecosttomicrotrenchwouldlikelybebetweenthelowandhighcostsestimatedependingon
thesurfacerestorationrequired.
37
BackboneandDistributionPlantSplicingincludesalllaborrelatedtofibersplicingof
outdoorfiberopticcables.
BackboneHub,Termination,andTestingconsistsofthematerialandlaborcostsof
placinghubsheltersandenclosures,terminatingbackbonefibercableswithinthehubs,
andtestingbackbonecables.
FTTPServiceDropandLateralInstallationsconsistsofallcostsrelatedtofiberservice
drop installation, including outside plant construction on private property, building
penetration, and inside plant construction to a typical backbone network service
demarcation point. This cost also includes all materials and labor related to the
termination of fiber cables at the demarcation point.We assumed a take rate of 100
percentforstandardfiberservicedrops,becausetheutilitymodelmeansthatevery
residenceandbusinesswillbeserved.
2.4.3 Outside Plant Construction Costs

Intermsofoutsideplant,theestimatedcosttoconstructtheproposedFTTPnetworkis$760
million,or$2,050perpassing.45Asdiscussedabove,themodelassumesamixtureofaerialand
undergroundfiberconstruction,dependingontheconstructionofexistingutilitiesintheareaas
wellasthestateofanyutilitypolesandexistinginfrastructure.Table10andTable11providea
breakdownoftheestimatedoutsideplantcostsbytypeofarea.(Notethatthecostshavebeen
rounded.)
Table10:EstimatedOutsidePlantCosts(DirectionalBoringCostEstimate)
Distribution
Costper CostPer
Area Plant TotalCost Passings
Passing PlantMile
Mileage
EntireCity 1,600 $703,292,000 374,000 $1,880 $440,000
HighDensity 20 $10,875,000 26,000 $420 $544,000
Medium
400 $200,439,000 140,000 $1,430 $501,000
Density
LowDensity 1,180 $491,978,000 208,000 $2,370 $417,000
45
ThisistheHighCostScenario.Thepassingcountincludesindividualsingleunitbuildingsandunitsinmulti
dwellingandmultibusinessbuildingsassinglepassings.InlargerMDUs,thecostestimateassumesthatexisting
wiringwillbeusedtoservetheunits.Weassumethateachunitwillcost$500perunitinlargeMDUs.
38
Table11:EstimatedOutsidePlantCosts(SidewalkRestorationCostEstimate)
Distribution
Costper CostPer
Area Plant TotalCost Passings
Passing PlantMile
Mileage
EntireCity 1600.0 $762,858,000 373,480 $2,040 $480,000
HighDensity 20.0 $11,686,000 25,960 $450 $644,000
Medium
400.0 $211,365,000 139,940 $1,510 $531,000
Density
LowDensity 1,180.0 $539,807,000 207,580 $2,600 $459,000
Costsforaerialandundergroundplacementwereestimatedusingavailableunitcostdatafor
materialsandestimatedlaborcostsforplacing,pulling,andboringfiberbasedonconstruction
incomparablemarkets.
Barring unknown economies of scale and inflation rates, as well as any sort of phenomenon
restrictingmaterialavailabilityandcosts,materialcostsweregenerallystraightforward.
We estimated labor costs associated with the construction of fiber based on similar FTTP
constructionprojects,informationfromserviceprovidersintheCity,andthecoststheCityhas
paidforotherfiberconstruction.
Aerialconstructionentailstheattachmentoffiberinfrastructuretoexistingutilitypoles,which
could offer significant savings compared to allunderground construction, but increases
uncertaintyaroundcostandtimeline.Theutilitypoleownersandotherattachedproviderscan
imposecostsanddelaysrelatedtopoleremediationandmakereadyconstructionthatcan
makeaerialconstructioncostprohibitiveincomparisontoundergroundconstruction.
While generally allowing for greater control over timelines and more predictable costs,
undergroundconstructionissubjecttouncertaintyrelatedtocongestionofutilitiesinthepublic
rightsofway and the prevalence of subsurface hard rockneither of which can be fully
mitigated without physical excavation and/or testing. While anomalies and unique challenges
willariseregardlessofthedesignorconstructionmethodology,thelargescaleofthisprojectis
likely to provide ample opportunity for variations in construction difficulty to yield relatively
predictableresultsonaverage.
Weassumeundergroundconstructionwillconsistprimarilyofhorizontal,directionalboringto
minimizerightofwayimpactandtoprovidegreaterflexibilitytonavigatearoundotherutilities.
Thedesignmodelassumesasingletwoinch,HighDensityPolyethylene(HDPE)flexibleconduit
39
overundergrounddistributionpaths,anddualtwoinchconduitsoverundergroundbackbone
pathstoprovidescalabilityforfuturenetworkgrowth.
2.4.4 Central Network Electronics Costs

Centralnetworkelectronicswillcostanestimated$112million,or$300perpassing,basedonan
assumedtakerateof100percent.46(Thesecostsmaybephasedinassubscribersareaddedto
thenetwork.)Thecentralnetworkelectronicsconsistoftheelectronicstoconnectsubscribers
totheFTTPnetworkatthecore,hubs,andcabinets.Table12belowliststheestimatedcostsfor
eachsegment.
Table12:EstimatedCentralNetworkElectronicsCosts
NetworkSegment Passings CostperPassing Subtotal

CoreandDistributionElectronics 373,000 $100 $37million
FTTPAccessCabinetElectronics 373,000 $200 $75million
CentralNetworkElectronicsTotal 373,000 $300 $112million
2.4.4.1 Core Electronics

The core electronics connect the hub sites and connect the network to the internet. Core
electronicsconsistofhighperformancerouters,whichhandlealltheroutingbothontheFTTP
network and to the internet. The core routers should have modular chassis to provide high
availability in terms of redundant components and the ability to hot swap 47line cards and
modularintheeventofanoutage.Modularroutersalsoprovidetheabilitytoexpandtherouters
asdemandforadditionalbandwidthincreases.
Thecostestimatedesignenvisionsredundantringsbetweenthecoresitesrunningnetworking
protocolssuchashotstandbyroutingprotocol(HSRP)toensureredundancyintheeventofa
corefailure.Additionalringscanbeaddedasbandwidthonthenetworkincreases.Thecoresites
wouldalsotietobothhubsusing10Gbpslinks.Thelinkstothehubscanalsobeincreasedwith
additional10Gbpsand40Gbpslinecardsandopticsasdemandgrowsonthenetwork.Thecore
networkswillalsohave40GbpstointernetserviceprovidersthatconnecttheFTTPnetworkto
theinternet.
Thecostofthecoreroutingequipmentforthetwocoresitesis$8million.Thesecostsdonot
includetheserviceprovidersOperationalSupportSystems(OSS)suchasprovisioningplatforms,
46
Thetakerateaffectstheelectronicsanddropcosts,butalsomayaffectotherpartsofthenetwork,astheCity
maymakedifferentdesignchoicesbasedontheexpectedtakerate.InCTCsfinancialanalysis,wewillexamine
howthefeasibilityoftheprojectdependsonarangeoftakerates.
47
Ahotswappablelinecardcanberemovedandreinsertedwithouttheentiredevicebeingpowereddownor
rebooted.Thecontrolcardsintheroutershouldmaintainallconfigurationsandpushthemtoareplacedlinecard
withouttheneedforreconfirmation.
40
fault and performance management systems, remote access, and other operational support
systemsforFTTPoperations.Serviceprovidersand/ortheircontentprovidersmayalreadyhave
thesesystemsinplace.
2.4.4.2 Distribution Electronics

ThedistributionnetworkelectronicsatthetwohubsitesaggregatethetrafficfromtheFDCsand
send it to the core sites to access the internet. The core sites consist of highperformance
aggregationswitches,whichconsolidatethetrafficfromthemanyaccesselectronicsandsendit
to the core for route processing. The distribution switches typically are large modular switch
chassis that can accommodate many line cards for aggregation. The switches should also be
modulartoprovideredundancyinthesamemannerasthecoreswitches.
This cost estimate assumes that the aggregation switches connect to the access network
electronicswith10Gbpslinkstoeachdistributionswitch.Theaggregationswitcheswouldthen
connecttothecoreswitchesoversingleormultiple10Gbpslinksasneededtomeetthedemand
oftheFTTPusersineachservicearea.
Thecostofthedistributionswitchingequipmentfordistributionhubsis$29million.Thesecosts
donotincludeanyoftheserviceprovidersOSSorothermanagementequipment.
2.4.4.3 Access Cabinet Electronics

TheaccessnetworkelectronicsattheFDCsconnectthesubscribersCPEstotheFTTPnetwork.
WerecommenddeployingaccessnetworkelectronicsthatcansupportbothGPONandActive
Ethernet subscribers to provide flexibility within the FDC service area. We also recommend
deployingmodularaccessnetworkelectronicsforreliabilityandtheabilitytoaddlinecardsas
moresubscribersjoinintheservicearea.Modularityalsohelpsreduceinitialcapitalcostswhile
thenetworkisunderconstructionorduringtherolloutofthenetwork.
Thecostoftheaccessnetworkelectronicsforthenetworkis$74million.Thesecostsinclude
opticalsplittersattheFDCsandarebasedonatakerateof100percentundertheutilitydelivery
model.
2.4.5 Subscriber Electronics and Service Drop Installation Costs (PerSubscriber

Costs)
Eachactivatedsubscriberwouldalsorequireafiberdropinstallationandsubscriberelectronics,
whichwouldcostroughly$2,600to$2,800persubscriber,or$970millionto$1.0billiontotal.
ThesecostsaresummarizedinTable13.
41
Table13:PerSubscriberActivationCostEstimates
ConstructionandElectronicsRequiredto
EstimatedAverageCost
ActivateaSubscriber
DropInstallationandMaterials $1,800to$2,000
SubscriberElectronics(ONTandOLT) 500
ElectronicsInstallation 200
InstallationMaterials 100
Total $2,600to$2,800
Thenumbersprovidedaboveareaveragesandwillvarydependingonthetypeofpremisesand
theinternalwiringavailableateachpremises.
2.4.5.1 Customer Premises Equipment and Related Subscriber Electronics

Thepersubscriberexpensesforelectronicsarethecostoftheopticalnetworkterminal(ONT)at
thepremises,aportionoftheopticallinetermination(OLT)costsatthehub,thelabortoinstall
andconfiguretheelectronics,andtheincidentalmaterialsneededtoperformtheinstallation.
TheONT,alsoknownascustomerpremisesequipment(CPE),isthesubscribersinterfacetothe
FTTPnetwork.Forthiscostestimate,weselectedCPEsthatprovideonlyEthernetdataservices
(however, there are a wide variety of CPEs offering other data, voice, and video services).
Assuming a 100 percent take rate under the utility delivery model, weestimated the CPE for
customerswillbe$290million.ForlargeMDUsweassumetheuseofinternalwiring.Thecost
estimateassumestheuseofanEthernetswitchoraG.FastorG.hnnetworkaccessmultiplexer
andnetworkterminal.
2.4.5.2 Drop Installation

Thedropinstallationcostisthebiggestvariableinthetotalcostofaddingasubscriber.Ashort
aerialdropcancostaslittleas$250toinstall,whereasalongundergrounddropinstallationina
congestedundergroundarealikeSanFranciscocancost upwardof $5,000. Inthismodel,we
estimateanaverageof$1,800to$2,000perdropinstallation,foratotalof$680millionto$740
million.
Toestimatethecostofthedrops,wedeterminedtheaveragedroplengthforeachsubscriber.
The average drop length is calculated using a variety of variable and assumptions including
averagehousesetback,averagehomefrontage,andnearesttapplacement.Lessdenseareas
havelongerdropshowevertheytendtohavemoreaerialconstructionwhichislessexpensive.48
48
TheFTTPfiberdistributionplantisapproximatelyhalfaerialandhalfunderground,thereforehalfofthedrops
willbeaerialandhalfthedropswillbeunderground.Table8providesthebreakdownofaerialplantbetween
densityareas.
42
Weassumedacostofapproximately$400fortheinstallationofanaerialdroporpullingadrop
throughexistingconduit.Inundergroundareas,weassumedacostof$50to$65perfootfor
undergroundconduitplacement.MDUslargerthan20unitsreceiveasinglelateraltoservethe
entirebuilding.Thefollowingtablessummarizethedropcosts(andhavebeenrounded).
Table14:EstimatedDropandLateralCosts(DirectionalBoringEstimate)
Average
TotalDrop TotalDrop DropCost
Area Passings Drop
Costs Footages PerPassing
Length
EntireCity $680,000,000 373,000 23,000,000 133 $1,800
HighDensity $11,000,000 26,000 300,000 98 $700

Medium
$200,000,000 140,000 8,800,000 124 $1,700
Density
LowDensity $490,000,000 207,000 13,900,000 142 $2,100
Table15:EstimatedDropandLateralCosts(SidewalkReplacementEstimate)
Average
TotalDrop TotalDrop DropCost
Area Passings Drop
Costs Footages PerPassing
Length
EntireCity $741,000,000 373,000 23,000,000 125 $2,000
HighDensity $19,000,000 26,000 300,000 98 $700
Medium
$260,000,000 140,000 8,800,000 124 $1,900
Density
LowDensity $462,000,000 207,000 13,900,000 142 $2,200
The cost of underground construction drives the average drop. For example, an installation
requiring 30 feet of underground construction costs $1,900 ($400 for the drop at $1,500 for
conduitinstallationat$50perfoot).Whereas,inasuburbancommunitywhereadropcouldbe
plowedthroughagrassyyardthecostwouldonlybe$700assumingaconduitinstallationcost
of$10perfoot.
43
3 The Purely Public Broadband Option Involves High Risk

InevaluatingtheoptionsbeforetheCity,theProjectTeamconsideredapurelypublicmodel,in
whichtheCitywouldbuild,own,andoperateafibertothepremisesnetworkitself.Thisisa
highriskandhighrewardproposition,witharespectabletrackrecordincommunitiesacrossthe
country.Inourobservation,however,thisisadifficultmodeltoreplicate,particularlyforacity
thatdoesnotownitselectricutilityoronethatdoesnothavetheinternalexpertisetobuildand
operateaninternetenterprise.Aggressiveantimunicipaleffortsbyincumbentphoneandcable
companiesmakethismodelevenmorerisky.Thepurelypublicmodelhasneverbeenundertaken
byacityofthesizeorprofileofSanFrancisco.
AndtheeconomicsofthemodelarelikelytobechallenginginSanFrancisco.Assumingthelower
costestimatedevelopedbytheengineers,theCitywouldrequirefinancingofalmost$1.29billion
inbondsanda$20millionloan.Tobreakeven,theCitywouldrequirea47percenttakeratein
both the business and residential markets. If implementation costs approach the high cost
estimate,developingaprojectthatisfinanciallyfeasiblebecomesevenmorechallenging.The
totalbondinginthefirstfouryearsincreasestoroughly$1.43billion,plustheinitialloanof$20
million.Thetakeraterequiredtobreakevenincreasesto53percent.
3.1 Background
Over the past 25 years, about 100 local communities have built hybrid fibercoaxial (HFC)
networks(thearchitectureusedbycablecompanies)orFTTPnetworkstocomprehensivelyserve
residentialandbusinessmarkets.Inalmosteverycase,thesenetworkshavebeendeployedin
smalltomidsizetownsinlargelyruralareas.Some,butnotall,ofthesetownsalreadyhadcable
modemservice,butmanyofthemwereunservedorclosetounservedbybroadbandservice.
3.1.1 Track Record of Public Broadband Efforts in FTTP

TheeconomicsofFTTPareextremelychallenginggiventheveryhighcapitalcostsandthemodest
revenuespossible,particularlyinlightofcompetitionfromlesserbroadbandtechnologies.Unlike
otherpublicutilitiessuchaswaterandsewer,citycommunicationsnetworksdonotoperatein
amonopolyenvironment,andanumberofcompetitortechnologies,howeverinferior,doexist.
These include farlowerbandwidth options such as DSL, cable modem service, and wireless
service.(Incontrast,someofthemunicipalFTTPnetworkswerebuiltadecadeormoreagoata
time when there may not have been much or any competition in those rural towns.) These
alternativetechnologiesdonotofferthesamefutureproofscalabilityandspeedsasfiber,but
cancertainlyofferrobustcompetitionwithrespecttopriceandotherfactors.
44
TremendoussuccesseshavebeenachievedbypublicFTTPnetworkssuchinLafayette,Louisiana;
Chattanooga,Tennessee;andWilson,NorthCarolinaallofwhicharemunicipalelectricutilities
thatachievedsubstantialefficiencies.49
Themostdramaticsuccessesofthesenetworksareinthebenefitsthatdonotnecessarilyshow
uponfinancialstatementssuchasenhancedproductivity,innovation,education,healthcare,
companyrecruitment,andrelatedbenefitsthatareamongthereasonsforthecommunications
investmentinthefirstplace.
AmodestsubsetofthemunicipalFTTPandHFCnetworkshavestruggledfinanciallyforarange
of reasons, including the challenges created by incumbent opposition to municipal networks,
whichinsomestateshasrestrictedmunicipaloptions(forexample,inUtah,municipalitiesare
statutorilyprohibitedfromofferingretailservicesalawthatwaspassedatthebehestofthe
incumbent providers, and that has created significant economic challenges for municipal
networks)butalsoongoinglitigation,antimunicipalPRcampaigns,andrelatedeffortstodelay,
obstruct,andincreasethecostsofmunicipalefforts.
The significant challenging economics of competitive broadband have also played a factor in
creatingrealfinancialchallengesforasubsetofthemunicipalnetworks.Insomecommunities,
someofthefunctionsrequiredtooperateacompetitiveinternet,cable,andphoneenterprise
havesimplyprovenchallengingforpublicsectorentitiesinparticularsuchcriticalelementsas
superiorcustomerserviceinacompetitivemarketischallengingevenforprivatecompanies,and
hasprovensoforsomepublicentitiesaswell.
3.1.2 Track Record of Public Broadband in California

Ahandfulofmunicipalnetworksthathavestumbledhaveemergedandbeenwidelypublicized
by antimunicipal incumbents and their allies. In California, there exist a number of very
successful municipal efforts, particularly those with more modest or incremental goals. For
example, the city of Santa Monica built fiber optics beginning in the late 1990s to connect
governmentbuildingsandhasbeensuccessfulinexpandingitsnetworkandofferingservicesin
thecommercialmarket,thoughithasnottodateenteredtheresidentialmarket.Similarly,the
cityofPaloAltosmunicipalelectricutilityhasseentremendoussuccessinleasingtotheprivate
sectordarkfiberstrandsonitsbackbonenetwork.
49
ChristopherMitchell,BroadbandAttheSpeedofLight:HowThreeCommunitiesBuiltNextGeneration
Networks,InstituteforLocalSelfReliance,April2012,http://www.ilsr.org/wpcontent/uploads/2012/04/muni
bbspeedlight.pdf(accessedMarch2017).Seealso:JulieKendrick,TheBroadbandChallenge,TheLine,July18,
2012,http://www.thelinemedia.com/features/broadband071812.aspx(accessedMarch2017).
45
Ubiquitous broadband to the home has proven more challenging. The city of Alameda, for
example,builtahybridfibercoaxialcableandinternetnetworkthatitlatersoldatasubstantial
losstoComcastfollowinganumberofyearsofchallengingfinancialoutcomes.
3.1.3 Advantages of Having a Public Electric Utility

Themajorityofthepublicnetworksthathavebeendeployedweredeployedbymunicipalelectric
utilities.Thiscorrelationisnotsurprisingforanumberofreasons.First,communitiesinwhich
theprivatesectordidnothaveabusinesscaseforelectrificationarewherelocalgovernments
chose to build public power. Not surprisingly, those same communities did not see adequate
privatesectorinvestmentinbroadband,either,andthuschose,inbothcases,toinvestforthe
benefitofthebroadercommunity.
Second,thechallengeofundertakingapublicfacingcommunicationsprojectisreducedfora
municipalelectricutilityrelativetoalocalgovernmentthatisnotalreadyapowerprovider.A
rangeofelementsofacommunicationsnetworkoverlapthoseofapowernetwork,includingthe
polesonwhichtheinfrastructureisbuilt,thefacilitiesinwhichnetworkhubsarelocated,the
skillsandequipmentoffieldstaff,andeven,insomecases,thebilling,operating,andcustomer
service systems that support the service offerings. A minority of the municipal, publicfacing
networkswerebuiltbylocalitiesthatwerenotpowerutilities.
3.2 Benefits
Whileitcomesatconsiderablecost,municipalbroadbandoffersdemonstrableandimportant
benefits,includingmunicipaldecisionmakingandcontrolwithregardtocriticalfactorssuchas
insuringthatthenetworkismadeuniversallyaccessibletoallmembersofthecommunityrather
thanbuildingonlytoareaswithahigherreturnoninvestment.
Municipal ownership also delivers considerable flexibility and opportunity to expand, change
businessmodels,andotherwiseinnovateasnecessarytomeettheneedsofthecommunity.
Anadditionalandcriticalareaofbenefitisinregardtopricingandtheaffordabilitythatiscreated
byacompetitivedynamic.ThemunicipalFTTPandHFCnetworkstodatehavetendedtobein
marketswherethelikelihoodofathirdprivatecompetitoremergingwasslimtononexistent.In
fact, some of these networks have developed in markets where only one or no broadband
competitorsexisted.Asaresult,thesemunicipalnetworkshavealmostuniversallydelivereda
levelofcompetitionthatwasinconceivableotherwise.
3.3 Challenges
At the same time, municipal networks present considerable challenges. In addition to the
financial risk entailed in building and operating a venture that competes with private sector
service offerings in some cases, a number of other challenges present themselves. Municipal
46
networksaregenerallyonlypossiblewhenacityisinapositiontoissuebondstoraisecapitalto
buildthenetwork.Inmanycases,evencitieswithoutstandingcreditratingsareunabletobond
to support a communications network because of other critical priorities, such as shoring up
pension funds, building schools and public safety facilities, and the many other functions of
governmentforwhichbondingcapacitymustbereserved.Andbondingcapacityislimitedinany
city,nomatterhowgooditscreditratingorhowprosperousitseconomy.
Competingneedsforcapitalarethusaconsiderablechallengearoundmunicipalefforts.Atthe
same time, additional challenges can be created in multiple functional areas associated with
construction,operations,andmaintenanceofacommunicationsenterprise.Thechallengescan
rangefromtheconsiderabledifficultyofhiringnetworkengineers,particularlyintheBayAreaof
Californiawherenetworkengineersareinsuchdemandandcommandveryhighprivatesector
salariesthataredifficultforthepublicsectortomatch.
3.4 Financial Analysis

Tofullyunderstandthepotentialviabilityofthepurelypublicmodelforthisinitiative,theProject
Teamdevelopedafullfinancialmodelandanalysisoftheeconomicsofthetraditionalmunicipal
ownershipandoperationsmodel.(SeeAppendixJ.)
Underthismodel,andassumingthelowercostestimatedevelopedbytheengineers,theCity
wouldrequirefinancingofalmost$1.29billioninbondsanda$20millionloan.Tobreakeven,
theCitywouldneedalmosthalfofallresidencesandbusinessestobuyservicesaveryrobust
takerate.
ThebasecaseanalysisassumesthattheCityoffersthreedataservices:
1Gbpsresidentialserviceat$80permonth
1Gbpssmallcommercialserviceat$90permonth
1Gbpsmediumcommercialservice(withservicelevelagreements)at$240permonth50
TheseservicepricesareabovethoseofGoogleFiberinothermarkets($70for1Gbps)butare
lowerthansomeotherproviderspricing.Forexample,Comcastsrecentlyannouncedresidential
1 Gbps service in San Francisco is priced at $160 per month, with introductory promotional
pricingat$70permonth,plusanequipmentrentalfeeof$10permonth.
Themodelassumesthatsubscribershipfordataserviceswillrampupoveryearsonethrough
four,andthenremainsteady.Themodelassumesthatsubscribershipforvoiceandvideoservices
50
Mediumcommercialservicereceivesaloweroversubscriptionrate,thatis,lesscustomerssharingthe
connection,decreasingtheinstancesofnetworkcongestionreducingoverallspeeds.
47
will ramp up over the same period, but then decline as adoption of overthetop (OTT)
alternativesincreases.
Ifthetakerateassumedinthisbasecaseanalysisweretodrop,oriftheCitywereunableto
maintainservicepricesatthelevelsassumedinouranalysis,theenterprisewouldaccumulate
significantdeficitsovertime.
As an example, our sensitivity analysis indicates that if the take rate were to drop by 10
percentagepoints,theenterprisewouldquicklyturnfromanunrestrictedcashbalanceofabout
$406,000 in year one to a negative cash balance. By year five, the unrestricted cash balance
deficitwouldbeover$61.9million,growingtoadeficitofmorethan$324millionbyyear20.
We also found that if competitive pressures were to force the City to reduce the residential
internetservicefeeby$5permonthfromthelevelsassumedinthebasecase,theenterprise
wouldnotbefinanciallyfeasible.Byyear10,theunrestrictedcashbalancedeficitwouldbemore
than$58.2million,growingtoadeficitofjustunder$126millionbytheendofyear20.
48
4 The Purely Private Option Will Not Meet the Citys Goals of Open,
Ubiquitous Fiber to All Homes and Businesses
With respect to the key goals identified for this initiativeopen access and direct fiber
connectivitytoeveryhomeandbusinessinSanFranciscoprivatedeploymentaloneisunlikely
tosuffice.Thoughthereareemergingpatternsofnewdeploymentbycompetitorsandupgrades
byincumbents,noneoftheprovidersinSanFranciscooffersanopenaccessnetworkandnone
ofthemplanstobuildfiberonaubiquitousbasis.
Thereexistslittleagreementaboutthedegreetowhichprivatecapitalisemergingtoexpand
broadband competition and improve broadband speeds. Nationally, while the incumbent
communicationscompaniesclaimtobeinvestingextensivelyinupgradedandnewbroadband
facilities,robustnewfiberinvestmenttendstobeconcentratedinmarketsthatalreadyfeature
strongcompetition(suchasthosewhereGoogleFiberdeployedextensiveresidentialfiber)or
more incremental in nature (for example, incremental expansion of the cable industrys fiber
backbonetosupportnewcablemodemtechnologies,orincrementalexpansionofthephone
companysfiberfootprinttoreachlargebusinesscustomers).
SanFranciscosresidentsandbusinesseshavearelativelybroadrangeofservicesavailable,as
comparedtootherurbanareas.However,ascomparedtoFTTPareas(suchasVerizonFiOSand
GoogleFibercitiesandtheFTTPnetworksinmajorcitiesinAsia),SanFranciscoisamongthe
Americancitieswithoutcitywideresidentialfiber.
InSanFrancisco,noneoftheCitysincumbentwiredproviders(AT&T,Wave,andComcast)has
indicatedanyplanstodeployfiberonaubiquitousbasis,althoughnewcompetitionisemerging
insmaller,selectareasthankstonewinvestmentbycompanieslikeSonic,Monkeybrains,and
Webpass.
VerizonoffersFiOSinpartsofSouthernCaliforniabuthasnotindicatedanyplansofextending
theservicetoSanFrancisco.GoogleFibersexpansionappearshalted,atleastfornow,though
GoogleAccessWebpassunitisexpandinginthemultidwellingunitmarketdowntown.
ComcasthasupgradeditsentireSanFranciscoservicefootprint51tothenextgenerationofcable
modemtechnology,whichenablesspeedsofupto1Gbpsinthedownstreamdirectionunder
optimal conditions. AT&T is also upgrading its infrastructure to fiber in many parts of San
Francisco,butitsunlikelythattheupgradewillbecitywideorubiquitous.AT&Tsupgradesin
many cities thus far have been incremental and concentrated in select neighborhoods,
51
ComcastinformedCTCandCitystaffinaSeptember2017meetingthatitsnetworkreaches98.9percentof
residencesand56percentofbusinessesinSanFrancisco.
49
sometimesbasedonwhereAT&Talreadyhasfiberandcancosteffectivelyextendit,oronareas
whereAT&Thasreasontobuildfibertoservelargeenterprisecustomers.
Insummary,whileComcastsupgradehasimprovedspeedsthroughouttheCity,andwhilethe
othercompaniesareallincrementallygrowingorupgradingtheirexistingfacilities,theprivate
sectorisnotcurrentlypoisedtodeliverubiquitousserviceorubiquitouscompetitionthroughout
SanFrancisco.Theredoesexistnewinvestment,whichmayincreasewithtime,buttheresults
will be uneven and inconsistent, with some but not all areas seeing substantially improved
services.
4.1 AT&T
AT&TistheprimaryincumbentlocalexchangecarrierinSanFranciscoandmuchofCalifornia,
whereitoffersDSLandothercopperlineservicestomostoftheCityandprovidesfiberservices
suchasMetroEthernettobusinesses.Becausethecostofthefiberservicesissubstantiallyhigher
thantheDSLservices,smallandmediumsizedbusinessesmayhavedifficultyaffordingthese
enhancedservices.
DSL represents a relatively lowbandwidth form of broadbanda network of roads, not

superhighways.DSLrunsontelephonenetworkcopperwires,whichsimplycannothandlethe
same capacity as fiber or even as Comcasts hybrid fibercoaxial (HFC) networks. As capacity
requirementsincrease,DSLislikelytofallfurtherbehindcable.52EvenwithnewerDSLversions,
suchastheVDSLfibertothecabinetserviceprovidingdownloadspeedsofapproximately50
Mbps,20itsspeedissignificantlylowerthatwhatisavailableovercableorfiber.Fromatechnical
standpoint, DSL is a shortterm solution in a market where bandwidth needs are growing
exponentiallyandhigh,symmetricalcapacityisincreasinglyneededforsmallbusinessesandfor
popularapplicationslikevideodownloads,videogaming,andvideoconferencing.Agingcopper
plantisnotcapableofmeetingthesegrowingneedsinthelongrun.
Inaverypositivedevelopment,AT&Thasrecentlystartedupgradingsomeareasinafewdozen
citiesthroughoutthecountrytofibertothepremises53andwehaveobservedAT&Tupgrading
infrastructureinsomepartsofSanFrancisco.Itsnotclearhowextensivethefiberupgradein
SanFranciscowillbe,andexperienceinothercitiessuggestsitisunlikelytobeubiquitous.
52
ThelimitationsofDSLareillustratedbyVerizonsinvestment,overthepastdecadeorso,tosupplementitsold
copperphonenetworkswithnewFTTPnetworksinlimitedmetropolitanareaswithinitsexistingfootprint,andby
AT&Tslimitedbuteffectiveexpansionoffiberinsomemarketswhereitfacescompetition.
53
100%FiberNetworkPoweredbyAT&TFiberExpandsin17MetrosSeeifItsNearYou,AT&T,March23,
2017,http://about.att.com/newsroom/att_fiber_expands_to_17_metro_areas.html;
http://www.theverge.com/2017/3/24/15056428/attcontinuestoexpanditsfiberinternetservice(accessed
June2017).
50
4.2 Comcast
In San Francisco, Comcast operates a hybrid fibercoaxial system that ably competes against
otherofferingsintheexistingmarketplace.Comcasthasalsoindicatedthatitwillupgradeits
networkinSanFranciscotothenextgenerationofcablemodemtechnology,knownasDOCSIS
3.1, within the next year or two. 54 That upgrade, if it is also supported by additional fiber
construction,willenableComcasttooffer(upto)gigabitspeedsinthedownstreamdirection
tomanySanFranciscoconsumers.Pricingforthisproductinothermarketshasthusfarexceeded
$150.
Comcastsnetwork,however,islimitedbyitslackoffiberevenwithadvancedelectronicsand
software,itssystemcannotkeeppacewiththepotentialspeedsoffullyfibernetworkssuchas
those operated by Verizon, Google Fiber and, in limited areas, AT&T and CenturyLink. Cable
systemsareboundbytheinherentlimitationsofthecoaxialcablethatrunsfromtheirnodesinto
the home. An additional limitation arises from the shared nature of cable modem service
bandwidthwithinaneighborhoodissharedratherthandedicated.Asaresult,promisedspeeds
ofupto1Gbpsdownstream(andconsiderablylessupstream)maybesignificantlydecreased
byonesneighborssimultaneoususeoftheircablemodems.
Importantly,theDOCSIS3.1upgradeprovidesthepotentialforhigherspeedstonearlyallSan
Franciscoresidents,becausetheupgradecoverstheentireservicearea,notonanincremental
or partial basis (for technical reasons, an electronics upgrade such as DOCIS 3.1 is done on a
systemwidebasis).
Incontrast,however,theDOCSIS3.1technologywillnotbeavailableforthoseinsmallbusiness
areas that are outside Comcasts existing footprint. Because the traditional footprint of cable
companies was residential only, for purposes of selling oneway cable video products, cable
networks frequently do not reach into business areas where there exists insufficient for new
infrastructureinvestment.Asaresult,somebusinessareasdonothavethebenefitofComcasts
networksandextensionoftheinfrastructuretothosebusinessescanbeprohibitivelyexpensive.
Comcast also offers Gigabit Pro, its fibertothepremises symmetrical 2 Gbps residential
service, in limited parts of San Francisco. Gigabit Pro is only available to customers within a
quarter to a third of a mile of Comcasts existing fiber backboneand the service require
extensionofthatfiberthecustomerspremises,whichcantakemanyweeks.Asoflate2016,
54
JonBrodkin,Comcastsgigabitcablewillbein15citiesbyearly2017,ArsTechnica,November2,2016,
https://arstechnica.com/informationtechnology/2016/11/comcastsgigabitcablewillbein15citiesbyearly
2017/(accessedJune2017).
51
Comcast was charging approximately $1,000 for installation and service activation. Monthly
servicefeesareapproximately$300.55
4.3 Etheric Networks

EthericNetworksisaprivatelyowned,fixedwirelessISP,basedinRedwoodCity,withcoverage
areas spanning from Salinas to Berkeley. Since 2003, the company has deployed licensed
microwaveantennastoconnectcustomerstoitsfiberbackbone.56Thecompanyalsoprovides
directfiberconnectionsforbusinessesinselectlocations.
Itsresidentialservicecomeswitha$400to$440startupfee,coveringthecostofnewantennas
and their installation. Then service costs $85 to $350 per month for guaranteed minimum,
symmetrical speeds ranging from 1 Mbps to 25 Mbps. They also have more expensive small
business/homeofficeservicewithlowerlatencyandgreaterburstpotentialavailableinsome
areas.ThisSOHOservicecosts$160to$700permonthwithguaranteedspeedsbetween3Mbps
and 25 Mbps. Etherics business service levels range from $200 to $1,500 per month, with
guaranteedspeedsrangingfrom2Mbpsto100Mbps.57
Ethericreportsthatitcontinuestodeployhigherspeedwirelesstechnologiesastheybecome
available,butthecompanyhassharednoplanstofurtherexpanditsfiberbackboneatthistime.
4.4 Monkeybrains
MonkeybrainsisaSanFranciscobasedISPthatstartedofferingdialupservicein1998.Itfocuses
exclusivelyonservingtheSanFranciscoarea,butdoesnotofferserviceinmuchofthewestern
region of the City. Over the years, it has continued to improve its service offerings as better
technologieshavebecomeavailabletoconnectcustomerstoitscorefibernetwork.In2009,it
beganusingmicrowavetechnologytoconnectendcustomerstoitscorefibernetwork.Itnow
hasanetworkofmorethan1,000networkaccesspointsacrosstheCity,whichsendsignalsalong
lineofsightconnectionstoantennasoncustomerpremises.
Monkeybrainsstandardresidentialservicecosts$35amonth,aftera$250antennasetupfee.
In2015,itusedacrowdfundingcampaigntonegotiateabulkbuyingdiscountonmillimeterwave
radios. Through the crowdfunding campaign, customers could cover the cost of the $2,500
antenna,andasareward,Monkeybrainspromisedtodeliver300to500Mbpstotheirhomeor
55
JonBrodkin,Comcastsgigabitcablewillbein15citiesbyearly2017,ArsTechnica,November2,2016,
https://arstechnica.com/informationtechnology/2016/11/comcastsgigabitcablewillbein15citiesbyearly
2017/(accessedJune2017).
56
RobertMullins,EthericBootstrapstobuildwirelessbroadbandnetwork,SiliconValleyBusinessJournal,
November30,2003,http://www.bizjournals.com/sanjose/stories/2003/12/01/story5.html/(accessedJune2017).
57
EthericNetworks,https://ethericnetworks.com/(accessedJune2017).
52
officeatthesamemonthlypriceof$35.Throughthecampaign,273backershelpedthecompany
raised$420,501,enoughtopurchasealmosttwiceasmanyantennasasplanned.Thecompany
used the extra funds from the project to deploy additional antennas on the roofs of schools,
homelessshelters,andculturalcentersacrosstheCity.58
While Monkeybrains standard service remains 20 Mbps/10 Mbps, it also offers business
packageswithsymmetricalspeedsrangingfrom20Mbpsto1+Gbps.59Theirservicelevelswill
likely improve as wireless technologies get better and more affordable, but the company
leadershipacknowledgesthatatsomepointtheywillhavetoinvestmoreheavilyintofiber.The
foundershavereportedlystatedthattheyhavenoplanstoinvestinlastmilefiberlinksuntil
thereissomesortofrelaxationintherightofwayaccess[offiber].60
FCC Form 477 data are not available for Monkeybrains, but our understanding of the
MonkeybrainsservicefootprintisthatitislimitedtocertainareasofsinglefamilyhomesinSan
Francisco.
4.5 Wave
WaveisaKirkland,Washington,basedISPwithanetworkthatincludesmetropolitanareason
theWestCoastfromtheBayAreatoSeattle.Theprivatelyheldcompanywasrecapitalizedfor
morethan$1billioninNovember2012,61andraised$130millionin201562andanother$125
millionin2016.63Itisusingthesefundstoexpanditsfiberandcoaxialfootprintthroughnew
constructionandacquisitions.
58
RickPaulus,CananIndieISPbeatComcast?PopularMechanics,March14,2016,
http://www.popularmechanics.com/technology/infrastructure/a19741/cananindieispbeatcomcast/(accessed
June2017).
59
Monkeybrains.net,https://www.monkeybrains.net/business.php(accessedJune2017).
60
RickPaulus,CananIndieISPbeatComcast?PopularMechanics,March14,2016,
http://www.popularmechanics.com/technology/infrastructure/a19741/cananindieispbeatcomcast/(accessed
June2017).
61
WaveDivisionHoldingsRaises$1.052Billion,PlanstoaccelerateExpansionofWestCoastBroadband/Fiber
OpticNetwork,PressRelease,WallerCapital,November15,2012,http://www.wallercapital.com/wavedivision
holdingsraises1052billionplanstoaccelerateexpansionofwestcoastbroadbandfiberopticnetwork/
(accessedJune2017).
62
MikeFarrell,WaveRaises$130MforFiberBuild,MultiChannelNews,May21,2015,
http://www.multichannel.com/news/broadband/waveraises130mfiberbuild/390795(accessedJune2017).
63
WaveRaises$125MilliontoContinueAcceleratedFiberNetworkConstructionandCommercialBroadband
ServicesExpansion,PressRelease,Wave,September14,2016,http://residential.wavebroadband.com/wave
raises125milliontocontinueacceleratedfibernetworkconstructionandcommercialbroadbandservices
expansion/(accessedJune2017).
53
Wave began offering gigabit residential internet service in San Francisco in June, 2015. It has
focuseditsfibereffortsonlargeMDUs,anditdoesnotsharethelocationofitsexistingfiber
plant.Whereavailable,theservicecosts$80permonth.
ConsumerswhodonotliveinabuildingwithWaveGgigabitservicemaystillbeabletoreceive
up to 110 Mbps service over Waves hybrid fiber coaxial (HFC) network. 64 The company also
offersbusinessspeedsupto1Gbps,butservicelevelsabove250Mbpsareonlyavailableinselect
locations,orwithaddednewconstructionfees.65
WhileitisunclearhowmuchnewfiberconstructionWavehasplannedfortheSanFrancisco
area,thecompanyrecentlyputoutajoblistingforafiberconstructionleadinSanFrancisco.66
ThecompanywilllikelycontinuetoexpandfibertoMDUs,buthasnotindicatedanyintentionto
expanditsFTTPservicetosinglefamilyhomes.
The following map (Figure 10) illustrates the reach of Waves services, and is based on self
reporteddataprovidedbyWavetotheFCC.67
64
SanFrancisco,CACableTV,Internet,andPhoneService,Wave,
http://residential.wavebroadband.com/resources/cities/sanfranciscocableinternetphone.php(accessedJune
2017).
65
HighSpeedInternet,Wave,http://residential.wavebroadband.com/internet/(accessedJune2017).
66
FiberConstructionLead:WaveBroadband,SanFranciscoBayArea,JobListing,LinkedIn,
https://www.linkedin.com/jobs/view/249977325(accessedJune2017).
67
ThemapisbasedonFCCForm477dataasofJune30,2016,https://www.fcc.gov/general/broadband
deploymentdatafccform477(accessedJune2017).
54
Figure10:WaveServiceinSanFrancisco
4.6 Webpass
WebpassisawirelessISP,foundedinSanFranciscoin2003.Thecompanyusesmillimeterwave
antennastoofferhighspeedservicetocustomersinMDUs.WhenAlphabetsaccessdivision,
GoogleAccess,purchasedthecompanyinJune,2016,Webpasshadinfrastructuretoservesome
residentialMDUsandofficebuildingsinSanFrancisco,Oakland,SanDiego,Miami,Chicago,and
Boston.68
68
AGreatDay,WebpassBlog,October3,2016,https://webpass.net/blog/agreatday(accessedJune2017).
55
Sincetheacquisition,GoogleAccesshasscaledbackitsFTTPefforts69andappearstobemoving
forwardwithWebpasswirelessstrategy.70WebpasshaslaunchednewserviceinDenver,and
appearstobepreparingtolaunchinSeattle.71
Webpassoffersitscustomersspeedsrangingfrom100Mbpsto1Gbpsfor $60 permonth.72

Webpasscontinuestoexpanditsservicestonewbuildingsinthemarketsitserves.Inorderto
beeligibleforWebpassservice,abuildingmusthaveatleast10unitsandbewiredwithEthernet
cabling.73ItappearsunlikelythatWebpasswillprovideservicetosinglefamilyhomesinthenear
futureduetothehighcostofthemillimeterwaveantennasthatthecompanyreliesonforits
lastmileconnections.74
FCCForm477dataarenotavailableforWebpass,butourunderstandingoftheWebpassservice
footprintisthatitisconcentratedinMDUsinthedowntownareaofSanFrancisco.
4.7 Sonic
SonicisaSantaRosabasedISPthathasbeeninbusinesssince1994.Historically,Sonicoffered
DSLserviceoverAT&Tscopperphonelines,butthecompanyhassteadilyexpandeditsnetwork
acrosstheBayArea,andbeganofferingfiberbasedservicetoresidentialcustomersinlimited
areas in 2011, with fiberbased business service soon to follow.75After building smaller FTTP
projectsinNorthandEastBay,SonicannouncedinFebruary2016theprovisionofGigabitfiber
69
JonBrodkin,GoogleFiberdivisioncutsstaffby9%,pausesfiberplansin11cities,ArsTechnica,October25,
2016,https://arstechnica.com/informationtechnology/2016/10/googlefiberlayingoff9ofstaffwillpause
plansfor10cities/(accessedJune2017).
70
JackClark,GooglePlotsCheaperWirelessFuturetoExpandFiberProject,BloombergTechnology,July28,
2016,https://www.bloomberg.com/news/articles/20160728/googleplotscheaperwirelessfuturetoexpand
fiberproject(accessedJune2017).
71
JonBrodkin,GoogleFibermakesexpansionplansfor$60wirelessgigabitservice,ArsTechnica,February22,
2017,https://arstechnica.com/informationtechnology/2017/02/googlefibermakesexpansionplansfor60
wirelessgigabitservice/(accessedJune2017).
72
JonBrodkin,GoogleFibermakesexpansionplansfor$60wirelessgigabitservice,ArsTechnica,February22,
2017,https://arstechnica.com/informationtechnology/2017/02/googlefibermakesexpansionplansfor60
wirelessgigabitservice/(accessedJune2017).
73
MorePlacesToGrabGigabit,WithWebpass,BlogPosting,Webpass,January30,2017,
https://webpass.net/blog/moreplacestograbgigabitwithwebpass(accessedJune2017).
74
BruceBrown,GoogleFiberIsntJustFiberAnymoreNowItOffersWirelessService,Too,DigitalTrends,
October5,2016,http://www.digitaltrends.com/web/googlefiberwebpasshybridinternetfiberwireless/
(accessedJune2017).
75
MartinEspinoza,Sonicincreasesinvestmentinfibernetworks,ThePressDemocrat,January18,2015,
http://www.pressdemocrat.com/news/3310907181/sonicincreasesinvestmentinfiber?artslide=0(accessed
June2017).
56
serviceavailableinpartsoftheSunsetandRichmondneighborhoods.76Overthepastyear,Sonic
hascontinuedtobuildfiberinfrastructureinSanFrancisconeighborhoodswhereconstruction
costsarelowest(generally,areaswithoverheadutilities,whichtendtoofferlowercoststhan
undergroundconstruction).SonicsleadershipsuggestedinrecentmeetingswiththeCitythat
thecompanywillexpanditsFTTPfootprinttoadditionalareasofSanFranciscowithaerialutilities
overthenextyearorso.Eveninaerialareas,Sonicsconstructionisnotuniformandtendsto
skip areas where costs and complexity are higher. We thus do not expect Sonics fiber
constructiontobeubiquitous,evenintheneighborhoodswhereitplanstobuild.
Whereavailable,Sonicsgigabitserviceis$40permonth($58includingtaxesandfees),77and
includesafreephoneline.
ThefollowingmapillustratesthereachofSonicsservices,andisbasedonselfreporteddata
providedbySonictotheFCC.78
76
DaneJasper,SonicFusionGigabitFiberNowAvailableinSanFrancisco,BlogPosting,Sonic,February25,2016,
https://corp.sonic.net/ceo/2016/02/25/sonicfusiongigabitfibernowavailableinsanfrancisco/(accessedJune
2017).
77
BernieArnason,SonicBringsDisruptivePricingforGigabitServicesAheadofGoogleFiber,Telecompetitor,
February26,2016,http://www.telecompetitor.com/sonicbringsdisruptivepricingforgigabitservicesaheadof
googlefiber/(accessedJune2017).
78
ThemapisbasedonFCCForm477dataasofJune30,2016,https://www.fcc.gov/general/broadband
deploymentdatafccform477(accessedJune2017).
57
Figure11:SonicServiceinSanFrancisco
4.8 Factors That Can Impact Private Deployment Patterns

Every company obviously makes unique decisions about deployment based on its technology
choices, available capital, business plan, and return on investment requirements. In an
environmentwheretheprivatesectorhascompletecontroloverthatdynamic,thereexistsa
widerangeoffactorsthatwilldeterminedeploymentpatterns.
Amongtheseareproximitytoexistingbackboneandbackhaulfiberplant.AT&T,forexample,
tendstooffergigabitservicesaroundthecountrytomultidwellingunitsandsinglefamilyhomes
thatarelocatedincloseproximitytoitsexistingfiberoptics.
Anotherfactorisproximitytoareaswithsignificantpotentialenterprisebusiness.Aswiththe
formerfactor,manycompanieswillhaveinvestedinfiberinfrastructuretoareaswherelarge
enterprise customers that make big monthly purchases of services are located. As a result,
neighborhoods with some proximity to these enterprise customers tend to offer better
economicsfornewresidentialdeployment.
58
Theavailabilityofaerialutilitypolespaceisafurthercriticalfactor,solongasutilitypolesarein
reasonablygoodconditionanddonotrequirecostlyreplacementorextensivemakereadyto
preparethemforanewattachment.
Closelyrelatedtowhetheraerialutilitiesareavailableornotisthecostfactorassociatedwith
density.Thehigherthedensityinaneighborhood,thelowerthepercustomerconstructionor
upgradecosts(allotherthingsbeingequal);obviouslythesavingsassociatedwithdensitycanbe
negatedbythepotentialhighercostsofexpensivepolereplacementandmakereadyorcostly
undergroundconstruction.
Anotherrelatedfactoristhetypeofhousingintheneighborhood,andwhetheritissinglefamily
orMDUs.ConstructiontoMDUscanbemorecosteffective,assumingthattheinsidewiringin
thebuildingisusablefornewbroadbandpurposes,andthattheserviceproviderhasaccessto
themarketintheMDU.(ThisisfrequentlynotthecasebutSanFranciscobecamethefirstcity
inthecountrytopassalawregulatingatenantsrighttochoosecommunicationsservicesand
accesstotheinsidewiringownedbythepropertyowner.)
Finally, demographics play a significant role in private sector investment calculus around
broadbandupgradesorexpansion.Again,allotherthingsbeingequal,familieswithhigherlevels
ofeducation,multiplebreadwinners,multiplechildrenofinternetuseage,andsoontendtobe
highrevenuecustomersforbroadbandinternetservice.Neighborhoodsandcitiesinwhichthese
potential customers are concentrated are of particular interest to private investors. In San
Francisco,thesedemographicfactorswillplaylessofarolethaninothercities,wheretheoverall
demographicfactorsarelessattractivethaninSanFrancisco.
Thetotalityoftheseandotherfactorswillimpactprivatedeployment.Inamodelinwhichthe
City leaves broadband investment entirely to the private sector, other than through some
facilitation or other lowcost strategies, these factors rather than the Citys policy priorities
around digital opportunity, digital equity, and universal access will determine deployment
patterns.
4.9 Benefits and Challenges of the Private Investment Model

The foremost benefit of the private investment model is private investment and lack of City
financialrisk.Totheextentthattheprivatemarketdeliversnewbroadbandcapabilitiesandnew
broadbandcompetition,theCityisinapositiontopartiallyorincrementallymeetitspolicygoals
without risk to the taxpayers. That said, as is discussed throughout this document, private
investmentisanunevenandincrementalsolution,ratherthanacomprehensiveanduniversal
approach.
59
Alongwiththebenefitofnothavingtofundorfinanceabroadbandnetwork,theprivatesolution
candelivertheconsiderablebenefitsofprivatesectorinnovationandcreativity.Particularlyina
singular market like San Francisco of early adopters and technologyadept citizens, private
networksintheorywillrespondtothecapabilitiesofthemarketthroughserviceinnovationand
creativity.Wenote,though,thatthebroadbandmarkethashistoricallyseenlessofthiskindof
innovation than one would want, given lack of competition and opportunity for network
operators to use the monopoly or duopoly status to protect existing business models and
revenuestreams.
The modest emerging competition in some parts of San Francisco will hopefully continue to
impact that static dynamic, as has been the case to some extent in markets throughout the
country since the early Google Fiber announcements stimulated modest new investment by
incumbents.ButtheincrementalincreasesincompetitioninsomepartsoftheCitywouldnot
havethesametransformativeimpactwithregardtocompetitiveservicedeliveryandinnovation
aswouldacomprehensive,universalcitywidedeployment.
Amongtheconsiderablechallenges,fromtheCitysstandpoint,ofleavingbroadbandtoalargely
noncompetitive private market is the deeply uneven benefit to some communities and
neighborhoodsrelativetoothers.Aswehavedocumentedhere,withtheexceptionofComcast,
privatesectordeploymentpatternsarenotcomprehensiveorconsistentthroughouttheCity.
This has an impact not only in regard to services and availability, but also in regard to the
affordabilitythatispresumablycreatedbyrobustcompetition.
4.10 Risk of Consolidation and Reduction in Competition

In the current environment, broadband competition faces political challenges. Given the
regulatory environment in Washington, we are moving toward far greater consolidation and
monopolizationofcommunicationsinfrastructureownership.
Forexample,weanticipateaveryfriendlyregulatoryapproachtowardmergers,giventhecurrent
leadershipattheFederalCommunicationsCommissionandDepartmentofJustice.Indeed,we
arealreadyseeingawaveofconsolidationamongenterpriseproviders.Thatconsolidationbegan
a number of years ago and is likely to accelerate, as demonstrated by CenturyLinks pending
acquisitionofLevel3.
Weexpectthisconsolidationtoreachevenfurtherintotheindustry,aswell.Theprospectof
VerizonsacquisitionofeitherCharterorComcast,forexample,wouldnotonlyreducewireline
broadbandcompetition,butwouldalsothreatenthepromisedmobilecompetitionthatcable
companiesareplanningtooffertomobilephonecompanies.
60
5 P3 Structure Could Enable the City to Achieve the Goal of Open,

Ubiquitous Fiber
P3representsapromisingmeansbywhichtomeetSanFranciscosgoalsforthisinitiative.Policy
goals such as ubiquity, openness, and pricing can potentially be negotiated with the private
partnerbecausetheCitysinvestmentbuysitinfluence.
5.1 P3 Background
P3 is a new term in broadband planning, 79 as that term is used internationally to signify an
alternativedeliverymechanismforfinancingconstructionandoperationsofapublicasset.
The formal P3 Structure is largely untested in the U.S. with respect to broadband. To our
knowledge,theKentuckymiddlemileprojectistheonlyformalbroadbandP3tohaveclosed80
and the only other current P3 project contemplated for FTTP (other than this initiative) is a
nascentprojectinTopeka,Kansas,andthesurroundingcounty.81P3hasbeenutilizedforlast
milebroadbandinpartsofAsiaandEurope.
Inotherinfrastructureareas,theformalP3mechanismhaslongbeenapreferredtoolforpublic
infrastructureprojectsinpartsofEuropeandAsia,andhasgainedconsiderablecurrencyinthe
U.S.inrecentyears.Indeed,itappearstobepotentiallycontemplatedforthemuchdiscussed
potentialTrumpinfrastructurebillthatmayemergeoverthenextyear.
The formal P3 mechanism has been used most frequently in the areas of transportation and
public transit, and is increasingly now being used for public projects like libraries, schools,
government buildings, prisons, and convention centers. Formal P3s have tended to be
concentrated in the United States in projects that are at least in part selffundingprojects
79
Thetermpublicprivatepartnershipisusedinarangeofwaysincludingtorefertolessformalarrangements
suchasfacilitationbypublicentitiesofprivateinvestment,suchasGoogleFibersrelationshipwithKansasCity,
Austin,andothers.ButforpurposesofthisdocumentandtheBroadbandforSanFranciscoInitiative,weusethe
termtorefertoaformalnegotiatedstructureunderwhichapublicentityselectsaconcessionairetodeliver,over
somesubstantialperiodoftime,somesubsetofthenecessaryelementsofaninfrastructureproject.
80
KentuckyWiredisthefirstofitskindinthecountry.Thenetworkwillconnect1,100governmententitiesovera
3,400milefiberopticbackboneandexcesscapacitywillbemadeavailabletoprivatesectorcompaniesforlease,
development,andinnovation.Kentuckywillretainownershipofthenetwork;theconcessionairewillbuild,maintain
andoperatethenetworkfor30years.Theprojectisnotableforitspoliticallybipartisansupport.KentuckyWired
wasfirstannouncedbyRep.HalRogers,aRepublican,andthenGov.SteveBeshear,aDemocrat,in2014.Gov.Matt
Bevin,aRepublican,announcedhissupportfortheprojectafterhiselectionin2016.KentuckyWins'Dealofthe
Year' for KentuckyWired Fiber Optic Project, News, CTC Technology & Energy, December 8, 2015,
http://www.ctcnet.us/blog/kentuckywinsdealoftheyearforkentuckywiredfiberopticproject/(accessed June
2017);KentuckyWiredStatewideBroadbandNetworkInitiativeMovingForwardinEasternKentuckyandBeyond,
PressRelease,KentuckyWired,September16,2016,https://goo.gl/QxzjYx(accessedJune2017).
81
JEDOBoardMeeting,Minutes,December14,2016,CityofTopekaandShawneeCounty,JointEconomic
DevelopmentOrganization,http://www.jedoecodevo.com/Meeting
Documents/2016Documents/12.14.16Agenda%20Packet.pdf(accessedJune2017).
61
wherearevenuestreamisassociatedwiththeinfrastructureproject,andthoserevenuescanbe
appliedtothepublicentitysfinancialobligationtotheconcessionaire.
Thus, for example, in the case of a toll road, a hypothetical P3 might anticipate certain toll
revenuesoveraperiodoftimethattheconcessionairewouldcollectandkeeppotentiallywith
aminimumrevenueguarantee(fromthestatedepartmentoftransportationthatisthesponsor
oftheP3)intheeventthattollrevenuesdonotmeetthenegotiatedtargets,andpotentiallywith
revenuesharingbetweenpublicandprivateentitiesintheeventthatrevenuesexceedcertain
targets.
EachP3isobviouslyunique,withacustomizedmechanismforaddressingrevenuerisk,aswell
aslongtermrevenueupside.SimilarlycustomizedforeachP3wouldbeallocationofafullrange
of risks, including construction risk and performance risk. Public entities that sponsor
infrastructureP3sfrequentlydosoinordertoefficientlyallocatesuchrisksasbesttheycan.
Thus,forexample,aformalP3couldpotentiallyallocateperformancerisktotheprivatesector
to get the advantage of presumed private sector efficiencies and operational capacitywhile
possibly still allocating some construction risk to the public partner, particularly if public
processessuchaspermitting,accesstoland,andsoon,arecontributorstorisk.
AnotherprimaryreasonforpublicsponsorshipofP3sisascenarioinwhichtheP3providesaccess
tocapitalforthepublicinfrastructureproject.Privatecapitalofthissortisnotnecessarilylow
cost.Indeed,itisalmostalwaysgoingtobehighercostcapitalthanwouldbepublicbonddebt.
Butintheeventthatthisisthemostaccessibleoronlyaccessibleformofcapital,thisisauseful
toolforagovernmenttoproceedwithprojectsthatotherwisewouldhavetowait.
5.2 The Core Structure of the Formal P3

Atitscore,theconcessionairemodelrepresentsapublicgrantofrightstoaprivatepartnerthat
contracts to finance, build, and operate a broadband network, with the concession lasting a
considerable period of time (likely 30 or more years). The private partner agrees to provide
financing,construction,operations,andservicedeliveryoverthenetwork.
Tofundallthisactivityandinvestment,thelocalityagreestoanarrangementthatguaranteesa
certain level of revenue to the concessionaire. This revenue mechanism could take any form
legallyavailabletothelocality.Dependingonstateandlocallaw,thelocalitycouldprovidea
direct revenue stream out of general funds, through a specific fee or tax, or otherwise.
Alternatively,thepartnershipcouldbestructuredsuchthattheprivatepartnersrevenuestream
derivesfromenduserpayments(servicefeesfordata,voice,andvideoservices,paidbynetwork
customers), with a minimum revenue guarantee by the locality such that any shortfall in
projectedrevenueswouldbemadeupbythelocality.Hybridarrangementsarealsopossible,in
whichtheprivatepartnerspaymentsarefundedpartiallythroughapublictaxorfeeandpartially
62
throughendConnectionFeesforservices.Inanyevent,theprivatepartnerisunlikelytoassume
allormostofthedemandandrevenueriskinaP3arrangement.
Topartiallyoffsetpublicriskinthisregard,P3arrangementsalsofrequentlyincludeprovisionfor
revenuesharingintheeventthatnetworkrevenuesexceedtheminimumamountsguaranteed
totheprivatepartner.
Some communities that have considered P3s have sought easy out scenarios in which the
localitycancancelthepartnershiporchoosenottoappropriatenecessaryfundsinanygiven
year,andthusdepartitsfinancialobligationsatanytime.Butwithoutthepublicentityslong
termfinancialcommitment,thisapproachisunlikelytobemarketable.
5.3 Applying the Formal P3 Model to Broadband Creates Some Risks

TheapplicationoftheformalP3modeltobroadbandisstillnewanduntestedintheU.S.with
respecttolastmilebroadband,whichisnotanareaoftraditionalpublicsectoroperations.Thus,
forexample,unlikeroads,bridges,andsubwaysystems,broadbandrepresentsaninfrastructure
categorythathaslargely(thoughnotexclusively)beentheprovinceoftheprivatesector.Asa
result,aformalP3inthecontextofbroadbandfacescertainanalyticalandpracticalchallenges
thatarenotexperiencedinthesamewayinotherP3sectors.
Asaresult,traditionalP3riskmodelingformunicipalinfrastructureisinsufficientasappliedto
broadband. In a water, sewer, or tollway project, one typically considers traditional financing
factorsandthetypeofdemandthatexistsinamonopolisticenvironment.Furthermore,inthose
infrastructurecategories,theservicesprovidedarelargelystableandpredictable(awaterutility
willprovidewaterforitsentirelifetime,andatollroadwillprovidetransitforvehiclesforits
entirelifetime).This,however,isnotthecaseinbroadband,anindustryinwhichthewaysin
whichconsumersusetheproductisconstantlyinfluxandtheservicesprovidedoverthenetwork
are predictable only in the short to mediumterm. Note, for example, the decline of cable
televisionasaserviceasconsumersmovetointernetbasedvideostreamingservicessuchas
Netflix andHulu. Similarly, wireline phone service, onceauniversallyadopted service, is now
usedbyasfewashalfthehomesintechnologyadeptmarkets.
Indeed,applyingP3tofibertothepremisesintroducesarangeofcriticalmarketanddemand
risks. Because right alongside the Citys network, private competitors operate their own
networks,eventhoughnoneoffersubiquitousfiber,sotheservicestheyofferwillnotmatch
what is possible over the FTTP network that the City envisions. And both the Citysponsored
networkandprivatecompetitorshavetograpplewith,andadaptto,constantandsubstantial
changesinconsumerdemand,servicepreferences,andtechnologychange.Thesefactorsare
dramaticallydifferenttothoseimpactingrevenueriskinaformalP3inamoretraditionalsector
suchaswaterorpublictransit.
63
6 P3 Model Analysis
This section of the report provides an analysis regarding application of alternative delivery
modelstotheCitysproposedFTTPproject.P3shaveneverbeforebeenappliedtoFTTPprojects
intheUnitedStates,thoughtheideahasbeenconsideredinanumberofcommunities.Asa
result,thisisafirstofitskindanalysis.Inthissection,wedescribetheCitysguidingprinciples
andtheprinciplesofrigorousfinancialmodelingthatshapedtheanalysisandrecommendations.
We summarize our analysis and the alternatives we considered, including the typical
infrastructureP3model(theClassicP3Model),anddescribeourrecommendationsfortwo
alternativemeansbywhichtoapplyP3tofibertothepremises.
6.1 Framework Goals Used to Develop Potential Project Delivery Models

WesetasidestandardthinkingaboutbroadbandP3stodevelopframeworkgoalsthatwould
guideourassessmentofpotentialapproachesand,ultimately,thecreationofacompletelynew
broadbandP3modelthatmeetstheCitysgoalsandcreatesafeasibleinvestmentvehiclefor
privatepartners.Ourthreeframeworkgoalsareasfollows.
6.1.1 Create a Broadband Market That Provides LongTerm Best Value for
Taxpayers, Users, and the City
We seek to create longterm best value for the City by creating a level playing field and
competitivemarketattheserviceleveland,ideally,atthedarkfiberlevel,whichwebelievewill
leadtoreducedcostsandimprovedservicesforusers.Toachievetruecompetition,themodel
wouldneedtosupporttheentryofmultipleserviceproviders.
SomelevelofCityand/ortaxpayersupportwillbenecessaryforaprojectofthissizeandnature
butenablingadditionalISPsandapplicationproviderstoenterthemarketwilllikelyyieldmore
revenuesandlessenthefinancialburdenontheCityanditstaxpayers.
To ensure that this principle remains in place over the long term, we seek flexibility to
accommodatechangesovertimeintermsoftechnology(whichwillimproveovertime)aswell
assupplyanddemandforservices.
6.1.2 Create a Feasible Business Case for the Project

Investments in the project will be significantand the projects revenues will need to be
sufficient to not only recover the investments, but also to cover operational costs and
compensatepartiesforabsorbingrisks.
The guiding financial question, then, is how to minimize life cycle costs as much as possible,
generateenoughrevenues,andreducetheprojectsriskprofile.Theseoutcomesarenotagiven,
but can beoptimized by scoping the project with these needs in mind, and by structuring its
fundingmechanisms,financingstructure,anddeliverymodel.
64
6.1.3 Create a Structure That Can Be Financed

Asafundamentalstartingpointinourplanning,werecognizethattheCitydoesnotwantto
retainallrisksandfinancetheprojectitself;rather,theCityseekstouseprivatefinancingandto
transfersomeriskstotheprivatesector.
Thatsaid,constructionandrevenueriskforaprojectsuchasthisareverysignificant;thishigh
riskprofilemaybedifficultorimpossibleforfinanciersandinvestorstoaccept.
Asaconsequence,financingcostsmaybeextremelyhigh,orthedealmaynotbefinanceableat
all. Therefore, the project needs to be structured such that it leads to a risk profile that is
acceptabletofinanciersandinvestors.
6.2 Classic P3 Model Illustrates Limitations of Broadband P3 Approaches

Asabenchmark,webeganouranalysisofbroadbandP3optionswiththeclassicP3modela
middlemilevariationofwhichwasfirstsuccessfullynegotiatedforabroadbandinitiativeby
theCommonwealthofKentuckyingrantingamiddlemilefiberconcessiontoMacquarieCapital
in2014.82Inouruseofthetermclassicmodel,wearereferringtothefactthatthereexistsa
generalframeworkforP3sacrossdifferentassetclassesandsectorssuchaspublicutilities,social
infrastructure,publictransit,andtransportation.InthecaseoftheKentuckyproject,theasset
class(fiberoptics)isofcoursedifferent,asarethetypeofentitiesthatwillleaseandusethe
asset,buttheconceptandgeneralP3structurearethesame.
Figure12:ClassicP3Model

AsisillustratedinFigure12,theCitywouldnegotiatealongtermcontractwithaP3broadband
concessionaire. The concessionaire would be responsible for delivering a turnkey solution,
82
CTCwasanindependentconsultanttotheCommonwealthofKentuckyinitsnegotiationswithMacquarie.
65
includingtheoutsideplantinfrastructure,lightingofthenetwork,andprovisionofalloperations
andservices.Inotherwords,theconcessionairewoulddesign,construct,finance,operate(i.e.,
deliver services to end users and/or sell capacity to competing ISPs to deliver services), and
maintainthenetwork.TheCitywouldsponsortheinfrastructure,83buttheconcessionairewould
controlaccesstoitundernegotiatedandagreedupontermswiththeCity.
TheCitywouldachievesomeofitsprimarygoalsunderthismodelincludingtheconstruction
of a futureproof fiber optic broadband network and delivery of services to residents and
businesses.TheCityscontractwiththeconcessionairewouldalsoshieldtheCityfromsomelong
termprojectrisks.Theconcessionairewouldholdthelongtermperformancerisksassociated
with the project, as well as revenue risks (e.g., commercial revenue from base level service
providers,leasepaymentsfromotherinternetandapplicationproviders).
Thatsaid,pursuingtheClassicP3Modelhassomeshortcomings:
1. Limitedcompetitionreduceslongtermvalue
Because the concessionaire is completely responsible for services over the infrastructure
delivering base level services to all households and businesses, additional retail services, and
wholesale services to service providersthis model limits competition. This is an issue for
creatingalevelplayingfieldwithrespecttoretailservices,andforincentivizingacompetitive
market that can effectively drive down prices for end users and offer the other benefits of
competition.
Inaddition,theremaybelimitedabilitytoaccommodatechangesovertimeduetothelongterm
natureofP3contracts.Thismaynotbeasimportantofanissueatthedarkfiberlevel,asfiber
represents a longterm asset and has a longer replacement life cycle. But for investments in
networkequipment,operations,andserviceprovision,theeconomiclifeofthisequipmentis
onlysevento10yearswhichishardtoaccommodatewithinalongtermP3concessionaire
model.
2. Uncertainrevenuesincreaserisktoprojectbusinesscase
Potentialcommercialrevenuesunderthismodelareuncertain,duebothtotheunpredictable
natureofbroadbandequipmentandnetworktechnologiesandthematurityoftheSanFrancisco
market.
In this model, for example, the concessionaire would hold risks associated with commercial
revenue from base level service providers, and lease payments from other internet and
83
ThetermsponsorcomesfromthetermProjectSponsorwhichistheentitywiththeoverallaccountabilityfor
theproject.Inthiscase,theCityofSanFranciscoisthesponsoroftheproject.TheCityhasultimateresponsibility,
andisaccountabletothepublic/taxpayers.
66
applicationproviderswhichareuncertainrevenuestreams.Thisisunfavorablefortheprojects
businesscase.
DedicatingrevenuefromuserscouldreducetherevenueriskprofilefortheP3concessionaire,
butwouldonlypartiallymitigatetherisk.
3. Riskmitigationrequiredtoattractinvestorsandfinanciers
Due to the uncertain revenue streams created by this model, financing a project using this
approachwouldbequitechallenging.Itisunlikelythatfinanciersandinvestorswillbewillingto
investunlesstheywouldhaveaquasimonopoly,includingratesettingpowers,and/orsome
typeofcontributionfromtheCity(e.g.,aguarantee,risksharingmechanism,orpartialpayment)
tomitigaterisk.
67
7 Alternative P3 Approach: Dual P3s for Dark and Lit Fiber

To ameliorate the Classic P3 Models shortcomings in the broadband environment, and to
maximizethepotentialbenefitsoftheP3structureforfibertothepremises,theProjectTeam
developedanewP3approachforconsiderationbytheCity.
TheapproachisadualstructurethatincludestwoseparateP3s(DualP3Model).TheDualP3
Model(Figure13)seekstoaddressthechallengesinherentintheothermodelsbyreimagining
theestablishedinfrastructureP3andcreatingtwoseparateP3sfortheprojectoneforthedark
fiberinvestments(DarkFiberP3)andonefortheequipment,operations,andserviceprovision
investments(LitFiberP3).
Figure13:DualP3Model
7.1 Overview of the Dual P3 Model

The Dual P3 Model recognizes that the most critical challenge in the Classic P3 Model is the
disparate risk profiles associated with the dark fiber investments on the one hand, and the
equipment/network operations/service provision investments on the other. Combining these
twodistinctriskprofileswithinoneP3ishighlychallenging,andwouldencumberthebusiness
casefortheproject.Asaresult,theDualP3ModelcomprisestwoseparateP3contracts:
3. DarkFiberP3:Includesdesign,build,partialfinance,andoperationsandmaintenance
(O&M)ofdarkfiberandotherpassiveoutsideplantsuchascommunicationshutsand
splitters. This is a longterm P3 contract (e.g., 30 years) for dark fiber/outside plant
investments;itpotentiallyleveragespartialpublicfinancing.
68
4. Lit Fiber P3: Includes design, build, and O&M of network electronics investments, and
provisionofabaselevelofservice,aswellasprovisionofwholesaleservicestoRetail
ServiceProviderswhowillcompeteoverthenetwork.ThisisashortertermP3contract
designed to enable responsiveness to technology change and the need to replace
networkelectronicseverysevento10years.
Thismodelrecognizesthattheriskprofileforconstructingandmaintainingafibernetworkis
different from the risk profile associated with providing services over the long term in a
commercial broadband market. The P3 for the dark fiber/outside plant investments can be
separatedfromtheP3forelectronicequipment,networkoperations,andprovisionofbaselevel
servicebecauseinterrelationshipsbetweenthetwoarelimited.Thisapproachresultsinamore
explicitdistinctionbetweenthedarkfibernetworkproviderandtheoperator/serviceprovider
onthatnetwork,withtheintenttoenablemoretechnologicalflexibilityandcompetitionatin
networkoperationsandserviceprovision.
7.1.1 Potential Participating Entities

TheDualP3Modelassignsrolestoawiderangeof participants. The CityandCountyofSan
Francisco would be the sponsoring agencyproviding an oversight role, and paying the
concessionaireforservicestotheCity.TheCityisultimatelyaccountabletotheresidentsand
taxpayersoftheCity,whoaretheintendedusersandbeneficiariesofthisproject.
Residents and businesses in the City of San Francisco would fund some of the infrastructure
through some form of assessment, utility tax,propertyrelated fee, parcel tax, or any type of
approvedrevenueinitiative.84Ifapropertyrelatedfeeisapplied,thenprovisioningofbaselevel
serviceswillneedtobeincludedintheP3tojustifythefee.85
Apublicbroadbandentity(PublicBroadbandEnterprise)wouldbecreatedbytheCitytomanage
the multiple P3 agreements, organize leasing of the assets, and oversee lease payments and
otherfinancialtransactions.ThisrolecouldbetakenbyaCityagency(existingornewlycreated)
orsomeotherpublicentity,potentiallyevenincluding(limited)privateinvolvement.86ThePublic
BroadbandEnterprisecouldalsoreceiveuserspropertyrelatedfeesorparceltaxpaymentsand
issuepublicfinancing,ifthatapproachisselected.Alternatively,receiptoffeesandissuanceof
84
InconsultationwiththeCityAttorneysoffice,theProjectTeamdevelopedamatrixandbusinessanalysisof
potentialfundingmechanismsthatisincludedintheappendicestothisreport.
85
SeeSection7.3formoredetail.
86
Inordertohelpmaximizethevalueatthesystemlevel,itwouldbeanoptiontohaveinvolvementfromprivate
shareholders.ApotentialwaytostructurethiswouldbetostartoffbysellingasmallpercentageofthePublic
BroadbandEnterprisesshares.Privateownershipcouldpotentiallybeexpandedovertime(whentheriskprofile
decreasesandthevalueofthePublicBroadbandEnterprisesharesisexpectedtoincrease).
69
financing could be fulfilled by a nonprofit corporation that works in parallel with the Public
BroadbandEnterprise.87
Internetandapplicationserviceproviders(ISPsandASPs)wouldincludeanyentitythatmakes
apaymentforaccesstodarkorlitfiber.Thesemayincludelargecorporations,institutions(e.g.,
universities),andserviceprovidersofarangeofsizes,includingsomethatareverysmallbutwho
are able to compete over the network by purchasing wholesale services from the Lit Fiber
Concessionaire.
The dark fiber design, build, finance, maintain (DBfM) concessionaire (Dark Fiber
Concessionaire) will provide investments in dark fiber and other outside plant assets and,
followingconstruction,outsideplantoperationsandmaintenance(O&M).
Abaselevelserviceconcessionaire(LitFiberConcessionaire)willprovideequipmentandO&M
investments,andwilldeliverbothbaselevelservicestohomesandbusinessesandwholesalelit
fiberservicestoother ISPs.Importantly,thispartnerwillbeprecludedfromdeliveringretail
servicesotherthanbaselevelservices.
7.1.2 Financial Flows

TheDualP3ModelissimilartotheClassicP3Modelinthatitwillincludesomefinancialfeatures
that were crafted to address issues that are universal to the Broadband for San Francisco
Initiative,regardlessoftheP3model.Thismodel:
AttributesafixedpartialpaymentfromtheCitytotheproject(i.e.,theCitywillcommit
tosomelevelofdedicatedfundingoftheproject)tomitigatetheconcessionaireshigh
risksofuncertainprojectrevenues
ProvidesaminimumrevenueguaranteefromtheCity
AllowsforupfrontleasepaymentstobeagreeduponbytheCityandotherinternetand
applicationproviders
The City will be in a position to make a contribution for data services it receives from the
concessionaireandthatarecurrentlyprocuredfromotherpartiesintheformofleasedcircuits.
Theamountofthatcontributionwouldpotentiallybebasedonavoidedcosts,andwhilewehave
not been provided data in this regard, based on experience in other cities, we estimate the
amounttobeinexcessof$10millionperyear.
87
Formoreinformationon6320corporations,see:
http://www.financingtransportation.org/funding_financing/legislation_regulations/state_legislation/non_profit_6
3_20.aspx
70
SanFranciscosresidentsandbusinesseswillalsoprovidesometypeofdedicatedfunding,for
boththeDarkFiberP3andtheLitFiberP3.Ifapropertyrelatedfeeisused,thenprovisionof
baselevelservicesisneededtojustifythefee,andprovisionofbaselevelserviceswillneedto
beincludedintheP3.
TheLitFiberConcessionairewouldberesponsibleforacombinationofequipmentinvestments,
maintenance,andbaselevelserviceprovision.Additionalfinancialflowsinclude:
CitypaymentstotheDarkFiberConcessionaire
o Milestonepayments:PaymentsfromthepublicentitytotheDarkFiber
Concessionaireforthedarkfiberconstruction
o Availabilitypayments:PeriodicpaymenttotheDarkFiberConcessionairefor
makingtheinfrastructureavailableatagreeduponstandardsforthedurationof
theP3contract
CitypaymentstotheLitFiberConcessionaire
o Servicepayment:PeriodicpaymenttotheLitFiberConcessionaireforproviding
servicesattheagreeduponstandardsforthedurationoftheP3contract
Intermsoffinancing,webelievetheDarkFiberP3wouldrequirepartialpublicandpartialprivate
financing,giventhehighcostofthedarkfibernetwork.Webelievethattherequiredinvestments
fortheLitFiberP3canbedeliveredthroughprivatefinancing.
7.1.3 Key Contractual Elements
7.1.3.1 Terms / periods

No technological flexibility is necessary at for the dark fiber/outside plant because fiber
technologyisnotexpectedtobealimitingfactor;therefore,alongtermcontractissuitable.
Incontrast,fornetworkequipment,operations,andserviceprovision,arelativelyshortcontract
termwouldallowforflexibilitytomodifystandardsovertimetoaddressfuturetechnological
changes.Theeconomiclifeofnetworkequipmentissevento10years,dependingonthetypeof
equipmentso this P3 concession should ideally be reprocured every seven to 10 years to
maintain technological currency. This regular revision of technical standardsand continuous
competitionwill increase the likelihood that San Franciscans will receive a high quality of
serviceandthelatesttechnicalstandardsatattractiverates.
To further promote competition, the equipment that is part of the Lit Fiber P3 would be
transferredtothePublicBroadbandEnterpriseattheendofthecontracttermandwouldbe
availabletotheLitFiberConcessionaireselectedforthenewterm;detailedinformationabout
71
theequipmentanditsconditionwouldbepartoftheRFPforthenextLitFiberP3soastoenable
allbidderstobuildtheinformationintotheirplanningandbids.
7.1.3.2 Elements
Asisdiscussedindetailintheprocurementstrategysectionofthisreport,werecommendthat,
priortolaunchingtheP3procurements,theCityundertakealeasingprocesstosecurethirdparty
revenuesfordarkfiber(andpotentiallylitfiberwholesaleservices).Thisapproachwouldfurther
improvefinancialfeasibilityoftheprojectandensureamorelevelplayingfieldforP3bidders.88
7.1.3.3 Allowable activities/generating additional revenues

IntheDualP3Model,concessionairesmaydeliveronlyspecificservices;theyareprohibitedfrom
engaginginotheractivities.
Tomaximizerevenues,revenueincentivescanbeincludedineithertheLitFiberP3orbothP3s.
P3biddersshouldhaveanincentivetosecureadditionalrevenuesduringthebiddingprocess;
securingmorerevenuesupfrontcanreducethefeeortaxtobechargedforestablishingthefiber
networkanddeliveringabaselevelservice.
P3concessionairesshouldalsohaveanincentivetogeneratemorerevenuesovertime.Providing
anincentiveratherthantransferringallrevenuerisktotheP3concessionairewillpresumably
encourage the concessionaire to maximize revenue risk over time, without creating an
excessivelyhighriskprofilethatresultsinveryhighfinancingcosts.
7.2 How the Dual P3 Model Addresses the Project Principles

Splitting the project into two separate P3s maximizes longterm value by increasing
competition.HavingaseparateLitFiberP3thatisreprocuredeverysevento10years(matching
theeconomiclifecycleoftheelectronicsinthenetwork)maximizescompetition,creatingmore
longtermvaluefortheCity.ReprocuringtheLitFiberP3willalsohelptocreatealevelplaying
field among competing service providers, and incentivize a competitive market that can
effectivelydrivedownpricesforendusers.
Thismodelcreatesafeasiblebusinesscase,withmorecertaintyregardingrevenue,throughan
upfrontsaleorpreleaseofdarkfiberandothernetworkcapacity,inwhichtheCitywillseekto
securecommercialrevenuespriortoprocurementsuchthatallpotentialbiddershavesome
dedicated commercial revenue to include in their proposed projects. Absent this preP3 sale,
potentialcommercialrevenues(darkfiberleasefees,feesforwholesalelitaccesstocustomers)
would be highly uncertain, due both to the unpredictable nature of broadband technology
development and broadband services, as well as the relative maturity of the San Francisco
market.SplittingtheprojectintotwoseparateP3salsomitigatessomeoftherisksassociated
88
Section9.1providesdetailregardingthisconsideration.
72
withlongtermuncertaintyforrevenuestreamsbymakingashortertermP3fortheLitFiber
portionoftheproject.
Thismodelalsocreatesastructurethatcanlikelybefinanced.Byseparatingtheprojectinto
twoP3s,therecommendedmodelalsoseparatestwodistinctriskprofilesmakingtheproject
moreattractivetoinvestorsandfinanciers,whocanchoosetoinvestintheP3withariskprofile
mostappropriatetotheirriskappetite.Additionally,theactionstakentocreatemorecertainty
regardingrevenuescanalleviatesomeofthepressurethatfinanciersandinvestorswouldhave
putontheprojecttohaveaquasimonopoly,includingratesettingpowers,and/orsometype
ofcontributionfromtheCity(e.g.,aguarantee,risksharingmechanism,orpartialpayment)to
mitigaterisk.
7.3 Financial Analysis and Models of Dual P3

AspartoftheanalysisfortheBroadbandforSanFranciscoproject,theProjectTeamanalyzed
thefinancialfeasibilityoftheDualP3Model,basedonourconstructioncostestimatesforfull
FTTPdeployment.Thissectionisadescriptionofthefinancialanalysis,includingassumptions,
andpreliminaryfinancialresultsandsensitivities.
Financialfeasibilityanalysisistypicallyconductedtounderstandthefinancialimplicationsofthe
project.Specifically,thefinancialfeasibilityanalysisofthisprojectseekstoanswerthefollowing
questions:
Giventhecapitalandoperationscostsfortheproject,whatistheexpectedresidential
feeperpassing?
Giventhecapitalandoperationscostsfortheproject,whatistheexpectedbusinessfee
perpassing?
Howdocostvariations(forexample,lowincomesubsidyorincreasingcapitalcosts)
influencethesefees?
Howdorevenuevariations(forexample,fromupfrontcommitmentsorfromCity
resources)influencethesefees?
WhatistheexpectedannualavailabilitypaymenttotheP3concessionaires?
Currently,theprojectisatthepreprocurementstage,whereanswerstothequestionsabove
can help decisionmakers understand whether to proceed with the project. Specifically, the
expectedfeesperpassinginfluencetheConnectionFees.TheleveloftheConnectionFeecan
helpdecisionmakersunderstandthefeasibilityofthisprojectfromapoliticalstandpoint.
73
Thefoundationofthefinancialfeasibilityanalysisisafinancialmodelthatwastailorbuiltforthis
projectand based on a base case consisting of preliminary assumptions. More indepth
project preparations are needed, including an industry sounding and a preprocurement sale
processtofirmupassumptionsoncosts,potentialupfrontrevenues,publicresources,subsidy,
financingconditions,andsoon.Thefinancialmodelandsubsequentfinancialfeasibilityanalysis
donotincludeananalysisoftheCitysfinancialresourcesforthisproject.
7.3.1 HighLevel Overview of Cash Flows in Broadband P3 Model

OuranalysisisbasedontheDualP3model,consistingoftwoseparateP3s,asdescribedabove.89
Figure14:DualP3Model

The Dual P3 Model also includes a Public Broadband Enterprise (PBE). The PBE manages the
multipleP3agreements,organizesanupfrontleasesale,andoverseesleasepaymentsandother
financialtransactions.ThisrolecouldbetakenbyaCityagency(existingornewlycreated)or
someotherpublicentity.
Keyfinancialflowsinthismodelconsistofthefollowing:
89
DarkFiberP3:Includesdesign,build,partialfinance,andoperationsandmaintenance(O&M)ofdarkfiberand
otherpassiveoutsideplantsuchascommunicationshutsandsplitters.ThisisalongtermP3contract(e.g.,30years)
fordarkfiber/outsideplantinvestments.
LitFiberP3:Includesdesign,build,financingandO&Mofnetworkelectronicsinvestments,andprovisionofabase
level of service, as well as provision of wholesale services to retail service providers who will compete over the
network.ThisisashortertermP3contractdesignedtoenableresponsivenesstotechnologychangeandtheneed
toreplacenetworkelectronicseveryfiveto10years.
74
CitypaymentstotheDarkFiberConcessionaire
o Milestonepayments:PaymentsfromthePBEtotheDarkFiberConcessionaire
forthedarkfiberconstruction
o Availabilitypayments:PeriodicpaymenttotheDarkFiberConcessionairefor
makingtheinfrastructureavailableatagreeduponstandardsforthedurationof
theP3contract
CitypaymentstotheLitFiberConcessionaire
o Availabilitypayment:PeriodicpaymentfromthePBEtotheLitFiber
Concessionaireforprovidingservicesattheagreeduponstandardsforthe
durationoftheP3contract
All City payments to both concessionaires will come from the PBE. The PBE will be funded
through:
ConnectionFeesthepaymentsfromhouseholdsandbusinessestofundtheP3s,most
likelystructuredintheformofapropertyrelatedfee,orparceltax
PossibleservicefeesfromtheCityforuseofthebroadbandnetworksimilartoexisting
coststhattheCityincursforcurrentbroadbandservices
Potentialpayments/revenuesassociatedwiththisprojectinclude:
Leasepaymentsfordarkandorlitfiberfromlesseesthatarecommittedupfrontviathe
preprocurementsale
Additionalleasepayments/revenuesconfirmedbyeitherconcessionaire
7.3.2 Assumptions per Entity
7.3.2.1 General Assumptions

TheProjectTeamassumesthattheprojectwillbeginafteranappropriateprojectpreparation,
analysis,andprocurementperiod.Themainassumptionforthebasecaseisthattheprojects
costs are funded by Connection Fees and potential additional revenueshowever, these
potentialadditionalrevenuesareunknownatthistime.Themodeliscurrentlyaflatmodel,
assumingnoindexingofcostsorrevenuesforinflation.
75
7.3.2.2 Assumptions for Dark Fiber P3
7.3.2.2.1 Timing
Constructionofdarkfiberisassumedtooccuroverthreeyears.Asdarkfiberinfrastructureis
installed,itwillbemadeavailableforuse.Therefore,anappropriateproportionofassociated
operationsandmaintenancecostswillbeincurredasdarkfiberinfrastructureisinstalledover
the threeyear construction period. Logically, the availability payment for the dark fiber
infrastructure, made to the concessionaire, will also be provided in proportion with the
percentage of infrastructure made available. We assume a construction rampup profile as
follows:year1,30percent;year2,50percent;year3,20percent.Theoperationsperiodofthe
infrastructureis30yearsfromcompletionofconstruction.
7.3.2.2.2 Capital Expenditures

Forthebasecaseusedinthisfinancialfeasibilityanalysis,thetotalestimatedcapitalexpenditure
isapproximately$1.5billion.Thesecostsrepresenttheequipment,material,andconstruction
laborassociatedwithbuilding,andimplementingadarkfibernetwork.
Table16:BaseCaseDarkFiberP3CapitalExpenditure
DarkFiberP3
BaseCase
CapitalExpenditure Description
Cost
Elements
Engineering,QA/QC,generaloutsideplant
construction,specialcrossings,backbone
$776
Outsideplant anddistributionplantsplicing,backbone
million
hub,termination,andtesting,FTTPservice
dropandlateralinstallations.
Assumesresidentialdrops(354,807
Customerpremises $728
passings)andbusinessdrops(18,674
equipment million
passings)
7.3.2.2.3 Operational Expenditures

Operations expenditures for the dark fiber infrastructure have various timing profiles. Some
operational expenditures are linked to the ramp up profile of the capital expenditures (for
example, for fiber and network maintenance), while others are linked to a separate rampup
profile: year 1, 50 percent; year 2, 50 percent (for example, for office expenses). Annual
operationscostafterconstructionisapproximately$23million.Thisincludeslabor,insurance,
offices, facilities lease, locates and ticket processing, contingency, fiber and network
maintenance,legal,consulting,educationandtraining,andpoleattachments.
76
7.3.2.2.4 Revenues
Revenues for the Dark Fiber P3 consist of two main sources, milestone payments during
construction,andavailabilitypaymentsasthedarkfiberinfrastructureismadeavailable.Forthe
basecase,theProjectTeamassumedtwomilestonepaymentsattheendofthesecondandthird
yearofconstruction,totaling$1.4billion.ThisassumptionwasmadebasedontheProjectTeams
professionaljudgementthattheP3marketwouldexpectthatthemilestonepaymentfromthe
Citywouldfundapproximately90percentofcapitalcost.90
Availabilitypayments will be transmitted to the Dark Fiber Concessionaire from the PBE. It is
optimizedtocoverforallcosts,exceptforthemajorityofconstructionwhichiscoveredwith
themilestonepaymentsasdescribedabove.Availabilitypaymentswillcommenceinyear1,as
30percentofthedarkfibernetworkisdeployed,andwillrampuptofullpaymentof$36million,
inlinewiththecapitalexpenditurerampupprofile.
Forthebasecase,noadditionalleasepaymentsfromlessees,oradditionalpaymentsfromthe
Cityareassumed.
7.3.2.2.5 Financing
BasedontheProjectTeamsexperiencewithP3transactionsinothersectors,weexpectthatthe
marketwillbereceptivetoabroadbandinfrastructureprojectthatfeaturespartialpublicand
partialprivatefinancingmeaningthatthepublicsectorwillfundaportionofthecapitalcostof
theproject.Basedonexperiencesinother P3 projects,theProjectTeamanticipatesthatthe
public sector can fund approximately 90 percent of the capital cost of the project while still
enjoying the benefits of private financing related to effective risk transfer and performance
incentives. As such, we have assumed milestone payments from the PBE to the Dark Fiber
Concessionaireequalto90percentofthecapitalcost.
PrivatefinancingforP3projectstypicallycomprisesofacombinationofprivateequityanddebt.
Based on the Project Teams experience with other P3 projects, we expect that the private
financestructurewillbehighlyleveraged,with90percentdebtand10percentequityinthebase
case.TheProjectTeamanticipatesthattheDarkFiberP3concessionairewilluseabridgeloanto
meettheportionofthefinancingthatwillberepaidwithmilestonepayments.
7.3.2.3 Assumptions for Lit Fiber P3
7.3.2.3.1 Timing
TheLitFiberP3concessionisassumedtohaveasevenyearcontractdurationwithsubsequent
LitFiberP3procurements.Forthisfinancialanalysis,theProjectTeamchosetomodelonlythe
firstlitfiberconcessionbecauseitisdifficulttopredictmarketexpectationsandassumptions
90
PleaseseeSection7.3.2.2.5foradetailedexplanation.
77
seven years into the future for a market such as broadband that has such fastchanging
technologicalrequirements.
Constructionoflitfiberinfrastructureisassumedtobeginin2020,oneyearaftertheDarkFiber
P3starts,withatwoyearconstructionduration.Asdarkfiberinfrastructureisinstalled,itwillbe
made available for useand the Lit Fiber Concessionaire will follow with its construction
accordingly. Therefore, an appropriate proportion of associated operations and maintenance
costswillbeincurredasthelitfiberinfrastructureisinstalledoverthetwoyearconstruction
period.
The operations period of the infrastructure is five years from completion of construction
totalingtoasevenyearcontractduration.
7.3.2.3.2 Capital Expenditures

Forthebasecaseusedinthisfinancialfeasibilityanalysis,thetotalestimatedcapitalexpenditure
fortheLitFiberP3isapproximately$360million.Thesecostsrepresenttheequipment,material,
andlaborassociatedwithbuildingandimplementinglitfiberinfrastructureoverdarkfiber.
Table17:BaseCaseLitFiberP3CapitalExpenditure
LitFiberP3 BaseCase
Description
CapitalExpenditureElements Cost
Corenetworkequipment,distributionand
Networkequipment $112million
accessequipment(GPON)
Lastmileandcustomer CPEforresidential,CPEforbusinesses,and
$247million
premisesequipment additionalannualreplacementcapital
Reinvestmentsandmajormaintenanceofthelitfiberinfrastructureisimplicitlyincludedbyusing
thesameavailabilitypaymentthatisrequiredtocompletelyrecovertheinvestmentsinthefirst
contract term for subsequent contracts. This allows for the complete renewal of the lit fiber
infrastructureeverysevenyears.

Operationsexpendituresforthelitfiberinfrastructureisassumedtostartin2020,oneyearafter
thecommencementofthedarkfibercontract.Weassumearampupprofileasfollows:2020
50percent,and202150percent.
Annual operations costs after construction is estimated to be approximately $21 million. This
includeslabor,insurance,utilities,officeexpenses,facilityleases,locatesandticketprocessing,
contingency, legal fees, consulting fees, education and training, and vendor maintenance
contracts.
78
7.3.2.3.4 Revenues
Revenues for the Lit Fiber P3 consist of two main sources: milestone payments during
construction, and availability payments as the lit fiber service is madeavailable. For the base
case,theProjectTeamassumedtwomilestonepaymentsattheendofeachconstructionyear,
totaling $147 million. This assumption was made based on the Project Teams professional
judgementthattheP3marketwouldexpectthatthemilestonepaymentsfromtheCitywould
fundapproximately50percentofcapitalcost.91
Availabilitypaymentsforthelitfiberinfrastructureandservicewillalsobeprovidedinproportion
withthepercentageofinfrastructuremadeavailable,commencingin2020with50percentof
theavailabilitypayment,furtherrampinguptothefullavailabilitypaymentinyear2021.
Forthebasecase,neitheradditionalleasepaymentsfromlessees,noradditionalpaymentsfrom
theCity,areassumed.
7.3.2.3.5 Financing
JustlikefortheDarkFiberP3,weexpectthatthemarketwillbereceptivetoaLitFiberP3project
thatfeaturespartialpublicandpartialprivatefinancingmeaningthatthepublicsectorwillfund
aportionofthecapitalcostoftheproject.BecausetheLitFiberP3hasashorterduration,and
lowercapitalcosts,theProjectTeamanticipatesthatapproximately50percentofthecapital
costoftheprojectwillneedtobefundedbythepublicsector,andassuchwehaveassumed
milestone payments from the PBE to the Lit Fiber Concessionaire equal to 50 percent of the
capitalcost.
Private financing for P3 projects typicallycomprise a combination of private equity and debt.
Based on the Project Teams experience with other P3 projects, we expect that the private
financestructurewillbeleveragedwithapproximately90percentdebt,10percentequity.Other
financingconditionsareassumedtobesimilartotransactionswithaverysimilarriskprofilethat
theProjectTeamhasobservedincomparableP3sectors.
7.3.2.4 Assumptions for Public Broadband Entity

ThePBEmanagesthemultipleP3agreements,organizesthefiberleasingprocess,andoversees
lease payments and other financial transactions. This role could be taken by a City agency
(existing or newly created) or some other public entity. The PBE could also receive users
propertyrelated fees or parcel tax payments and issue public financing, if that approach is
selected.Alternatively,receiptoffeesandissuanceoffinancingcouldbefulfilledbyanonprofit
corporationthatworksinparallelwiththePBE.Therefore,thisentitywillnotincuranycapital
91
AdetailedexplanationcanbefoundinSection7.3.2.2.5.
79
expenditure,butwillhaveoperationalexpendituresandrevenues,asdetailedinthefollowing
sections.

Operational expenditures for the PBE include three categories of costs: payments to the P3
concessionaires,operationscosts,andcostofsubsidiesfortargetpopulations.Paymentstothe
P3concessionariesinclude:milestonepaymentstobothDarkFiberandLitFiberConcessionaires,
availabilitypaymentsfortheDarkFiberP3,andavailabilitypaymentsfortheLitFiberP3.
Currently, operational costs for running this department or entity (e.g., labor costs) have not
beenincludedduetolackofinformation.Thisinformationshouldbeincludedinfutureupdates
oftheanalysisandincludecostsforstaffing,runningthedepartment,collectingfees,andsoon.
Theoverallprojectincludessubsidiesfortargetpopulations.Thebasecaseassumptionisthat15
percentofresidentialpassingsareeligibleforsubsidy.Inordertoprovidesubsidiesthatcanbring
the residential Connection Fee down to $10 per month, the PBE will need to provide $486
annuallypereacheligibleresidentialpassing.Theannualcostofthesubsidyatfulldeployment
isapproximately$26million.
7.3.2.4.2 Revenues
Revenues for the PBE are anticipated to include Connection Fee revenue, additional lease
revenues,andcontributionsfromtheCity.However,forthisbasecase,theProjectTeamhasonly
includedConnectionFeerevenue,asadditionalleaserevenues,andcontributionsfromtheCity
areunknownatthispoint.
ConnectionFeerevenueisanticipatedtonet$266millionperyearbyyear4(fulloperations).
ConnectionFeesareanticipatedtobeimplementedintheformofdedicatedrevenueprovided
bytheCityandareaccountedforasaperpassingfee.Thetotalnumberofresidentialpassings
assumedis345,807.
Business passing fees were estimated using the assumptions shown in Table 18 below. For
businesspassings,thereisanadditionalassumptionof2percentbaddebt(inabilitytocollectthe
ConnectionFee).
80
Table18:BaseCaseBusinessPassingFeeAssumptions
Maximumnumber Multiplier
No.ofpassings
ofemployees (%)
4 24,478 100
9 9,445 150
19 4,375 200
49 2,373 300
99 827 300
249 478 500
499 150 500
999 54 700
4,999 66 1,000
9,999 6 1,000
9,999andabove 13 1,000
7.3.2.4.3 Financing
TheProjectTeamanticipatesthattheCitywillneedtoissuebondfinancinginordertomeetits
substantialobligationsintheearlyyearsoftheprojectduetothemilestonepaymentsforcapital
expenditurestothetwoP3concessionaires.BasedontheProjectTeamsexperienceasfinancial
advisorsforbothmunicipalgovernments,andP3transactions,theProjectTeamanticipatesthat
thebondissuanceassociatedwiththisprojectcan,andshould,bestructuredasarevenuebond.
ThisisduetotheassumptionthatConnectionFeesstructuredasparceltaxesorpropertytaxes
willbededicatedtotherepaymentofthisbond.
Thebondservicedurationisassumedtobe27years,withmajordrawdownsinyears2and3of
theprojectinordertopaymilestonepayments.Asistypicalofrevenuebonds,weassume3.5
percentinterest,andminimumDSCRof1.2.
Currently,theProjectTeamdoesnotanticipatethatadebtservicereserveaccountisneeded
duetothesignificantcushionthattheConnectionFeeprovides.
7.3.3 Results
OnekeyfinancialoutputofthebusinesscaseforthisFTTPinitiativeistheConnectionFeethat
willberequiredtomaketheprojectfeasible.Financialindicatorsthatinformthatoutputarethe
payments that the P3 concessionaires will need to cover their investments and operational
expenses,meettheirfinancingobligations,andmakeanappropriatereturnonequityvisvis
thevariousprojectrisksthattheyareacceptingundertheP3agreements.
81
Alltheinputspresentedintheprevioussectionareuncertain,whichiscommonforthisstageof
theprojectdevelopment.Thisiswhytheresultsareshowninrangesratherthanexactoutcomes.
For the purpose of the sensitivity analysis we show a base case, which is based on the
assumptionsasdescribedintheprevioussection.
Theuncertaintiescanbefurtherreducedandtherangeswillbecomenarrowerastheresultof
furtherprojectpreparation.Theresultsofthepreprocurementsaleofcapacitywillsignificantly
reducetheuncertaintyandsowilltheultimateresultsoftheP3procurements.
Tocontributetoabetterunderstandingoftheimpactsofthevariousuncertainties,thissection
alsopresentsasensitivityanalysis.
7.3.3.1 Key Results
7.3.3.1.1 Dark Fiber P3

The operating cash flow of the Dark Fiber P3 has two phases: in the first three years, the
milestonepaymentsfromthePBEarereceived,funding90percentofthecapitalexpensesinthe
firstyears.Aftersubstantialcompletion,thedarkfiberhaslimitedoperationsandmaintenance
expensesandreceivesanavailabilitypayment.

ThefinancingcashflowoftheDarkFiberP3startswithaseniordebtandequitydrawdownof
$151millioninyear1ofthecontract.Seniordebtinterestandfeesstartinyear2.Seniordebt
repayment and payout of dividends start in year 4 and add up to approximately $10 million
annuallyuntiltheendofthecontract.
82

The expected milestone payments and availability payments that the Dark Fiber P3
Concessionaireneedstomakeitsbusinesscaseworkarethefollowing:
Table19:EstimatedDarkFiberP3MilestonePaymentandAvailabilityPayment

$1.087billion
DarkFiberMilestonePayment $1.358billion
$1.629billion
$27million/year
DarkFiberAvailabilityPayment $36million/year
$45million/year
AsTable19illustrates,a$1.36billionmilestonepaymentatsubstantialcompletion(i.e.,paidin
years 2 and 3), in combination with annual availability payments of $36 million for 30 years,
would cover all of the project costs(including the subsidy)under the base case assumptions.
Becausetheseresultsareuncertain(whichistypicalforthisstageofprojectdevelopment),the
sensitivityanalysisshowsthelikelyrangeofrequiredpayments,basedonthemainuncertainties.
Under the most conservative, worstcase scenario, a milestone payment of $1.63 billion in
combinationwithannualavailabilitypaymentsof$45millionwouldcoveralloftheprojectcosts.
The Project Team developed the twopart payment approach (i.e., a milestone payment in
combinationwithanavailabilitypayment)tooptimizetheprojectsfinancingstructure.Covering
alloftheprojectcostswithoutthemilestonepayment,forexample,wouldrequireanannual
availabilitypaymentof$129millionover30yearsunderthebasecasescenario.Underthemost
conservative, worstcase scenario, the required annual availability payment would be $173
million.Inbothofthesescenarios,theCitywouldfacesignificantlyhigherfinancingcosts.

83
7.3.3.1.2 Lit Fiber P3

Operating cash flow of the Lit Fiber P3 starts in year 2 of the project. In the first two years,
milestone payments are received, financing 50 percent of the capital expenditures of this P3
contract.Availabilitypaymentsstartinyear2,rampinguptothefull$62millioninyear3,then
continuingthroughouttherestofthecontractperiod.Operatingcostsrampupinyear1and2
ofthecontract(year2and3oftheprojectperiod)andstabilizeat$21millionuntiltheendof
thecontractperiod.

ThefinancingcashflowoftheLitFiberP3ischaracterizedbytwoyearsofdrawdownsofequity
anddebt,withminorinterestpaymentsforseniordebt,followedbyfiveyearsofoutflowsinthe
formofinterest,debtrepayment,anddividends.Inyear1ofthecontract,$101millioninsenior
debtand$40millioninequityaredrawndowntofinancecapitalexpenditures.Thesepayments
peakinyear4ofthecontractat$76million,followedbythreeyearsofcombinedpaymentsof
about$38million.
84
TheexpectedmilestonepaymentsandavailabilitypaymentsthattheLitFiberP3Concessionaire
needstomakeitsbusinesscaseworkarethefollowing:
Table20:EstimatedLitFiberP3MilestonePaymentandAvailabilityPayment

$123million
LitFiberMilestonePayment $147million
$176million
$42million/year
LitFiberAvailabilityPayment $62million/year
$76million/year
7.3.3.1.3 Public Broadband Entity

ThePBEsoperatingcashflowismainlydefinedbytherevenuesfromConnectionFeesandthe
paymentstothetwoP3s.ThemajorcashoutflowsarethetwomilestonepaymentstotheDark
FiberP3inthesecondandthirdyearafterthestartoftheproject.Thesecausenetrevenueto
benegativeuntilyear4ofthecontract,withthemostnegativevaluereaching$712millionin
year2.Netrevenuestabilizesaround$77millionperyearinyear9oftheproject.

PBEsfinancingcashflowshowsmajorcashinflowsinyear2and3ofthecontractof$713million
and$675million,respectively.Thisinflowistheseniordebtdrawdownthatisrequiredforthe
milestonepaymentstotheP3contractsinthoseyears.Fromyear4oftheprojectonwards,the
financingcashflowis$85million,comprisedofinterestandfeesfortheseniordebtaswellas
repaymentofthisdebtfacility.
85

ThePBEneedstochargethefollowingConnectionFeesinordertomakeitsbusinesscasework:
Table21:EstimatedConnectionFees(BaseCaseandBestCase)

ConnectionFeeresidential $51/passing/month $26$67/passing/month
ConnectionFeebusiness(average) $73/passing/month $38$97/passing/month
The range of potential Connection Fees is relatively wide, due to a series of uncertainties, as
furtherdiscussedinthenextsection.
7.3.3.2 Sensitivity Analysis

Thissectionprovidesanoverviewofthesensitivityanalysis,reflectingthemainuncertaintiesin
theproject.TheindicatorforwhichthesensitivitiesareshownistheresidentialConnectionFee,
whichis$51perpassingpermonthunderthebasecase.Worstcaseisdefinedastheassumptions
leadingtoaConnectionFeethatishigherthanunderthebasecaseassumptions;bestcaseis
defined as the assumptions leading to a lower Connection Fee than under the base case
assumptions.
86
Table22:SensitivityAnalysisofDualP3ResidentialConnectionFees
WorstCase BestCase
Variable Sensitivitydescription Connection Connection
Fee Fee
Capex Thecapitalcostestimatesareuncertain,duetothe $59 $42
currentstageoftheprojectpreparationand
design.Thevariabilityofthecostestimatesis+/
20percent,meaningthattheworstcaseis20
percenthigherthantheexpectedvalue(DarkFiber
highcostestimateandLitFiberbasecasecost
estimate)andthebestcaseis20percentlower
thantheexpectedvalue.
Opex Theestimatesofoperationalandmaintenance $53 $48
costsareuncertain,duetothecurrentstageofthe
projectpreparationanddesign.Thevariabilityof
thecostestimatesis+/20percent,meaningthat
theworstcaseis20percenthigherthanthe
expectedvalue(DarkFiberhighcasecostestimate
andLitFiberbasecasecostestimate)andthebest
caseis20percentlowerthantheexpectedvalue.
LitFiber ThebasecasecontracttermfortheLitFiberP3 $51 $47
contract concessionissevenyears.Thisisalsotheworst
term case,whereasthebestcaseis10years.
Subsidy Underthebasecase,lowincomeresidents(15 $52 $49
percentofthetotalpopulation)willpaya$10
ConnectionFee.Underaminimumsubsidy
scenario,lowincomeresidentswillpaya$20
ConnectionFee;underthemaximumscenario,low
incomeresidentswillpaynothing.
Additional Thebasecasedoesnotassumeanyrevenuesfrom $51 $44
revenues theCityoranyadditionalrevenues.Thisisalsothe
worstcase.Inthebestcase,theannualadditional
revenuesare$25million.Notethatthisscenariois
anillustrationonly;itisnotbasedonanyresearch
andishighlyuncertain.Furtherprojectpreparation
andthepreprocurementleasingprocesswill
reducetheuncertaintyandareintendedtolockin
additionalrevenues.
87
WorstCase BestCase
Variable Sensitivitydescription Connection Connection
Fee Fee
Financing Thefinancingcostsaredeterminedbyafewkey $54 $47
financingconditions.Theworstcase/basecase/
bestcasevaluesforthesefinancingconditionsare
asfollows:
Gearing:85percent/90percent/91percent
DSCR:1.25/1.20/1.15
Interest(commercialdebt):6.5percent/5.5
percent/4.5percent
Interest(revenuebond):4.5percent/3.5
percent/2.5percent
ROE:15percent/13percent/11percent
Thesensitivitiesarecalculatedforthecombination
ofminimumandmaximumvalues.
All Thissensitivityshowsthecombinationofallworst $67 $26
combined caseassumptionsandallbestcaseassumptions
combined.
TheoverallrangeoftheresidentialConnectionFeeis$26to$67foracombinationofallthebest
caseandworstcaseassumptions.Thecapexuncertaintiesarethemostsignificant,followedby
thefinancinguncertaintiesandthesubsidyforlowincomeresidents.Furtherprojectpreparation
willcontributetoreducingtheseuncertainties.
7.4 Structuring the Lit Fiber P3 to Achieve City Goals

IntheDualP3Model,theLitFiberP3shouldbestructuredtoservetheCitysdefinedPrinciples
andgoals.Amongothermatters,theinitiativemustdeliverabaselevelofbroadbandserviceto
everySanFranciscohouseholdandbusiness;definingwhatthatbaselevelofserviceshouldbe
isthekeychallengeofstructuringtheLitFiberP3.92
Inevaluatingapproachestoachievingthegoalofdeliveringabaselevelservice,weconsidered
fouroptionsfortheLitFiberP3thatreflectarangeofcostsandoutcomes:
Option1:LitFiberP3thatwoulddeliverlowlevelinternetservicetoall
92
TherequirementforabaselevelofservicesisbasedonguidancefromtheCityAttorneysofficethatuseofa
propertyrelatedfeefortheConnectionFeerequiresprovisionofaserviceintoeveryhomeandbusiness.The
propertyrelatedfeeandotherpotentialfundingmechanismsarediscussedintheappendices.
88
Option2:LitFiberP3thatwoulddeliverrobust,higherspeedinternetservicetoall
Option3:LitFiberP3thatwoulddeliverrobust,higherspeedinternetserviceto
lowincomeresidents
Option4:LitFiberP3thatwoulddeliveruniversalaccesstoahighspeedCivic
Network,withinternetprovidednotbytheLitFiberConcessionairebutratherby
ISPs
TheCitysgoalsforopennessandcompetitionarebestservedbyOption4whichwouldinclude
robustaccesstoahighspeedCivicNetworkforeveryhouseholdandbusinessintheCity.Atthe
sametime,theLitFiberConcessionairewouldprovidewholesaleaccesstocompetingISPsthat
would make possible robust competition with very low startup costs for ISPs that choose to
provideretailservicesoverthenetworklitupbytheLitFiberConcessionaire.Thisapproachis
newanduntestedinaU.S.citythesizeofSanFrancisco,buthasthepotentialtoachievethe
Cityskeygoals(includingequity,competition,civicopportunity,andinnovation)whileavoiding
someofthesignificantpitfallsandfinancialfailingsoftheotherapproaches.
Inthesectionsbelow,wedescribetheprosandconsofeachoption,andtheanalysisthatledto
ourrecommendation.
7.4.1 Option 1: LowLevel Internet Service to All

ThefirstoptionweconsideredwouldrequiretheLitFiberConcessionairetodeliveralowlevel
internetservice(i.e.,atalevellowerthanBLAsrecommended25Mbpsdownstream,3Mbps
upstream) to every home and business in the City. Achieving equity through ubiquity, this
approachmakesadirectconnectionbetweentheConnectionFeeandservice;everyproperty
ownerwouldpayafee,andeveryhouseholdandbusinesswouldreceiveservice.
ThisapproachpresentstwobenefitsintermsoftheCitysgoalsformarketbasedcompetition.
First,thelitnetworkwouldserveasaplatformfortheLitFiberConcessionairetosellwholesale
servicestocompetingretailISPs.Second,becausealowlevelinternetservicewouldbeslower
thanmanyhigherincomeresidentscurrentservices,itwouldbeunlikelytosatisfythoseusers
andthuswouldbeunlikelytotakesignificantmarketsharefromthecompetingISPs.
Thosecompetitivebenefits,however,arelargelyovershadowedbyseveralnegativeaspectsof
deliveringalowlevelinternetservicetoeveryhomeandbusiness.First,giventhehighcostof
deployingtheproposedbroadbandinfrastructure,thisoptionwouldbeanexpensivemechanism
fordeliveringalargelyinconsequentialserviceformanyresidents.Second,manyofthelower
incomeresidentswhomightbenefitfromthisservicemayalreadybeeligibleforlowcost,low
bandwidth programs such as Comcasts Internet Essentials programso this option would
duplicateanexistingservice.Finally,becausetheservicewouldbelowlevel,itwouldnotdeliver
89
thetransformativebenefitsoftruebroadbandtotheverycitizens,thoseoflowerincome,that
theBroadbandforSanFranciscoInitiativeisdesignedtosupport.
7.4.2 Option 2: Robust HighSpeed Internet Service to All

RecognizingthatthelowlevelserviceconsideredinOption1wouldnotdeliverthebenefitsof
truebroadbandtotheCityslowincomeresidents,wenextconsidereddeliveringrobustinternet
servicetoeveryhomeandbusiness.Inthisscenario,theLitFiberConcessionairewouldserve
eachresidentandbusinesswithatleast25/3service,thelevelofservicerecommendedbythe
BudgetandLegislativeAnalystandthatmeetstheFCCsdefinitionofbroadband.
Like Option 1, this option would deliver a service to every resident and businessachieving
equityandmakingthecaseforapropertyrelatedfee.Byincreasingtheminimumspeedofthe
baselevel service, this option would provide true broadband to lowincome residents. The
baselevel service would be an improvement over Comcasts and AT&Ts lowcost, low
bandwidthprograms,ratherthanaduplication.
But delivering robust internet service to every resident and business would have negative
consequencesforthecompetitivemarket.Becauseitwouldberobust,thisbaselevelservice
wouldlikelysatisfymanyuserswhowouldthenbeunlikelytopayforhigherlevelservicefrom
the competing ISPs that lease access on the platform lit by the Lit Fiber Concessionaire. By
eroding the potential market for ISPs, this option would dramatically change the economic
prospectsforthenetwork.
7.4.3 Option 3: Robust, HighSpeed Internet Service to LowIncome Residents

InouranalysisofOption2,wefoundthatprovidingarobustinternetservicetoallpassingswould
achieve equity in terms of the benefits of true broadband, but would make much more
challenging the projects potential to enable new competition and generate related income.
Drawingadistinctionbetweenresidentswhocanaffordtobuyrobustbroadbandandthosewho
cannot,wethenlookedattheeffectsofdeliveringrobustinternetasabaselevelserviceonlyto
lowincomeresidents.HigherincomeresidentsandbusinesspassingswouldpaytheConnection
Feebutwouldnotreceiveabaselevelservice.
ThisoptionamelioratessomeoftheshortcomingsofOption1andOption2.Mostimportantly,
intermsofequity,lowerincomeresidentswouldreceivethekindofhighbandwidthbroadband
service that is prioritized by the Broadband for San Francisco Initiative, and the benefits that
comewiththatlevelofservice.Andintermsofthebusinessmodel,thepotentialmarketfor
competingretailISPswouldnotbediminished,becausehigherincomeresidentswouldnotbe
receivingaservice.Thatsaid,Option3hassomesignificantshortcomings.LikeOption1,itwould
representanextremelyexpensivemeansofdeliveringservicetolowerincomeresidents.And
becausehigherincomeresidentsandbusinesseswouldnotbereceivinganybaselevelservice,
90
thesourcesoffundingfortheConnectionFeemightbereducedbecausenotallpayersofthe
Feearereceivingadirectservice.
7.4.4 Option 4: Robust Civic Network Service to All

InourP3analysis,wedeterminedthatseparatingthenetworklayers(darkfiberinfrastructurein
one P3, lit service in another) would solve some of the fundamental issues in terms of the
Broadband for San Francisco Initiatives risk profiles and financial projections. Similarly, this
optionsolvesthepuzzleofprovidingabaselevelservicebytreatinginfrastructureandinternet
serviceasseparateissues.
Inthisoption,asinthefirstthreeoptionsabove,aDarkFiberConcessionairewouldconstruct
infrastructuretopasseveryresidenceandbusinessintheCity.ALitFiberConcessionairewould
lightthefiberandsellwholesaleaccesstocompetingISPs.
Wherethisoptiondivergesfromthoseotherapproachesisinthedefinitionofthebaselevel
service.Ratherthanchoosinganinternetservicelevel,thisoptionenvisionsdelivering1Gbps
intranetaccesstoeveryhomeand businessandtherebycreatingaCivicNetworkthatwould
connectallSanFranciscoresidentsandbusinessestocivic,educational,healthcare,andnon
profit resources. Residents and businesses can also choose to purchase retail services (from
competing providers that would buy wholesale service from the Lit Fiber Concessionaire), in
additiontothefree,highbandwidthaccesstotheCivicNetwork.
Toaddresstheissuesofaffordabilityandequity,theCitycouldprovideasubsidytolowincome
residentstopurchaseretailinternetservicethoughwenotethatservicefeesmaybeverylow
underthisscenariobecausethetechnicalconfigurationwouldmakeiteasyandcheapformany
retailserviceproviderstocompeteoverthenetwork.
ThisoptioncheckstheboxesonboththeCitysprioritiesandtheprimarybenefitsachievedin
the other options. It would ensure ubiquitous access and, thus, would deliver value for the
Connection Fee. It would also make the dark fiber infrastructure a platform the Lit Fiber
Concessionaire to sell wholesale services to competing retail service providersand would
maintainthemarketforthoseproviders.
7.4.5 Technical Description of the Potential Civic Network

IntheOption4scenario,theLitFiberConcessionairewoulddesignandoperateaCivicNetwork
which will be made available at no extra charge to subscribers connected to the Citys FTTP
network.93ThisservicewillprovideaccesstocertainresourcestobemadeavailablebytheCity
aswellasaselfserviceportal,tobeoperatedbytheLitFiberConcessionaire,whichwillallow
93
AdetailedtechnicaldescriptionofthefunctioningoftheCivicNetworkandthepotentialtoenablerobustopen
accessandcompetitionoverthenetworkisincludedintheappendicestothisreport.
91
subscriberstosignupforservicesofferedbyRSPs.AccesstotheCitysCivicNetworkshouldbe
provisionedtoasubscriberportviaanEVCinthesamewaythatretailservicesareprovisioned.
TheselfserviceportalwillactasavirtualstorefrontforservicesbeingofferedontheCityFTTP
network.TheLitFiberConcessionairewillberesponsiblefordevelopingandoperatingabackend
servicewhichwillallowRSPstolisttheiravailableservicesandpricingandafrontendservice
whichwillallowsubscriberstoselectthoseservices.Dependingontheservicerequested,the
portalshouldeithertriggeranautomatedprocesswhichwillprovisiontherequestedserviceto
thesubscriberportandalerttheRSPtobeingbilling,orshouldautomaticallycreateaservice
requestwhichwillbesenttotheLitFiberConcessionaireNOCandtheRSPsoprovisioningand
billingservicesandbeperformedbytherespectiveentities.
TheCitysCivicNetworkwillbemadeavailabletoallsubscribersontheFTTPnetworkregardless
of whether they subscribe to retail services. Because different EVCs on the network are
separated,withtrafficforeachEVCbeingroutedbyeachRSPortheCivicNetwork,asubscriber
receivingtwoormoreserviceEVCswouldneedtohaveeachservicedeliveredonaseparate
physicalportorondifferentVLANsoverthesameport.Thismaybeareasonablesituationfor
subscriberswhoruntheirownbusinessclassroutersandswitches,butwouldbetoocomplex
formostresidentialsubscribersandmanybusinesssubscribers.
InordertoprovideeasyaccesstotheCitysCivicNetworkforsubscriberswhoalsosubscribeto
aretailservice,theagreementwithRSPsshouldrequireanocostinterconnectionwiththeCitys
CivicNetwork.ThiswillallowsubscribertrafficdestinedfortheCitysCivicNetworktoberouted
overtheRSPsnetwork.(SeeFigure15.)
Figure15:FTTPSubscribersCanAccessCivicNetworkDirectlyorviaanRSP'sNetwork
92
8 Single Dark Fiber P3 Approach

ThisoptionenvisionsamarkedlydifferentapproachtotheentireBroadbandforSanFrancisco
Initiative.
IncontrasttotheDualP3Model,inthisalternativeapproachtheCitywouldenterintoasingle
P3fordarkfiberonly(SingleDarkFiberP3)andplaynofinancialoroperationalroleinhowthe
fiber is lit or services provided over it. Rather, the City would focus its efforts on enabling
competition at the dark fiber layer. It would execute a Dark Fiber concession and the Public
BroadbandEnterpriseortheConcessionairewouldthenleasedarkfiberonawholesaleaccess
tocompetingISPs.TheISPswouldlightthefiberandsellretailservicestoanyinterestedresidents
orbusinesses.
8.1 Benefits
Solongasamarketemergesfordarkfiberleasesthatwouldenablenewcompetition,aSingle
DarkFiberP3Modeloffersgreatersimplicity,lowercost,andpotentiallyreducedcontroversy.
8.1.1 Reduced Complexity and Cost

WithaSingleDarkFiberP3,theCitywouldspenddramaticallylessbecauseitwouldnottakeon
thefinancialcommitmentnecessaryforaLitFiberConcession.Asthefinancialanalysisforthe
DualP3Modeldemonstrates,theConnectionFeefortheLitFiberP3isconsiderable.94Ifthat
componentisremovedandtheConnectionFeeislimitedtotheDarkFiberP3only,therevenue
necessarytosupportthemodelisgreatlyreduced.
DevelopingaSingleDarkFiberP3onlyalsoreducescomplexityandadministrativeburdenforthe
City.Unlikelitservicesandequipment,darkfiberisastable,predictableasset.Itisaninfinitely
scalable,longlife,futureproofassetthatwillcontinuetobeusableandusefulover30yearsor
moresolongasitisadequatelymaintained.Andfiberisuniversallyacknowledgedasthemost
robustandmostscalableofcommunicationsmedia,enablingscalingasnecessaryasbandwidth
needsgrowandessentialtosupportfuturegenerationsofwirelesstechnology,bothmobileand
fixed.Litservices,incontrast,changedramaticallyovertime,inresponsetomarketchangesand
serviceinnovations.Andtheequipmenttolightupthoseserviceswillalsochangeasthemarket
andbandwidthdemandsgrow.Broadbandequipmentmustalsobereplaced,dependingonthe
component,approximatelyeverysevento10years.AdministeringandenforcingaLitFiberP3
withalltheseelementsisanadministrativechallengethatisalleviatedbyeliminatingtheLitFiber
P3entirelyandleavingalltheseelementstothemarket.
The model also removes from the Citys purview network operations, where a considerable
amount of risk lies in managing technological and customer service aspects of the network.
94
TheConnectionFeeanalysisisdetailedinSection7.
93
RatherthandeterminingparametersandtermsfortheseareasofoperationthroughtheLitFiber
PG, the Citys role would be limited to the simpler dark fiber layer of the network. The
Concessionairealsowouldhavearolelimitedtobuildingandmaintainingdarkfiber,andthere
would be no need for a Concessionaire to deal with the more challenging areas of network
performance,servicedelivery,andcustomerrelationships.
8.1.2 Removes the City from the internet business or competing with private
companies
TheSingleDarkFiberP3ModelalsopotentiallyreducesthechallengesinherentintheCityswork
to promote competition and better broadband. Regardless of merits, City efforts to expand
broadbandandbroadbandcompetitionhavebeenaggressivelyopposedbyincumbentphone
andcablecompaniesnationwide.Inadarkfibermodel,theCity(anditsConcessionaire)playno
roleinprovidingservicestothepublic,butratherprovideadarkfiberplatformtoenablethe
private sector to enter the service business. Frankly, incumbents are likely to oppose this
initiativeregardlessofhowlimitedtheCitysroleis,butthecustomaryincumbentclaimofunfair
competitionfromthepublicsectorisundercutbythemodel.
Further, by removing the City from any role in shaping services, pricing, or performance
standards,theSingleDarkFiberP3preservesthefullretailmarketforcompetingISPsrather
thanusingaP3todeliversomelevelofservicestothepublic.
8.2 Challenges
Though simple and efficient, the Single Dark Fiber P3 Model reduces options with regard to
fundingandtocompetition.
8.2.1 Funding options

ThisapproachpotentiallyreducestherangeofoptionsbeforetheCityforfunding.Inparticular,
withoutaservicedeliveredtoeveryhomeandbusiness,apropertyrelatedfeecannotbeused
fortheConnectionFeetofundtheproject.95EliminatingtheLitFiberP3isefficient,butnarrows
theCityspathforward.
8.2.2 Competition considerations

ThecompetitionconsiderationsarecomplexwithrespecttocomparisonbetweentheSingleDark
Fiber P3 and the Dual P3 models. While the Single Dark Fiber P3 creates unprecedented
opportunityatthedarkfiberlayer,italsodoesnotdelivercompetitionaseffectivelyasaLitFiber
3potentiallycan.
ThoughaccesstodarkfibergreatlyreducesthebarriertoenteringtheSanFranciscomarketfor
newproviders,thereisstillconsiderablecostandcomplexityinlightingandoperatinganoptical
95
Arangeofpotentialfundingsourcesisanalyzedintheappendicestothisreport.
94
network over dark fiber. Only a sophisticated, wellfunded company will have the scale and
capacitytodoso.Giventhecostinvolvedandthepotentialcompetitioninthemarketcreated
by the dark fiber itself, its not likely that more than a small handful of entities would be
interestedinleasingdarkfiberinthenearterm.Indeed,themarket,dependingonhowthefiber
ispriced,maysupportonlyoneortwolesseesfortheforeseeablefuture.Overthelongterm,
however, market demand and structures may change and new opportunities for dark fiber
competitionmayarise.96
Incontrast,alitnetworkcanprovideveryinexpensiveopenaccesstomanyserviceproviders,
includingverysmalloperationswithmodestresources,atlowincrementalcost.97Lightingthe
networkinthefirstplaceiscostly,buttheplatformcanthenaddadditionalcompetingproviders
quicklyandcheaply.
Theunknownsregardingthedarkfibermarketalsoraisethequestionofwhetherthemodelwill
ensuresufficientcompetitionemergetosupporttheCitysaffordabilityanddigitalequitygoals.
Inthisscenario,theCitywouldnotbeabletousetheLitFiberP3toitselfprovidelowercost
products for lowincome residents. Rather, the City could provide a subsidy to lowincome
residentstoensuretheycanaffordretailservices.
8.3 Financial Analysis

TounderstandthepotentialConnectionFeenecessarytosupporttheSingleDarkFiberP3Model,
the Project Team developed a version of the financial model for dark fiber only. The model
suggeststhatthenecessaryConnectionFee,assumingnodarkfiberrevenues,willbeintherange
of$15to$37perresidentialpassingpermonth,withabasecasevalueof$28.
TheProjectTeamnextmodeledarangeofscenariostounderstandhowdarkfiberrevenuesmay
reducetheConnectionFeenecessary.Giventhatadarkfibermarketofthistype(connecting
everyhomeandbusiness)hasneverbeenattemptedinacityofthesizeofSanFrancisco,we
analyzeddatafromtwootherprojectsinsmallercitieswherethismodelhasbeenpioneeredto
understandthefinancialimplicationsforSanFranciscooftherevenuesgeneratedinthosecities.
The City of Westminster, Maryland, for example, has a 20yearcontract with Ting Internet in
which the city is obligated to finance and build all of the outside plant (including drops to
customerspremises).Tingisleasingfiberwithatwotieredleasepayment.Onemonthlyfeeis
96
InHuntsville,Alabama,wherethecitysmunicipalutilityownsafibernetworkwiththisbusinessmodel,Google
Fiberhasthusfarbeentheonlyentitytoleasefiberfromtheutility,eventhoughothersmaycanleasebasedon
HuntsvilleUtilitiesestablishedrates.Overtime,therewillpotentiallybeotherISPusersofthecitysfiber,
particularlytoservelargerbusinessesandinstitutions,thoughwequestionwhethertheeconomicsexistfor
anotherprovidertocompeteagainstGoogleFiberintheresidentialmarket,asleastintheshortterm.
97
ThetechnicaldescriptionforlowcostopenaccessoveralitnetworkcanbefoundinSection7.
95
basedonthenumberofpremisesthefiberpasses($6permonthperpassing);thesecondfeeis
basedonthenumberofsubscribersTingenrolls($17permonthpersubscriber).98
Inanothervariationonthismodel,Huntsville,Alabamasmunicipalelectricutilityisbuildinga
fibernetworkandGoogleFiberisleasingfiberthroughoutthecitytoserveresidencesandsmall
businesses.99GoogleFibersleaseisfor20yearsbasedonaratesheetthatprovidesforvarious
levelsofpricingbasedonamountsandvolume.Thepricingstructureiscomplexbutweestimate
GoogleFibersmonthlycosttobeapproximately$6to$7permonthperpassing.100
The financial model suggests that applying the Westminster and Huntsville pricing to San
FranciscowouldbringtheexpectedConnectionFeedownfrom$28permonthwithoutdarkfiber
revenuesto$20to$26permonth,respectively.
Thefollowingsensitivityanalysisreflectsthemainuncertaintiesintheproject.Theindicatorfor
whichthesensitivitiesareshownistheresidentialConnectionFee,whichis$28perpassingper
monthunderthescenariowithoutanyadditionalfees.Worstcaseisdefinedastheassumptions
leadingtoaConnectionFeethatishigherthanunderthebasecaseassumptions;bestcaseis
defined as the assumptions leading to a lower Connection Fee than under the base case
assumptions.
98
BecauseTingwillpayWestminsteramonthlyfeeof$6foreveryhomeandbusinesspassed,Tingisfinancially
obligatedtothecityfromdayone,evenifithasnocustomers.ThisstructuregivesthecityconfidencethatTingis
highlyincentedtosellservicestocovercosts.Tingwillalsopaythecity$17permonthforeachcustomeritserves.
Inlateryears,whenTingsrevenuehitscertainthresholds,Tingwillpaythecityasmallfractionofitsrevenueper
user.Thatmechanismisdesignedtoallowthecitytoshareinsomeoftheupsideofthenetworkssuccess.Inother
words, the city will receive a bit of entrepreneurial reward based on the entrepreneurial risk the city is taking.
Perhapsmostsignificantly,thereisalsoamechanismbuiltintothecontractthatensuresthatthetwopartiesare
trulysharingriskaroundthefinancingoftheoutsideplantinfrastructure.InanyquarterinwhichTingsfinancial
obligationstothecityareinsufficienttomeetthecitysdebtservice,Tingwillpaythecity50percentoftheshortfall.
Insubsequentquarters,ifTingsfeestothecityexceedthedebtservicerequirements,Tingwillbereimbursedan
equivalent amount. This element of the financial relationship made the deal much more attractive to the city
becauseitisacleardemonstrationofthefactthatitsprivatepartnerisinvestedwithit.
99
Huntsvilleisnotfinancingorbuildingthedropstothehome;itsobligationtoGoogleistoprovidefiberthat
passesthepremisesonly.
100
Incontrast,TingsobligationstoWestminsterarebasedinpartonhowmuchfiberitusesandinpartonhow
manycustomersitsecuresandrevenuesitgenerates.Asaresult,Westminsterwillhavelesspredictabilityand
certaintyaboutitsrevenuesfromTing,buthasthepotentialtoshareinupsideintheeventthatTingisvery
successfulinthatmarket.
96
Table23:SensitivityAnalysisofSingleDarkFiberP3ResidentialConnectionFees
WorstCase BestCase
Variable SensitivityDescription Connection Connection
Fee Fee
highcostestimate)andthebestcaseis20percent
lowerthantheexpectedvalue.

expectedvalue(DarkFiberhighcasecostestimate)
andthebestcaseis20percentlowerthanthe
expectedvalue.

ConnectionFee,andthosewhobuyinternet
service(expectedtobe80percentofthelow
incomeresidents)receiveanadditional$30subsidy
permonth.Underaminimumsubsidyscenario,
lowincomeresidentswillpaya$20Connection
Feeandtheadditionalinternetservicesubsidyis
$20;underthemaximumscenario,lowincome
residentswillpaynothingandtheadditional
internetservicesubsidyis$40.

anillustrationonly;itisnotbasedonanyresearch
andishighlyuncertain.Furtherprojectpreparation
97
WorstCase BestCase
Fee Fee
andthepreprocurementauctionwillreducethe
uncertaintyandareintendedtolockinadditional
revenues.

asfollows:
DSCR:1.25/1.20/1.15
percent/4.5percent
percent/2.5percent

combined.
The residential Connection Fee in the Single Dark Fiber P3 Model is $20 per month under a
scenarioassumingamonthlyperpassingfeeof$6andamonthlypersubscriberfeeof$6(ata
take rate of 50 percent for residential customers and 30 percent for business customers) in
additiontotheConnectionFee.Thetablebelowshowsthemainsensitivitiesunderthisscenario.
98
Table24:SensitivityAnalysisofSingleDarkFiberP3ResidentialConnectionFeeswithAdditionalFees
WorstCase BestCase
Fee Fee
highcostestimate)andthebestcaseis20percent
lowerthantheexpectedvalue.

expectedvalue(DarkFiberhighcasecostestimate)
andthebestcaseis20percentlowerthanthe
expectedvalue.

ConnectionFee,andthosewhobuyinternet
service(expectedtobe80percentofthelow
incomeresidents)receiveanadditional$30subsidy
permonth.Underaminimumsubsidyscenario,
lowincomeresidentswillpaya$20Connection
Feeandtheadditionalinternetservicesubsidyis
$20;underthemaximumscenario,lowincome
residentswillpaynothingandtheadditional
internetservicesubsidyis$40.

anillustration;itisnotbasedonanyresearchand
ishighlyuncertain.Furtherprojectpreparationand
99
WorstCase BestCase
Fee Fee
thepreprocurementauctionwillreducethe
uncertaintyandareintendedtolockinadditional
revenues.

asfollows:
DSCR:1.25/1.25/1.15
percent/4.5percent
percent/2.5percent
Takerate Inthebasecase,atakerateof50percentfor $21 $19

residentialcustomersand30percentforbusiness
customersisassumed.Inthebestcaseassumption,
thetakerateisincreasedto60percent
(residential)and40percent(businesscustomers).
Intheworstcaseassumption,thetakerateis
reducedto30percent(residential)and20percent
(businesscustomers).

combined.
Thesensitivityanalysisdemonstratesthatagoodstructuringoftheperpassingandpercustomer
fees in combination with upfront commitment of additional revenues can help reduce the
ConnectionFeesignificantly.
100
9 Recommendations for Implementation and Procurement Strategy

Thissectionofthereportconcernstheprocurementstrategyfortherecommendedapproachfor
theBroadbandforSanFranciscoInitiative.Insummary,thissectiondescribesthewaysinwhich
thispioneeringfibertothepremisesP3procurementwillbethefirstofitskindandthat,given
itsuniquenessandpioneeringnature,aprocurementstrategycustomizedforapplicationofP3
deliverymechanismstothenewP3areaofFTTPisnecessary.
We therefore suggest a procurement process grounded in the traditional P3 process but

customizedandexpandedtoaccountfortheuniqueaspectsofthiseffort,including(1)funding
authorizationrequirements,(2)uniquemarketcharacteristics,and(3)thelackoffamiliarityof
theinvestmentcommunitywithP3sinthebroadbandsector.
Werecommendthefollowingstepsintheprocurementprocess:
1. Industry Sounding. A preprocurement industry sounding to clarify and market the

projectandproposedP3structuretogeneratemarketappetiteaswellassolicitinputsto
furtherrefinethestructure.Theindustrysoundingcouldincludeafulldayeventorone
ononeconversations.
2. PreProcurementDarkFiberLeasing.Apreprocurementpreleaseorsaleofdarkfiber
andothernetworkcapacityinadvanceoftheP3RFPssoastoquantifyandsecuresome
revenues that would otherwise be speculative; this will enable (1) bids with lower
requirementsforCityfunding/guarantees,and(2)amorecompetitivebiddingprocessin
whichallbidderscanaccountfortheserevenuestreams.
3. P3 Procurement. In the event that the Dual P3 Model is adopted, a P3 procurement

processthatdistinguishestwoprocurements:onefordarkfiberandotheroutsideplant
infrastructure(DarkFiberRFP);andaseparateonefortheserviceprovider(LitFiberRFP),
toberunpartiallyinparallel.
4. IntegrationofRFPDevelopmentandIndustryConsultation/Negotiations.Althoughthe
CityhasbeencontemplatinganRFPdevelopmentprocessfollowedbynegotiationswith
one or more selected bidders, we recommend integrating these phases and building
consultationwithselectedbiddersintoamoreextensiveprocessofRFPdevelopmentand
vetting.ThisprocesswhichshouldbemoreextensivethaninotherP3sectors,giventhe
novelty of this effortwill enable the necessary level of interaction with potential
bidders,givingtheCityopportunitytorefinethesecomplexprocurements,whileenabling
thebidderstofurtherunderstandandcraftbidsforthisnew,unfamiliarassetclass.
101
Overall,theprocurementprocessshould:
Beginasearlyaspossibleinordertogatherindustryfeedbackbeforeseekingfunding
approval.Akeytaskforthisprojectistoensureamatchbetweentheleveloffunding
approved and what bidders will propose during the procurement process. In an ideal
situation,bidderswouldbeselectedandafirmunderstandingoftheleveloffinancial
commitment needed would be clear before funding authorization is sought. However,
bidderstendtorequirefundingapprovalinadvanceofsubmittingcomplexbidsandare
lesswillingtoputsignificanteffortintocompletingbidsforprojectsthattheyhavereason
todoubtwillbefunded.Therefore,startingengagementwithpotentialbiddersthrough
anindustrysounding,andinitiatingtheRFQandRFPprocessasearlyaspossibleinorder
toobtainsomelevelofinteractionandinformationfrombidders,willhelpprovidesome
claritytotheapprovalprocessandprovidedecisionmakerswithasmuchinformationas
is feasible at that stage of the process, even if binding bids have not already been
submitted.
Build in sufficient time and flexibility. A procurement timeline that accounts for the
pioneering and singular nature of this effort and thus builds in sufficient time and
flexibilitytovetallelementsofthebidrequirements;anextendedtimelinewilllikelyalso
berequiredbytheemergingstrategyofseparatingtheDarkFiberRFPfromtheLitFiber
RFP.
ApplyP3procurementbestpractices.Eventhoughthisprocurementwillbethefirstof
itskind,severalP3procurementbestpracticescanbeapplied,including:
o Shortlistthreeorfourqualifiedbidders(RFQPhase):Giventhetimeandresources
required for the City and bidders, it is important that the City selectively identify
biddersforadvancement.Theshortlistingofqualifiedbiddersisconsideredstandard
practice for any complex procurement and is particularly important for P3
procurements.OnlyhavingasmallgroupofprequalifiedbiddersallowstheCityto
focusitsresourcesonrespondingtoasmallernumberofbidders.Fromthebidders
perspective,becausethecostofdevelopingadetailedproposalduringtheRFPstage
is significant, they too prefer to compete among a handful of shortlisted teams. A
smallerpoolofcompetitorsincreasestheprobabilityofsucceedingattheRFPstage,
whichmotivatesbidderstoinvestresourcesindevelopingahighqualitybid.
o Communicatewiththebidderspriortobidsubmission(DraftRFPPhase):TheCity
canuseamultistagedraftRFPprocesstoincreasedialoguewiththepotentialP3
concessionaires,withtheobjectiveofcraftinganagreementthatadequatelyreflects
acceptablelongtermtransferofrisksandresponsibilitiesbetweenthepartiesinthe
102
mostefficientwaypossible.Whendonewell,adraftRFPprocessisanorganizedway
tosolicitinputfromprequalifiedbiddersandincreasefaircompetition.AdraftP3
agreementcanbeincludedinthedraftRFPdocumentstogainearlyindustryinputon
theproposedcontractualagreement,andtolimitnegotiationswithasinglepreferred
bidder.Thisprocessdecreasesthelikelihoodofmajorlastminuteamendmentstothe
finalRFP,whichcoulddistortthecompetitiveprocessandincreasetransactioncosts.
Ideally,italsoreplacesthepostbidnegotiationswithoneormoreselectedbidders
thatarecontemplatedbytheCitysoriginalplan.Giventhepioneeringnatureofthis
effortandthenewnessofthissectortoP3bidders,multipleiterationsofthisdraft
RFPprocessmaybeadvisable,andthetimelineshouldreflecttheneedforflexibility
andcreativityduringtheRFPdevelopmentprocess.
Theseissuesarediscussedinfurtherdetailbelow.
9.1 A PreProcurement Capacity Leasing Process Should Precede the P3

Procurement so as to Improve P3 Bidding Outcomes by Accounting for
Some New Revenue Opportunity
Onewaytodemonstratethepotentialviabilityoftheopportunityinadvanceofafullfunding
decision, both to policymakers and to potential bidders, relates to the second primary
consideration of procurement for this initiativethat is, the nature of the infrastructure and
servicesathand.
9.1.1 FTTP Represents an Untested Area for P3 Delivery, in a Complex and Ever
Changing Sector
Unlike traditional P3 areas, broadband and FTTP exist in a somewhat competitive market
(althoughitisamarketofunevenlevelsofservicetovariouspartsoftheCity,andwithservices
lesscapableandlessrobustthantheFTTPinfrastructureenvisionedbythisinitiative).Inaddition,
theuseofthefiberandthewaysinwhichconnectivityandtheinternetwillbeusedoverthe
courseofthelifeofthisnetworkareinsubstantialwaysunpredictable.
UnlikeawaterutilityP3thatwill,initsfirstyearanditslastyear,providewateroratollroad
that will throughout the life of the concession provide access to a navigable roadthe FTTP
networkwillbeusedinthemediumandlongterminwaysthatarelargelyimpossibletoproject
atthecurrentmoment.Generally,wecananticipateonlysomeoftheusesofthenetworkout
toperhapssevento10years.Consider,forexample,thattheadventofthecommercialinternet
was only 22 years ago. Facebook was founded only 13 years ago. Netflix started offering
streamingcontentonly10yearsago.
103
Asaresult,itisnotonlydifficulttodeterminehowconsumerswillusethisinfrastructureand
servicesoveritinthefuture,butitisalsoextremelychallengingtoanticipatewhatcompanies
willbeinterestedinwithrespecttotheuseoftheCitysnewinfrastructure.
9.1.2 The Opportunity Presented by This Initiative Is Unfamiliar to the Investment

Community
Inaddition,theopportunitytheCityisabouttopresenttotheprivatesectorisunprecedented.
TherehasneverbeforeexistedinanyAmericancommunityanopportunityforaprivateentity
toleasefiberorbroadbandinfrastructuretoreach100percentofthehomesandbusinessesin
thecommunity.Whileahandfuloflocalitieshavesoughttoofferopenaccessnetworks,those
wereonasmallerscalethanSanFranciscoandonlyconnectedaportionofthecommunity(those
customersthatpaidforservice)and,eventhen,generallyovernetworkswithoutthetechnical
robustnessofwhatiscontemplatedforthisinitiative.Incontrast,whattheCityhaselectedto
examineforthisinitiativeisimmediateconnectivityto100percentofthehomesandbusinesses
inthecommunity.
Inaddition,therehasneverbeforeexistedamechanismforcompaniestoleasedarkfiberon
suchascale.Becausethephoneandcableincumbentsdonotallowaccesstotheirdistribution
networkstothehomeandbusinessatthedarkfiberlayer(and,indeed,atanylayerinmany
cases), private sector internet entrepreneurs and other companies have never had the
opportunitypresentedhere.
Forallthesereasons,itisdeeplychallengingtoanticipatethekindsoflesseesthatmayemerge
eveninthefirstfewyearsafterthenetworkisbuilt,letaloneinthemediumandlongterm.
In sum, the opportunity the City is presenting to the private sector is unprecedented, and
thereforethereexistnoempiricaldatabywhichtoprojectorunderstandwhatkindsofpotential
revenuesmayemerge.Inaddition,P3investorswillnotacceptthekindofuncertaintythatisa
necessaryelementofthepioneeringnatureoftheasset.Rather,theywillbidwitharequirement
thattheriskpresentedbythisuncertaintyisbornebytheCity.
9.1.3 A PreProcurement Lease of Access to the Network Would Help to Mitigate

These Uncertainties
Giventheconfluenceofallthesefactors,werecommendasubstantialadditionalphasetothis
P3 procurementone that would occur in advance of an RFP for a P3 bidder. This first
procurementwouldbeaimednotattheP3investmentcommunity,butatthepotentialusersof
thenetwork,whowouldbeinterestedinleasingcapacityonthenetworktoreachresidences
andbusinessesinSanFrancisco.
104
Theprocurementwouldbedesignedtoquantifyatleastsomeofthepotentialdarkfiberleasing
revenuesandtoeffectivelycreatebindingmechanismsandrevenuestreamsfromthesebidders
thatwouldthenbecomepartoftheRFPandthebusinesspropositionofferedtotheP3bidders.
In other words, a preliminary procurement of anchor customers would allow the Citys P3
procurementtoquantifyandguaranteeasubstantialsourceofrevenuetothebidderswhich
wouldreducetherequiredfinancialcontributiontobeaskedofthetaxpayers.
OutsidetheP3context,thiskindofstrategyhasbeenusedonamuchsmallerscaleinafew
communitiesthathavepresolddarkfiberleasesbeforeconstructionsoastobringinrevenues.
TheCitycouldenticedarkfiberlesseesbyofferingpreconstructionpricingakintotherealestate
environment,andinthisway,couldconvertatleastsomeofthenetworksrevenuepotential
fromspeculativetoguaranteed.
Afurtherbenefitofthisstrategyarisesfromthefactthatidentifyingandsecuringtheserevenues
inadvanceofaP3procurementwillresultinamorecompetitiveprocurement,asallthebidders
willbeabletobuildintotheirbusinessmodelsthesepotentialrevenues.Incontrast,ifthese
revenuesarestillspeculative,asinglebiddermaybeabletoquietlyenterintoanagreementwith
apotentiallesseeoffiber;thiswouldskewthecompetitiveprocessbecauseotherbidderswould
notbeabletofactorthoserevenuesintotheirbids,eventhoughthoserevenuesarelikelyto
materialize.
Thecombinationofallthesefactorsleadsustorecommendaprocurementprocessthatincludes
contingentdarkfiberleasinginadvanceoftheP3procurement.
9.1.4 A PreProcurement Industry Sounding Should Be Used to Market the Project

and Solicit Inputs from the Private Sector
Due to the unique nature of this project, including an ambitious project scope, competitive
marketlandscape,andpotentially,adualP3structure,anindustryengagementevent(industry
sounding) is highly recommended. Industry sounding is used as a forum for providing
informationabouttheprojecttopotentialbidders,marketingtheproject,andsolicitinginput
frombidders.
9.2 The Procurement Timeline Should Be Developed in Light of the

Complexities and Newness of the Strategy
Theprocurementtimelineshouldaccountforthepioneeringandsingularnatureofthiseffort
and build in sufficient time and flexibility to vet and negotiate various elements of the bid
requirements.WeurgetheCitytobuildaprudentandriskmitigatingprocurementtimelineso
as to enable multiple stages of vetting of various elements of the bid documentation with
potentialbidders,aswellasexpertcommentontheeffort.
105
The newness and innovative nature of the Citys combination of technical scope (open,
ubiquitous FTTP) and delivery mechanism (P3) creates uncertainty and risk that can to some
extentbemitigatedbyaveryorganized,fair,andtransparentprocurementprocessthatstrives
to limit oneonone negotiations with a single bidder as much as possible. Therefore, we
recommendaprocessthathasbeenusedinP3procurementsglobally,andhasemergedasa
bestpractice.Thisprocessfeaturesadetailedrequestforqualifications(RFQ)processinwhich
biddersareshortlisted,andadraftRFPandfinalRFPprocessthatallowsfordialoguebetween
the City and shortlisted biddersresulting in robust competition amongst a small number of
qualifiedbiddersonafinalRFPandP3contractthatdoesnotnecessitateextensivenegotiations
afterselectionofpreferredbidder.
InapioneeringareasuchasFTTP,usingthisrecognizedP3processcouldreduceuncertaintyand
allowforrefinementandchangeofstrategiesortermsthatcanenhancevalueformoneyforthe
City,andproducefair,andtransparentcompetition.
TheneedfortimeisalsorequiredbythepotentialDualP3structureofseparatingtheRFPfor
thefibernetworkfrom theRFPfor operations andservice.Aspart of thisemergingstrategy,
there are two options for how to organize the dark fiber procurement, and the lit services
procurement:(1)sequentialprocurements,or(2)staggeredprocurements.Thefirstoptionisfor
theprocurementoftheP3stobedonesequentially,firstforthedarkfiber,andthenforthe
servicesP3.Thisoptionismoreinlinewiththetimingoftheprojectoverallandholdsfewer
procurementrisksrelatedtopossiblegamingoftheprocess,andchallenges.
ThesecondoptionisfortheprocurementoftheP3stobecompletedinastaggeredmanner,
suchthattheprocurementofthe DarkFiberP3begins,andonlyclosesaftertheLitFiberP3
procurementbegins.ThisstaggeredprocurementallowsfortheCitytousethedraftRFPprocess
withinbothP3procurementstoworkthroughcoordinationissuesbetweenthetwointerrelated
projects.ThisenablestheCitytoguardagainstlockingitselfintoaDarkFiberP3contractthat
maynotbeoptimalforthesubsequentLitFiberP3.Thisstaggeredapproachcanalsoprovidea
betterindicationofthefundingrequirementstotheCity,asinformationfrombothprocurements
canbeusedtoprovideafirmindicationoffinancialneedfromtheentire,integratedproject.
As such, although it will be more complicated to use a staggered procurement approach, we

recommendthattheCityusethestaggeredapproachintoordertoaddresscoordinationissues
withinthetwocompetitiveprocurementprocesses.
106
9.3 The P3 Procurement Process Should Begin as Early as Possible in Order to

Gain Inputs from Bidders Prior to Putting Together the Proposed Funding
Authorization
Funding authorization requirements and market viability necessitate a relatively extensive
timelinetopreparetheprocurementandeducatepolicymakersandthepublicregardingthe
project.
9.3.1 The P3 Procurements Should Be Initiated Before Full Funding Authorization

We recommend initiating the P3 procurements (but not complete procurement) before full
funding authorization. This will allow for an accelerated process, and for the City to use
informationgleanedfrombidderswithintheprocurementprocesstobemorecertainaboutthe
level of financial commitment that bidders will need from the City. Under this option, the
procurementprocesseswillbeinitiatedbeforefullfundingauthorization,andclosedafterfinal
fundingapprovalissecured.Ifproposalswouldbesubmittedpriortofundingauthorization,they
wouldcomewithsomanycaveatsandatsuchalevelofgeneralitythatthebiddingprocesswould
notresultinthekindofdefinitiveandefficientprocurementthatcouldfullyprotecttheCitys
interests.
9.3.2 The Bids Should Be Requested After Funding Authorization

Fundingauthorizationwillprovidecertaintytobiddersabouttheleveloffinancialcommitment
available to the project. Additionally, having funding certainty from the City will provide a
measureofcomforttoinvestorsastheyundertakethelevelofeffortandcostrequiredtobidon
this opportunity. Furthermore, prior to full funding approval, the City would presumably be
unabletoawardabindingcontract.
9.3.3 The Use of Stipends Will Encourage Bidder Participation

TheCitycanencouragebidderparticipationthroughamechanismthatisgenerallyusedinvery
largeP3procurementswhichistocompensatethebiddersfortheirefforts.Thiswouldlikely
be undertaken with the understanding that the winning bidders compensation from the City
would then be rolled in to the bid itself, and the Citys financial obligation for this particular
processwouldextendonlytothelosingbidders.
Thisisaviableandtestedwaytoconductalarge,significantscaleprocurement,andisfrequently
usedonceasubsetofbidders(thosethatareprequalified)areselected.Itwouldlikelyentail
paymentbetween$500,000and$2million,giventhenewnessofbroadbandP3sandthelevelof
effortnecessaryforeachbiddertodevelopnewpartnerships,models,andlegalanalysis.
107
Appendix A: Glossary
AdministrativeFeeTheportionoftheConnectionFeeallocatedtofundingtheadministrative
effortsofthePublicBroadbandEnterprise.
CommercialServiceFeesThefeechargedtoconsumersfordataservicesprovidedbyretail
servicesproviders.
ConnectionFeeThefeeassessedtohouseholdsandbusinessestofundtheP3s.
CPE(CustomerPremisesEquipment)Theequipmentatthesubscribersitethatconnectsto
theFTTPnetworkviathesubscriberport,suchasarouter,switch,orsettopbox.
DarkFiberConcessionaireTheprivateawardeeoftheP3forbuildingandmaintainingoutside
plantinfrastructure,includingfiberopticsandhubfacilities.
DarkFiberConnectionFeeTheportionoftheConnectionFeeallocatedtofundingtheDark
FiberConcessionaire.
DarkFiberLeaseFeesTheleasefeespaidbylesseesofdarkfiber.
DualP3ModelThedualP3structuredevelopedforthisinitiativethatwouldentailtwoP3s,
onefordarkfiber(DarkFiberP3)andoneforlitservice(LitFiberP3).
SingleP3ModelThealternativeP3structuredevelopedforthisinitiativethatwouldentaila
singleP3foropenaccessdarkfiber,withtheCityplayingnoroleinlightingthefiberoroffering
services.
EVC (Ethernet Virtual Connection) A service connection between two or more subscriber
portsorRSPaccessports.
FDC (Fiber Distribution Cabinet) A cabinet placed geographically between the network
aggregationswitchesorcentralofficewhichservesasanaggregationpointforfiberthatfeeds
subscriber premises. The cabinet houses either passive optical equipment (such as GPON
splitters)oractiveopticalequipment(suchasanOLT).
LitFiberConcessionaireTheprivateawardeeoftheP3forprovidingequipmentandlitservices
overdarkfiber.
LitFiberConnectionFeeTheportionoftheConnectionFeeallocatedtofundingtheLitFiber
Concessionaire.
108
NNI(NetworktoNetworkInterface)Aphysicalnetworkportthatactsasthedemarcation
pointbetweentwoserviceprovidernetworks;inthiscase,betweentheCityFTTPnetworkanda
serviceprovidernetwork.
OLT (Optical Line Termination) The upstream connection point (to the FTTP network) for
subscriberswhichconnectstotheONT.ThechoiceofanopticalinterfaceinstalledintheOLT
determines whether the network provisions shared access (one fiber split among multiple
subscribersinaGPONarchitecture)ordedicatedactiveEaccess(oneportforonesubscriber).
ONT(OpticalNetworkTerminal)Adevicethatsitsatthesubscriberpremisesandisconnected
tofiberfromtheFTTP.ThedeviceconvertsthenetworksignalbetweentheFTTPopticalfiber
andthecopperEthernetconnectionusedfortheCPE.
PublicBroadbandEnterprise(PBE)TheentityofSanFranciscogovernmentthatwillmanage
andadministertheP3contract(s),collectConnectionFees,andissuerelevantdebt.
QoS (Quality of Service) The means to ensure minimum performance levels for certain
networktrafficthroughprioritizationandreservationofresources.
ServiceProviderTheorganizationprovidinganetworkservice;mostcommonly,aninternet
serviceprovider(ISP).
SubscriberTheindividualororganizationpurchasinganetworkservice.
UNI(UsertoNetworkInterface)Aphysicalnetworkportthatactsasthedemarcationpoint
betweentheFTTPnetworkandasubscribernetwork.OnaFTTPnetwork,thiswouldusuallybe
acopperEthernetportonanONToranEthernetswitchwithanopticalupstreaminterface.
109
Appendix B: ValueforMoney Assessment and ValueforMoney Drivers

A valueformoney assessment is a structured comparison of delivery methods, usually
comparing a conventional delivery model (various single contracts for design, construction,
operations) with P3 delivery method (e.g., design, build, finance, operate, and maintainor
DBFOM).Theassessmentanswersthequestionwhichdeliverymethodprovidesthebestdeal
forimplementingaspecificprojectfromtheperspectiveofthetaxpayerorgovernment?
Avalueformoneyassessmentcanbeeitherqualitativeorquantitative.Forexample,forhighway
andsocialinfrastructureprojects,theresultofavalueformoneyassessmenttypicallyisanet
presentvalueorpercentagedifferencebetweenaconventionaldeliverymodelandaP3delivery
model.
However,anincreasingnumberofgovernments,intheU.S.andinternationally,havedecidedto
deemphasize the quantitative comparison or move away from the quantitative comparison
altogether;thisanalyticalshiftreflectsthelimiteddataavailabletojustifyexpecteddifferences
betweenconventionalandP3delivery,andanincreaseinclaimsregardingthemanipulationof
outcomeswithinthevalueformoneyassessment.
Thistrendhasledtoanincreaseinpurelyqualitativevalueformoneyassessments.Sincethere
is very little data on implementation of citywide broadband networks, our analysis of the
BroadbandforSanFranciscoInitiativedidnotincludeaqualitativecomparison.
Moreover,forthisinitiative,thereisnotreallyaconventionalmodelthatcanberealistically
implemented.Thismakesthetraditionalvalueformoneyassessmentinfeasible.
Thatsaid,thequalitativeargumentsaroundvalueformoneyarestillrelevantforstructuringand
assessing delivery models for the Broadband for San Francisco Initiative. The alternative
traditionalvalueformoneyassessmentincludedherediscussesthedriversofvalueformoney
andapplythoseinthediscussionandqualitativecomparisonofdeliverymodels.
110
Governance Mechanisms That Drive ValueforMoney

There are five major categories of governance mechanisms within P3 agreements that drive
valueformoney,asshowninFigure16:
1. Integrationandlifecyclecosting
2. Specificationsallowingforinnovation
3. Financialincentives
4. Competition
5. Efficientriskallocation
Figure16:P3ValueDrivers

Integration and Life Cycle Costing
Conventionaldeliveryofinfrastructurebypublicentitiestypicallyreliesonthepublicentityto
integrate multiple contractsfrom design, to construction, to a series of operations and
maintenancecontractsthroughoutthelifecycleoftheproject.Assuch,thepublicentityholds
riskrelatedtointegratingthesecontractedservices.
In P3 delivery, a private entity provides all the services for the life cycle of the infrastructure
throughonecontract,therebytakingontheriskofintegratingallthedifferentcomponentsof
theproject.IntegrationofthesedifferentcomponentsisafoundationforthevaluethatP3sare
expected to generate, because the integration aligns incentives for the private entity. For
111
example,theprivateentitymightmakespecificdecisionsinprojectdesignandconstructionthat
couldbesubsequentlyadvantageousinoperationsandmaintenance.
OutputBased Specifications That Allow for Innovation

In P3 contracts, specifications regarding what the concessionaire provides are outputbased,
leavingroomfortheconcessionairetodecidehowtodelivertheservicesthatarerequired.This
contrastswithtypicalcontracts,inwhichdesignandengineeringsolutionsareprovidedindetail,
andtheprivatepartiessimplyprovidetheservicetothedeterminedspecification.Outputbased
specifications,incombinationwithcompetitivepressurethroughabiddingprocess,incentivize
concessionairestocomeupwithcreativesolutionsthatseeklifecyclesavingsandhigherquality
ofservice.
Financial Incentives
AP3canalignpublicandprivateinterestsandincentivizetheprivatesectorsperformance
throughfinancialincentivesandpenalties.Positiveperformanceimprovestheprivatesectors
profits directly through timely, or higher, payments, or indirectly, through lower costs. Poor
performance triggers penalties, which will directly affect the financial performance of the
concessionaire.
Financialincentivesfortheconcessionairearestructuredasadirectrelationshipwiththeoutput
based specifications, and are linked to performance. A key to properly aligning financial
incentivesistomakesurethattheconcessionairehasmoneyatstake.Additionally,financing
alignstheinterestofdebtfinancierswiththeinterestsofthepublicentity.Theprivatepartywill
onlybepaid(andthedebtproviderrepaid)iftheproject/serviceisdeliveredwell.
Competition
The benefits of a P3 will only materialize if the concessionaire is procured through a fair,
transparent, and competitive process that features multiple, highlyqualified bidders. The
foundation of a competitive procurement is market appetite and capacity. A procurement
processthatfeaturesP3bestpracticessuchasshortlistingofprequalifiedbidders,usingadraft
RFP process, and limiting appropriations risk upfrontcan minimize transaction costs for all
partiesinvolved,andcandirectlyaffectmarketappetite.101
101
AnoteonP3transactioncosts:Limitingtransactioncostsisalsoimportantforthepublicsector.P3transactions
aretypicallymorecomplex,andusuallyrequiremoretime,andoutsidecounseltocompleteparticularlyifa
publicentitydoesnothavepriorexperiencewithprocuringP3s.HighertransactioncostsforP3sisexpecteddueto
thefactthataP3isessentiallyonecontractthatreplacesmultiplecontracts.Itisimportanttonotethatoverthe
lifecycleofaproject,transactioncostsassociatedwithconventionaldeliverymaybehigherthanaP3duetothe
needtoprocuremultiplecontracts,fromdesign,toconstruction,tofinancing,tovariousoperationsand
maintenancecontracts.
112
Anotheraspectofcompetitionistheprocurementevaluationcriteria.Thegoalofeverybidder
withinacompetitiveprocurementprocessistowin.Towin,biddersneedtoobtainthebestscore
visvistheevaluationcriteria.Therefore,theevaluationcriteriashouldfocusonthekeypublic
sectorobjectivesfortheproject.Inordertodothis,andbalancethegoalofcreatingthehighest
valueformoney, evaluation criteria are not simply pricebased, but based on bestvalue,
whichincludesachievingpublicgoals.
Efficient Allocation of Risks

ThecoreofaP3dealisefficientriskallocation.Riskallocationisbasedontheprinciplethatrisks
should be allocated to the party that can best manage the risks. Key to risk allocation is
recognitionthatsimplybecausetheriskisallocatedtoaspecificpartydoesnotmeanthatthe
riskdoesnotexist.
Also, regardless of whether the public or the private party assumes the risk, there is a cost
associatedwithbearingtherisk, whichisreflectedintheP3soverallvalueformoney.Some
risks, such as political risk or risks associated with budget appropriations, are typically best
managed by the public sector. Technical and interface risks such as errors and omissions in
design,orconstructiondelays,aretypicallybestmanagedbytheprivatesector.Notethatsome
risks,suchasforcemajeureandtechnicalchangesovertime,aretypicallysharedbetweenthe
publicandprivateparties.
ValueforMoney Analysis of Classic P3 Model

TheClassicP3ModeldoesnotcreatethebestvalueformanyprofileforbroadbandP3projects
because it does not efficiently allocate risks, limits competition, limits outputbased
specifications,andtypicallyinvolveshightransactioncosts.However,thismodeldoesallowfor
financialincentivesthroughperformancebasedpaymentsand/orrevenuesharing.
Efficientallocationofrisks:Mostrisksareallocatedtotheconcessionaireinthismodel.Given
thatefficiencyisdefinedasallocationofriskstothepartythatisthemostableandequippedto
manageandmitigatetherisk,theconcessionaireislikelynotthebestentitytomanageallrisks
associatedwithlongtermrevenueforservicesthattendtochangeovertime.
Integration and life cycle costing: This model does allow for maximumintegration of various
elementsoftheprojectinparticular,integrationbetweendarkfiberandlitservices.However,
aswenotedabove,thismodeldoescomewiththeriskofproducingaquasimonopoly,which
wouldlimitcompetitionforlitservices.
Outputbasedspecificationsthatallowforinnovation:Thismodelcanallowforoutputbased
specifications.However,itwillbechallengingtocreatespecificationsthataretrulyoutputbased
inthisP3model,becausetheeconomiclifeoftheequipmentisonlysevento10years.Thistype
113
ofinfrastructureishighlydynamic,whichmakesfutureoutputbasedspecificationsdifficultto
anticipate.
Financial incentives through performancebased payments and / or revenue sharing: This

modelcanbestructuredtoallowforfinancialincentives.
Transaction costs: We would anticipate high transaction costs under this model due to the
projectsfirstofitskindnature,thecomplexityoftheproject,andtheneedtodeterminethe
sourceofConnectionFees.
ValueforMoney Analysis of Dual P3 Model

The Dual P3 model seeks to create the best valueformany through efficient risk allocation,
incentivizingcompetition,andencouraginguseofoutputbasedspecifications.However,dueto
the complexity of this project, and the two procurement processes needed, high transaction
costsareexpected.
Efficientallocationofrisks:Risksareallocatedtotheconcessionairebestpositionedtomanage
them.Giventhatefficiencyisdefinedasallocationofriskstothepartythatisthemostableand
equippedtomanageandmitigatetherisk,theDarkFiberConcessionaireisthebestentityto
managerisksassociatedwiththelongtermprovisionofdarkfiberinfrastructure.Inthesame
vein,theLitFiberConcessionaireisthebestentitytomanagerisksassociatedwithrevenuesfor
theequipment,networkoperations,andserviceprovision.
Using a shorter concession term for the Lit Fiber P3 also inherently puts some of the risk
associatedwithrevenuesthatwillmostlikelychangeovertimebackintothehandsoftheCity
whichallowsfortheCitytoalsopartiallymanagethisrisk.
Integration and life cycle costing: This model does allow for maximumintegration of various
elementsoftheprojectinparticular,integrationbetweenthedarkfiber,networkoperations,
andserviceprovisionlevels.However,aswenotedabove,thismodeldoescomewiththeriskof
producingaquasimonopoly,whichwouldlimitcompetitionattheserviceprovisionlevel,and
potentiallyalsowithrespecttodarkfiber.
Outputbasedspecificationsthatallowforinnovation:Thismodelcanallowforoutputbased
specifications.EachP3contractwillincludeoutputbasedspecificationsthatareappropriatefor
theeconomiclifeoftheinvestment.TheshorterconcessiontermfortheLitFiberP3willallow
theP3contract(andsubsequentP3contracts)toadequatelyanticipatechangesoverashorter
timeduration,recognizingthehighlydynamicnatureoftheseinvestments.
114
Financialincentivesthroughperformancebasedpaymentsand/orrevenuesharing:Thismodel
can be structured to allow for financial incentives, including performancebased payments
and/orrevenuesharing.
Transactioncosts:WewouldanticipatehightransactioncostsundertheDualP3Modeldueto
theprojectsfirstofitskindnature,thecomplexityoftheproject,theneedtogotodetermine
the source of Connection Fees, and the fact that two complex procurements will need to be
completedinastaggeredfashion.
115
Appendix C: Risk Matrix

ThefollowingtablesummarizestheProjectTeamsassessmentofpotentialriskfactorsacrossa
rangeofprojectphases.
116
Category Description Risk Impact MitigationApproach Public Private
Priortotheengineeringphase,alertthevariouspole
ownersoftheprojectscopeandnegotiatea
Utilitypoleownersandattachedpartiesimpose
streamlinedprocessfortheconcessionairetoobtain
excessiveengineeringrequirementsandprovidepoor
poleattachmentinformationoverthecourseofthe
documentationforexistinginfrastructure,increasing High Medium 60 40
project.Constantlyreviseprioritiesandschedulesto
thedifficultyofaerialengineeringwhichresultsin
accommodateunexpectedconstructionroadblocks.
increaseddesigncostsandprojectdelays.
Constructundergroundwherenecessary.Establisha
onetouchmakereadypracticeforanewattacher.
WorkwiththeCityagenciestoeliminatenon
essentialpermitrequirements.Constantlyrevise
StringentCityagencypermitting,inspectionand
prioritiesandschedulestoaccommodateunexpected
restorationrequirementsincreasesthedifficultyof
High Medium constructionroadblocks.Concessionairecallsinutility 70 30
Design undergroundengineeringwhichresultsinhighdesign
locatesduringtheengineeringdesignprocessfor
costsandprojectdelays.
certainareasoftheCitywhereunderground
congestionisexpected.
Poordocumentationofexistinginfrastructureincrease DevelopaprocesstomakesureCityagenciesprovide
thedifficultyofundergroundengineeringwhich High Medium thebestavailableinformationonunderground 100 0
resultsinincreaseddesigncostsandprojectdelays. utilitiestotheconcessionaire.
Cityagencieschangeconstructionstandardsduring
WithacceptancefromtheappropriateCityagencies,
thecourseoftheprojectcausingengineeringdesigns
developasetofclearlydefinedconstruction
tobecomeinfeasibleornotoptimalfromacost Low Medium 100 0
standardsthatshallremainvalidthroughoutthe
perspective.Thiscouldresultinadditionalcostsand
durationoftheFTTPconstructionproject.
projectdelaysduetodesignrevisions.
117
WorkwithCityagencyorpoleownertoestablish
DelayedresponsesfromCityagenciesorpoleowners responsetimerequirementsforanyprocessthat
whentheengineeringcontractormakesinformation High Low currentlydoesnothavedefinedresponsetimes. 90 10
requestspertainingtoengineeringorpermitting. IdentifyappropriatePOCsforeachagencyandpole
owner.
Workwiththecityagenciestoeliminatenon
essentialpermitrequirementsanddevelopa
Cityagenciesand/orutilitypoleownersdonothave streamlinedprocessforthevariousCityagenciesto
thecapacitytoreviewengineeringandprovide reviewessentialpermits.
High Medium 100 0
permitsinatimelymannerforaprojectofthisscale, Phasetheplannedaerialconstructiontoaccountfor
causingdelaysinconstruction. expecteddelayswithpoleattachmentreview.
Prioritizethesubmissionofpoleattachmentpackages
sothatcriticalroutesarereviewedfirst.
Engineeringcontractorhasdifficultyobtaining Developplantoeffectivelyenforcenewlegislation
permissiontoaccesscommercialbuildingsandMDUs, High Medium thatrequiresbuildingownerstoallowaccesstoall 50 50
resultingindelaysorincompleteengineering. statelicensedproviders.
Unanticipatedconstructionrequirementsand/or
Concessionairecallsinutilitylocatesduringthe
obstructionsthatcannotbeidentifiedduringthe
Construction Medium High engineeringdesignprocessforcertainareasofthe 50 50
designphasethatresultinincreasedcostsandproject
Citywhereundergroundcongestionisexpected.
delays.
118
AreasoftheCitymayhavelimitedspaceinthe Concessionairecallsinutilitylocatesduringthe
undergroundROWresultinginredesignofroutes Low High engineeringdesignprocessforcertainareasofthe 0 100
whichcanincreasecostsanddelayedconstruction. Citywhereundergroundcongestionisexpected.
PoleownerssuchasPG&EandAT&Thaveamultiyear
schedulesforpolereplacementandmakereadywork. Negotiatewithpoleownerstomakeanexceptionto
Theydonotperformworkoutoftheirscheduled theirstandardprocess,allowingoutoforderworkto
order,nordotheyprovidedetailsofthescheduled beperformedontheCity'sproject.
work.Significantprojectdelaysforaerialbuildscould High High 50 50
occurwhilewaitingforpoleownertoperformmake Obtainthedetailedworkschedulefromthepole
readywork.Increasedcostsforengineeringredesign ownersoengineeringandconstructioncanbephased
andundergroundconstructionmayoccuriftheCity toaccommodatetheplannedmakereadyschedule.
buildsundergroundtobypassthedelays.
Ifneededtheconcessionairecanworkdirectlywitha
Constructioncontractorsareunabletoobtainfiber
fibercablemanufacturertoreserveaspotinthe
constructionmaterialsduetosupplyshortages Medium Medium 0 100
productionqueuetoensurefibercableisavailable
causingprojectdelays.
duringtheproject.
Localutilitylocatingservicesdonothavecapacityto Priortoconstruction,alertlocallocatingservicesof
keepupwithlocatesrequestsforaprojectofthis Medium Medium impendingprojectscopetoallowthempreparefor 50 100
scale,causingdelaysinconstruction. theinfluxoflocaterequestsasbestaspossible.
119
Requiretheselectedconcessionairetousemultiple
constructioncontractors,sothatcontactorswho
haveissueswithdamagingutilitiescanberemoved
fromtheprojectwithoutcompromisingtheproject
ProjectcontractordamagesCityownedorthirdparty
timeline.
utilitiescausingdelaysinconstructionandproject
Medium Medium 0 100
pushbackfromCityagenciesorthirdpartyutility Requireconcessionairetouseconstruction
owners. contractorswithexperienceworkingintheCity
and/oraproventrackrecordoflowhitrates.
Developadditionalqualitycontrolprocessesfor
contractorstoimplementaftertheyhitautility.
Designthenetworktohavebackuppoweratcore
Apoweroutageatcoresitesorhubfacilitiescausesa
Low Medium sitesandhublocations.Designnetworkwith 0 100
serviceoutageonthenetwork.
redundantroutingtominimizeserviceoutages.
EnsureSLAsareinplacetoguaranteequickresponse
Networksufferssignificantdamagefromearthquake andrepairtimesfordamagedassets.Designnetwork
Low High 50 50
Operations orothernaturaldisaster. withredundantroutingtominimizeserviceoutages
duringrepairs.
Ensurefirewallsandothersecuritymeasuresandbest
Acyberattackcausesaserviceoutagesorvulnerability practicesarebuiltintothenetwork.Developa
Medium Medium 0 100
tousersonthenetwork. responseplantodealwiththeattackandrestore
services.
120
Ensureallundergroundportionsofthenetworkhave
properlyregisteredwith811.
Collectaccurateasbuiltdocumentationattheendof
theconstructionphasetocreateaccessiblerecordsof
UtilityconstructionwithintheCitydamagesnetwork thenetworkinfrastructure.
High Low 80 20
fiberinfrastructurecausingoutages.
EnsureSLAsareinplacetoguaranteequickresponse
andrepairtimesfordamagedassets.
Designnetworkwithredundantroutingtominimize
serviceoutagesduringrepairs.
Networkbecomesoversubscribedduetonew Designadditionalcapacityintothenetworkto
Low Medium 0 100
developmentintheCityandpopulationincrease. accommodategrowth.
Ensureequipmentisinstalledtomanufactures
Themalfunctioningofequipmentplacedatornear specifications,local,andstateregulations.
customerpremisesresultinginpropertydamageor Low Medium Ensurevalidpermissionsareobtainedand 0 100
otherCity/concessionaireliability. documentedforanyinstallationoncustomer
property.
Fiberisthecornerstonetechnologyinwirelineand
Technologiessuchas5Gandsmallcellwirelessrender wirelessnetworks.ConcessionaireandtheCityneed
Low Low 20 80
acitywideFTTPnetworkobsoleteinthenearfuture. tobeawareoftechnologicalandmarketchangesand
keeptheserviceofferingsrelevant.
121
Weexpectsomelevelofinterestratefluctuation.To
reducetheexposureofthisriskfortheCity,adefined
maximumavailabilitypaymentcanbeusedintheP3
Interestrateriskafterfinancialclose High Low 100 0
agreements,useaconservativeinterestratefor
planningpurposes,andupdatedocumentswiththe
latestinterestrates.
StructuretheConnectionFeesuchthatcollections
aretiedtoanexisting,"triedandtrue"process.
AbilitytocollecttheConnectionFee Low High 100 0
Financial Currently,theConnectionFeeisstructuredasa
parceltaxwhichhasahistoryofreliablecollections.
Bank,market,underwriting,orequityissuesarisefor Ensureabilitytofinancethestructureduringthe
Low Medium 0 100
eitherconcessionairethatcandelayfinancialclose procurements.
Thepreprocurementauctiondoesnotsecureenough Undertakesignificantpremarketingandpublicizing
Medium Low 100 0
inleaserevenuetomaketheprojectfeasible ofthepreprocurementauction
Refinancingconditionsarenotgeneratingmaterial
Medium Low Useconservativeassumptionsregardingrefinancing. 50 50
upside
Continuecurrentpracticeofongoingoutreachto
incumbentsandcreateopportunityforparticipation;
Political Incumbentopposition High High 100 0
recognize,however,thatincumbentoppositionmay
persistregardlessofstepstheCitytakes
122
GainaclearunderstandingofallCityapprovalsthat
DelaystoCityapprovalsfortheP3contractsorthe
? Medium areneededforthisproject,andengagewith 100 0
project
stakeholdersearlyandoften.
Workwithlegislators,stakeholders,electedofficials,
etc.tofullyunderstandpublicconcerns.
Managingconcernsandissuesraisedbypublic Priortofinancialclose,thepublicwillownallofthis
Medium Medium risk.Postfinancialclose,theconcessionairewillwant 80 20
stakeholdersthatcanstoptheproject
tobeengagedineffortstomanagestakeholder
issues,andshapingpublicopinionabouttheproject,
andservicesprovided.
P3darkfiberprocurementdoesnotgarnersignificant
Engagethemarketearlytounderstandinterest,
marketinterest,anddoesnotgeneraterobust Low High 100 0
identifypossibleissues/showstoppers,etc.
competition
SetarealisticscheduleforthetwoP3s,includingan
DelaysintheprocurementprocessforeitherP3 High Medium 100 0
appropriatetimeframeforreview,approvals,etc.
Procurement
Carefullystructurethemarketandalignincentivesfor
Limitedcompetitioninsubsequentroundsofthelit
Medium Medium thefuturebaselevelserviceP3procurement(in57 100 0
fiberP3
years)
123
StructurethedarkfiberP3procurementwiththe
ProposalsreceivedinthedarkfiberP3procurement baselevelserviceP3inmind,andincludedialogue
Medium High 100 0
makeitverydifficulttoimplementthelitfiberP3 andtimedelaybetweentheprocurementsinorderto
identifyandaddressissues.
Bidderappealsoftheprocurementprocess,
Clearlyarticulatetherulesoftheprocurement
particularlyifitisseentoadvantageanincumbent,or Low Medium 100 0
process,tolimitopportunitiesforappeal.
limitsubsequentcompetition
Keepanactiveeyeonissuesassociatedwith
Distortionoflevelplayingfield,lackofacompetitive
collusion,andmanagepreprocurementauction
procurementduetocollusion,havingonebidder Medium High 100 0
properlytoavoidissuesoflargelesseesthatare
secureaverylargelease,etc.
dedicatedtooneproposer.
124
Appendix D: Direct Economic Benefits of Fiber Construction in San

Francisco
TodeterminethepotentialeconomicbenefitsoftheBroadbandforSanFranciscoinitiative,the
followingisananalysisofthepotentialdirectjobcreationfromthisinitiative,asconstruction
laborandnetworkstaffingrepresentsalargeportionofthetotalprojectcost
CTCengineersundertookanindependentanalysisoflikelyjobcreationbasedonourexperience
analyzing, testing, and designing communications networks and on our observations of fiber
deployments elsewhere. Our analysis is also based on our direct experience with network
projectsinothercitiesthataresubstantiallysimilartoSanFrancisco.
In our experience, there are two distinct categories of direct job creation from broadband
initiatives:(1)networkconstructionand(2)operations:
Thefirstcategoryofdirectjobswillonlyexistduringthethreeyearprojectconstructionperiod.
Thesejobsrelatetothelaborassociatedwithconstructionanddeploymentofthenetwork(e.g.,
technicians and construction workers who lay the broadband pipes). 102 This category of job
creationincludesthedirectlaborassociatedwithinstallingtheinfrastructurecomponentsand
equipmentthattakesthefibertotheuserlocation.Italsoincludesconstructionofwirelesstower
structuresandnetworkelectronics.
Fiber construction in San Francisco will have an immediate impact in the construction and
information technology sectors. Workers will engage in engineering, planning, and
environmentalcompliance.Duringtheintensivethreeyearconstructionperiod,workerswilldig
ditchesandtrenchesandplacefiberinthegroundandonutilitypoles.Additionalworkerswill
place and configure communications equipment, manage installation, repair supporting
infrastructure,andinteractwithcustomers.Someofthesedirectbenefitswillaccruetoworkers
fromoutsidethecommunity.Nonetheless,theseworkerswillbenefitthelocaleconomybecause
theywillpurchasefuel,eatatlocalrestaurants,andsleepinlocalhotels.
Basedonourexperience,weconservativelyestimatethat309jobsperyearwillbecreatedduring
thethreeyearconstructionphaseofaSanFranciscobuildout(Table25).
102
RobertD.Atkinson,DanielCastroandStephenJ.Ezell,TheDigitalRoadtoRecovery:AStimulusPlantoCreate
Jobs,BoostProductivityandRevitalizeAmerica,ITIF,Jan.7,2009,at2
(http://www.itif.org/files/roadtorecovery.pdf)(accessedMay10,2017).
125
Table25:EstimateofJobCreationDuringNetworkConstruction
Avg. Estimated Construction

Project
Item Unit Hours Total FTEs(3Year
Scope
PerUnit Hours BuildOut)
Project Management and
Miles 110 1,830 201,300 35
QualityControl
Physical Plant Engineering and
Miles 80 1,830 146,400 25
Permitting
Construction Miles 580 1,830 1,061,400 184
NetworkElectronicsDesignand
Subscribers 1 373,481 373,481 65
Deployment
TOTAL 309
Thesecondcategoryofdirectjobcreationincludesjobsassociatedwiththeongoingoperations
for both the dark fiber and the network electronics. The network will create skilled jobs as
workersinstallandactivatecommunicationsequipmentandothermaterialsthatareessential
fortheoperationofthefiberoptics.Weestimatethatatleast170jobswillbecreatedperyear
intheoperationsphaseforfibermaintenanceandnetworkoperations,including:
47FTEstooperatethefiberportionofthenetwork
8FTEcontractorsforlocatesandticketprocessing103
27FTEcontractorsforfiberrepairs104
88FTEsfornetworkoperations(e.g.,networksupport,networkmanagement,callcenters
andcustomersupport)
Thefirstthreecategoriesofoperationsjobsrelatetothemaintenanceofthedarkfibernetwork.
ThemajorityofthesejobsarelikelytobelocatedinorneartheCityofSanFrancisco,sincethey
entailmaintainingthephysicalaspectsofthenetwork,whichcannotbecompletedremotely.A
significantpercentageofthefinalcategory(networkoperations,whichincludesactivitiesrelated
toprovidingcustomerserviceandmanagingcallcenters)canbeperformedinotherpartsofthe
country,andperhapstheworld.
Insum,weestimatethatconstructionofthe1,830milefibernetworkwillsupport309jobsin
eachofthethreeyearsthattheprojectisconstructedandanadditional170jobsthroughoutthe
projectlife.
103
Basedonestimatedfeespaidtocontractorforservicesinyear5andthereafter($712,900),assuming90percent
oflistedexpensesattributedtolaborcosts,$60,000perFTE,anda40percentcostburdenperFTE.
104
Based on estimated fees paid to contractor for services in year 5 and thereafter ($15,040,400), assuming 25
percentoffiberrepairexpensesattributedtolaborcosts,$100,000perFTE,anda40percentcostburdenperFTE.
126
Appendix E: Technical Performance Metrics and Standards

Standard Specifications for Construction and Maintenance of the FTTP Network Fiber
The selected Dark Fiber Concessionaire shall provide the design, engineering, permitting,
procurement, material coordination, construction, testing, and completion of a fiber optic
network,primarilyconsistingof,butnotlimitedto,thefollowingtasks:
DevelopsystemlevelnetworkdesignsthatmeettheCitysgoals;
Performfieldwalkoutanddocumentationofallfiberroutes;
PrepareGISbaseddesignsandCADconstructionprintsoffinaldesigns;
Obtain all necessary permits, including environmental approvals and/or jurisdictional

determinations;
Obtainallnecessaryagreementsforutilitypoleattachmentorthirdpartyconduituse;
InitiationofutilitylocaterequeststhroughtheUSANorth811,andstrictadherencetoall
Californiautilitylocatinglaws;
Subsurface installation of conduit, primarily through the use of horizontal directional

drilling, microtrenching, and trenching, including existing utility locating through test
pitting,trafficcontrol,andpermanentsurfacerestoration;
Installation of underground handholes and ground rods, including permanent paved

surfacerestoration;
Placementoffiberopticcableandtracerwireinconduit;
Installationofnewmessengerstrandbetweenutilitypolesforaerialcableconstruction,
toincludeinstallationorrelocationofguywiresandanchors;
Lashand/oroverlashfiberopticcabletoaerialmessengerstrand;
Installationoffiberspliceenclosuresandfibersplicing;
Installationoftelecommunicationhutsandfiberdistributioncabinets;
Placementandassemblyoffiberterminationpanelsandrelatedhardware;
Opticalperformancetestingoffiberopticstrands;
Ongoingnetworkmaintenanceandrepair;and
127
Provisionandstorageofallmaterialsrelatedtotheaboveconstructiontasks.
TheConcessionaireshallberesponsibleforidentifyingandsecuringsuitablespaceformaterial
andequipmentstorageandstagingduringconstruction.
The Concessionaire will provide regular progress reporting, and will closely coordinate its
constructionschedulewiththeCity.TheConcessionaireshallprovideaprimarypointofcontact
totheCityforthedurationofthecontract,andshallbeexpectedtoattendregularprojectstatus
and management meetings. The Concessionaire shall provide daily progress reporting and
forecastingoftheconstructionlocationsforthefollowingworkdayduringactiveconstruction
phasesoftheproject,andshallprovideweeklyreportingofkeyprogressmetricstobedefined
bytheCity.
Network Design Specifications

ThedesignfortheFTTPnetworkshallbeinaccordancewiththegoalsandobjectivesexpressed
by the City. The design is limited to the physical layer of the network (conduit, fiber, hub
locations,etc.),anddoesnotincludenetworkelectronics,butmusttakeintoaccountcurrent
and emerging FTTP network technologies. It is the intent of the City for the FTTP physical
infrastructure to support any current or future mix of Passive Optical Network (PON), Active
Ethernet,and/orfuturetechnologystandard.
Network Construction Specifications

ConstructionofthenetworkshallbeincompliancewiththeNationalElectricalCode,theNational
ElectricSafetyCode,andapplicableindustrystandards,aswellasanyandallotherapplicable
Federal,Stateandlocallawsandregulations.Constructionmethodsandtechniquesusedshall
beinaccordancewiththerecommendedpracticesandprocedurespublishedbyleadingindustry
manufacturersandtradeassociations,includingbutnotlimitedtothefollowing:
Society of Cable Telecommunications Engineers (SCTE) Recommended Practices for

OpticalFiberConstructionandTesting;
TelcordiaBlueBookManualofDesignBuildersProcedures;
CaliforniaOccupationalSafetyandHealth(Cal/OSHA)Division;and
FederalOccupationalSafetyandHealthAdministration(OSHA)regulations.
Construction shall also follow all applicable local standards, including but not limited to the
following:
DepartmentofPublicWorksStandardSpecificationsandPlans;
128
DepartmentofPublicWorksOrderNumber178,940;
SanFranciscoMunicipalTransportationAgencyRegulationsforWorkingin
SanFranciscoStreets(BlueBook);and
CaliforniaPublicUtilitiesCommissionGeneralOrder95.
The contractors shall be aware of all standards and their application. Ignorance or lack of
knowledgeshallnotbeanexcuseforimproperworktooccur.Anyworkconstructedinviolation
ofanyapplicablecodeshallbecorrectedandreinstalledproperlyattheContractorsexpense.
Construction Material Specifications

Materialsusedintheconstructionofthenetworkmustmeetacceptableindustrystandardsand
shallmeettheCitysspecificationforeachindividualitem.
Conduitusedfortheconstructionofthenetworkwillgenerallyhaveanominaldiameterof2
inches,unlessotherwisedirectedbytheCity.Conduitshallbehighdensitypolyethylene(HDPE)
rollductmeetingapplicableASTMstandards.Allsuppliedconduitshallhaveasmoothinnerwall
and smooth outer wall. All conduit shall be prelubricated, and contain a preinstalled 1100
poundpolyesterpulltape.
When microtrenching is used to construct the network, microduct shall be used in place of
conduit.Microductshallbehighdensitypolyethylene(HDPE)thatconformstoASTMD335098a,
Type III, Category 5, Class B or C and Grade P34 per ASTM D124884, or meets equivalent
standards.UnlessotherwisedirectedbytheCity,microductshallconformtothefollowingsizes:
12mmnominaloutsidediameterand8mmnominalinsidediameter
14mmnominaloutsidediameterand10mmnominalinsidediameter
The fiber optic cable used for the construction of the network shall be in accordance to the
followingspecifications:
Ribbonopticalcableforbackboneapplications,containing12strandribbons,conforming
tointernationalstandardsITUTG.652.DandTelcordiaGR20.
Loosebuffertubeopticalcableforservicedropapplications,containing12strandbuffer
tubes,in12and24countsheaths,conformingtointernationalstandardsITUTG.652.D
andTelcordiaGR20.
Delivered on reels holding a contiguous fiber cable length of up to 18,000 feet. The
Contractor shall be expected to coordinate cable orders to ensure cable lengths are
129
providedasneededsupportcontiguouscablerunswithoutsplicing(notincludingmid
sheathsplices)accordingtofinalengineeringdesigns.
ComprisedofAllDielectric(AD)materials.
Comprisedofagelfreecabledesignincorporatingdrywaterblockingelements.
Marked,inpermanentwhitecharacters,with:
o Manufacturername
o Monthandyearofmanufacture
o Numberofopticalfibers
o Sequentiallengthmarkings,minimumofeverytwofeet,infeet
Specifications for Fiber Testing

Opticalperformancetestsshallbeperformedforallnewand/orrepairedcablestovalidatethe
optical performance of the entire link, as well as to verify that fiber splicing has occurred
accordingtosuppliedsplicematrices.ThistestingwillconsistofbidirectionalOTDRtesting,as
wellasdirectopticalattenuationandcontinuitytestingusingacalibratedopticalsourceand
powermeter.Thistestingshalloccuronlyafterfibersareterminatedonbothendsofalink,and
allintermediateconstructionand/orsplicinginvolvingthereentryofinstalledsplicecasesor
handling of the fiber optic cable iscompleted for a particular segmentunder test.Electronic
documentationofalltestresultsshallbeprovidedtotheCity.
Testingshallbedeemedsuccessfullycompletedif:(1)maximumfiberlossesmeetmanufacturer
specifications,withanallowanceforsplicesandconnectors;(2)individualsplicelossesdonot
exceed 0.1 dB; and (3) maximum mated connector losses do not exceed manufacturer
specifications. Testing will be performed by the Concessionaires contractors, and may be
observed by the City and other designated representatives of the City. The City may request
and/orperformadditionaltestingtoverifyresultspriortoacceptingtestdata.
AnOTDRshallbeusedtomeasureanddocumentsplicelossesandconnectorlosses.Tocorrectly
identifyabnormalitiesatashortrange,a100meterorlongerlaunchcableshallbeusedbetween
theOTDRandthefiberundertest.Bidirectionaltracesshallbeacquiredforeachfiber.Ifthe
connection of the launch cable to the patch panel requires optimization by the operator,
samplingacquisitionwillcommenceuponcompletionoftheoptimization.
130
Eachfiberwillbeidentified,andtheresultsofthetestforeachfiberwillberecordedasindicated
below.Thetestwillberepeatedforeachofthefiberslinkingaparticularsite.Alltestswillbe
madeat1310nmand1550nm.
SettingsontheOTDRshallreflectthefollowing:
The Refractive Index shall be set for the actual fiber utilized (commonlyused Corning
SMF28singlemodefiberhasarefractiveindexof1.4677at1310nm);
Pulsewidthnogreaterthan100ns(10m)forallfiberlengths;
Scatteringcoefficientspecifiedbythefibermanufacturerforeachwavelengthtested;
Aminimumof10,000samplingacquisitions(averages);
Maximumrangesettonomorethan10kmforallfiberlengthlessthan10km;
Maximumrangesettonomorethan25kmforfiberlengthsgreaterthan10km;and
Eventthreshold:0.05dB.
A uniform filenaming scheme for recorded data shall be used that comply with conventions
mutuallyagreeduponbytheCityandtheConcessionaire.
InstalledopticalfiberOTDRtestdocumentationshallinclude:
Totalfiberlength;
Individualfibertracesforcompletefiberlength;
Lossesofindividualsplicesandconnectors;
Lossesofotheranomalies;
Wavelengthtestedandmeasurementdirections;
Manufacturer,modelandserialnumberofthetestequipment;and
Nameandcompanyofthetechnicianperformingthetests.
Alldatacollectedateachlocationduringthetestsshallberecordedatthetimeofthetestsusing
electronicmeans.
131
Optical power meter measurements shall be made at the same time as the OTDR tests to
determineoverallfiberlossandto ensurethatfibershaveappropriateendtoendcontinuity
(fibersnotcrossed).Powermetertestingshallbeperformedatboth1310nmand1550nmand
shallreporttherelativelossofeachfiberstrand.
Fiber Maintenance Requirements

Fibermaintenanceandemergencyrepairservicesshallbemadeavailableona24x7x365basis.
The Dark Fiber Concessionaire shall have a reasonable inventory of materials and vehicles
normally supplied under the terms of this contract for maintenance requests and emergency
repairs.TheseitemsbelowdescribetheminimumlevelofservicetheDarkFiberConcessionaire
shallprovidetotheLitFiberConcessionaireandtheCity.Exacttermsshouldbedeterminedon
acasebycasebasisandstatedinanSLAwiththeLitFiberConcessionaire.
Meantimetoacknowledge:1hour
Meantimetorespond:4hour
Meantimetorepairfiber:24hours
TimetoacknowledgemeanstimefortheDarkFiberConcessionairetoacknowledgereceiptof
theservicerequestandverballycommunicateanacceptableplanofactiontoresolveservice
requestsubmittedbytheCityorLitFiberConcessionaire.
Time to respond means time to arrive at the location of fiber damage, with all necessary
equipmentandpersonnelresourcestomakerepairs.
Performance Requirements for the Lit Network

The selected Lit Fiber Concessionaire shall provide operation of the networks lit services,
primarilyconsistingof,butnotlimitedto,thefollowingtasks:
Transportofthenetworksservices;
DesignandoperationoftheCivicNetwork;
NOC (network operations center) services which include monitoring, troubleshooting,

andfixingthetransportnetwork;
ProvidingsupportandprovisioningservicesforRetailServiceProviders;and
EngagingtheDarkFiberConcessionairewhenafixornewfiberinstallationisrequired.
132
Lit Network Performance Specifications

TheLitFiberConcessionaireshallsupplyandprovisionnetworkelectronicscapableofdelivering
highlyavailabletransportforavarietyofservicesoverthedarkfiber.Transportservicesshall
meetorexceedaCommittedInformationRate(CIR)of50Mbpswiththeabilitytoprovidea
rangeofupstreamanddownstreamspeedsuptosymmetrical1GbpsspeedsfromtheProvider
Porttothepointofnetworkinterface.
Transport services provided by the Concessionaire shall meet the following minimum
performancerequirementsforanylinkbetweenaRSPportandasubscriberportorbetweenany
twosubscriberportsonthenetwork.
Networkavailability(monthly):99.9%
Packetloss:<0.3%
Latency(oneway):<10ms
Jitter(oneway):<3ms
Network Monitoring Requirements

TheLitFiberConcessionaireshallalsoprovideNOCservicestomonitorthenetworkandactas
thecentralpointofcontactforservicerequestsandtroublereportsfromserviceproviders.For
requeststhatfalloutsidetheirscoperesponsibilities,theNOCshallberesponsiblefordispatching
thecorrectteam.TheseitemsbelowdescribetheminimumlevelofservicetheNOCshallprovide
toRSPs.ExacttermsshouldbedeterminedonacasebycasebasisandstatedinanSLAwitheach
RSP.
Meantimetoacknowledge:15minutes
Meantimetoprovision(excludingfiberandelectronics):4hours
Meantimetorespond:4hours
Meantimetorepair(excludingfiber):6hours
Meantimetorepairfiber:24hours
TimetoacknowledgemeanstimeforaNOCrepresentativetoacknowledgereceiptofaservice
requestortroublereportsubmittedbyanRSP.Thisexcludesautomatedresponses.
TimetoprovisionmeanstimetoprovisionanexistingserviceEVCtoanexistingsubscriberport.
Thisexcludescasesthatrequirenewfiber,newservicedrops,ornewelectronics.
Time to respond means time to arrive on site with electronics or arrive at location of fiber
damage,whicheverisapplicable.
133
Appendix F: List of Information to Be Available to Bidders

This appendix identifies information deemed useful to a Master Developer looking to build a
ubiquitousbroadbandnetwork(suchasfibertothepremises)throughouttheCity.Thedatasets
identifiedinthisreportarecategorizedbythetypeofinformationtheycontain,including:
1. Cityandthirdpartyinfrastructure
2. PublicrightofwayandCityownedproperty
3. Potentialnetworkusers
4. Constructionstandardsandrestrictions
Foreachcategory,thisappendixidentifieswhatdataareavailable,recommendswhatadditional
datashouldbecollected,andproposesamethodforthecollectionoftheadditionaldata.
City and ThirdParty Infrastructure

Cityownedorthirdpartyinfrastructurethatexistswithinthepublicrightofway(e.g.,conduit
orutilitypoles)mayprovebeneficialtoaMasterDeveloperthatcanleveragethoseassetsto
reduceitsconstructioncostsandavoidthechallengesoftheCityscrowdedrightsofway.
Existing infrastructure could also present obstructions and potential complications during
construction.Undergroundutilitiessuchasgas,water,andsewertakeupspaceintherightof
wayandmaymakeconstructionofnewutilitiesdifficultinsomelocations.Additionally,street
furniture, which may provide value for wireless deployments, can also present obstacles for
certaintypesofundergroundconstruction.
Available Data
The City has some data on existing infrastructure, but most of the datasets provide limited
informationbeyondthelocationoftheasset.AMasterDevelopercouldusethisinformationto
getageneralsenseoftheassetsintheCity,butadditionalworkwouldberequiredtomakethe
datamoreuseful.Table26identifiesdatathatareavailableforusebyaMasterDeveloper.
Table26:AvailableInformationonExistingInfrastructure
ID# Dataset
2.1.1 SanFranciscoDepartmentofTechnologyFiberMap
1.1.2 SanFranciscoDepartmentofPublicWorksUtilityUndergroundingMap
2.1.3 SanFranciscoPublicUtilitiesCommissionConduit,Pullboxes,andStreetLights
2.1.4 SanFranciscoMunicipalTransportationAgencyConduitMap
2.1.5 SanFranciscoMunicipalTransportationAgencyRoutesandStops
2.1.6 SanFranciscoMunicipalTransportationOverheadLinesTrolleyPoleDatabase150813
2.1.7 SanFranciscoMunicipalTransportationMapofTrafficSignals
2.1.8 SanFranciscoDepartmentofPublicWorksMapofStreetTrees
134
San Francisco Department of Technology Fiber Network Map

This dataset provides a map of the municipal fiber network operated by the Department of
Technology(DT).Themapalsoidentifieswhichportionsofthenetworkareundergroundand
whichareaerial.ThisinformationmaybeusefulasaMasterDevelopermaybeabletoutilize
additionalconduitspaceontheundergroundfiberroutesoroverlashtotheCitysaerialroutes
toreduceconstructionexpenses.DTcanprovidethismaptotheMasterDeveloper.
San Francisco Department of Public Works Utility Undergrounding Map

This dataset is a geocoded map of where the power and telecommunications utilities are
underground.Thedatasetdoesnotcontainanyspecificinformationabouttheutilities,justthat
utilitiesarelocatedunderground.ThisdatasetmayinformaMasterDeveloperofwhereaerial
constructionmaynotbepossible.ThismapcanbemadeavailabletotheMasterDeveloperby
theSanFranciscoPublicUtilitiesCommission(SFPUC).
San Francisco Public Utilities Commission Conduit, Pullboxes, and Streetlights

ThisdatasetprovidesageocodedmapofallSFPUCownedconduitinstalledpriortoSeptember
2005,aswellasdataonSFPUCstreetlightsandpullboxes.Thisconduitinformationmaybeuseful
to a Master Developer that seeks to use existing conduit space to reduce its construction
expenses. The information on streetlight locations may be useful when looking at Smart City
applicationsorhybrid5Gsolutionsthatrequireequipmenttobemountedonstructures.This
mapcanbemadeavailabletotheMasterDeveloperbytheSFPUC.
San Francisco Municipal Transportation Agency Conduit Map

Thisdatasetprovidesgeospatialdataforthefibercableconduitusedbythetrafficsignalnetwork
(SFGo).AllotherinformationonSFMTAconduitiskeptinpaperrecords,whichmaynotcapture
alloftheexistingconduitroutes.IfSFMTAiswillingtoallowaccesstoitsconduit,thisinformation
may be useful to a Master Developer that seeks to use existing conduit space to reduce its
construction expenses. This dataset can be made available to the Master Developer by the
SFMTA.
San Francisco Municipal Transportation Agency Routes and Stops

ThisdatasetprovidesamapSFMTAsrouteandstoplocations.Whilethisdatasetdoesnotshow
conduitorfiberlocations,SFMTAslightrailnetworkhasconduitandfiberalongtheroutesand
mayprovideagoodstartingpointforfurtherinvestigation.Thedatasetalsoidentifieswhethera
stop is sheltered, which may help to identify street furniture where 5G equipment could be
mounted.ThisdatasetcanbefoundondataSFviathelinkbelow.
SFMTARoutesandStopsMap:https://data.sfgov.org/Transportation/SFMTAroutesandstops
forMarch2012/f5c38kkj
135
San Francisco Municipal Transportation Overhead Lines Trolley Pole Database 150813
This dataset provides estimated locations of poles supporting electric distribution wires for
trolleylineshowever,thedataare30yearsold.Itispossiblethatundersomecircumstances
thesepolescouldbeusedtosupportaerialfiberdeployment.Thisdatasetcanbeprovidedtoa
MasterDeveloperbySFMTA.
San Francisco Municipal Transportation Map of Traffic Signals

ThisdatasetprovidesmapsoflocationswheretheCityhastrafficsignals.Thisinformationmay
beusefultoaMasterDeveloperasitprovidesthegenerallocationoftrafficcabinetswhichcould
potentiallyserveashublocationsfortheFTTPnetworkoraSmartCitynetwork.Thetrafficsignal
mapcanbefoundondataSFviathelinkbelow.
TrafficSignalMap:https://data.sfgov.org/Transportation/MapofTrafficSignals/8xtasna8
San Francisco Department of Public Works Street Tree Map

ThisdatasetprovidesamapofthelocationsofDPWownedandmaintainedstreettrees.This
informationmaybeusefultoaMasterDeveloperbecauseknowingthelocationsofthetreeswill
allowtheMasterDevelopertoanticipatepotentialconflictswiththemicrotrenchingconstruction
methodologyorwheretreetrimmingmayberequiredduringaerialconstruction.Additionally,
thetreesmayposeanissuetoanypotentialhybrid5Gsolutionwithequipmentthatrequires
directlineofsight.ThestreettreemapcanbefoundondataSFviathelinkbelow.
DPWStreetTreeMap:
https://data.sfgov.org/browse?q=Street%20Tree%20Map&sortBy=relevance&utf8=%E2%9C%9
3
Additional Data Recommendations

WhiletheCityhassomedataonexistinginfrastructure,theexistingdatasetswouldbenefitfrom
ahigherlevelofdetail.Additionally,thecollectionofdataforseveralkeyutilitieswouldprovide
value.ThefollowingdatawouldbeusefulforaMasterDeveloper.
Collect Utility Pole Data from Other NCJPA Members

Having access to information on the locations of utility poles would be useful for a Master
DeveloperasitwouldhelpthemtoassessoptionsforaerialconstructionwithintheCity.Asa
memberoftheNorthernCaliforniaJointPoleAssociation(NCJPA),theCitycanrequestmapsof
polelocationsfromPG&E,AT&T,andotherowners.TheCitycouldprovidethisinformationtoa
selectedMasterDeveloperunderaNonDisclosureAgreement(NDA).Werecommendthatthe
City work directly with the pole owners to obtain this information if possible. Ideally, the
informationcouldbeprovidedinGISformataspartofacompletedatabaseofspatiallyaccurate
polelocations.However,itispossiblethepolelocationscanonlybeprovidedinpapermapbooks
oranotherformat.AttheveryleasttheCityshouldalerttheNCJPAmemberstothepossibility
136
ofaninfluxofdatarequests,andprepareaprocesstofacilitatethetransferofinformationfrom
thepoleownerstotheselectedMasterDeveloper.Exampleservicerequestscanbefoundin
AppendixB.
Identify ThirdParty Conduit Locations

IdentifyinglocationswherethirdpartyutilityownerssuchasPG&EorAT&Thaveexistingconduit
infrastructurewouldbeusefultoaMasterDeveloperthatmightwanttoleasespareconduit
spacetoreduceitsconstructionexpenses.WerecommendthattheCityrequestmapsidentifying
locationsofleasableconduitroutesfromPG&E,AT&T,andotherentitiesthatownconduitwithin
theCity.ObtainingthisinformationinGISformatwouldbeideal,butitisunlikelydatawillbe
availableinthatformat.TheCitycoulddigitizeanydatasetsprovidedonpapertocreateaGIS
databasethatcouldbeprovidedtoaMasterDeveloper.
Public RightOfWay and CityOwned Property

A Master Developer would be interested in understanding the boundaries for which new
infrastructure could be placed within. Useful information includes the identification of public
rightofway and City easements where conduit and fiber optic cable could be placed.
Additionally,informationidentifyingpublicpropertywherestructuressuchastelecomhutsor
cabinetscanbeplaced(i.e.,tohouseequipmentneededforthedistributionandoperationofthe
network)wouldbeimportanttoaMasterDeveloper.
Available Data
TheCityalreadyhasafairamountofdataavailableidentifyingboundarieswhereinfrastructure
couldbeplaced.Table27identifiesdatasetsthatareavailableforusebyaMasterDeveloper.
Table27:AvailableInformationonROWandPublicProperty
ID# Dataset
3.1.1 SanFranciscoDepartmentofPublicWorksRightofWayandEasementInformation
3.1.2 FacilitySystemofRecord(FSR)
3.1.3 MapofSchools
3.1.4 RecreationandParksProperties
3.1.5 ParcelswithPlanningDepartmentZoning
San Francisco Department of Public Works RightofWay and Easement Information

Thisdatasetprovidesrightofwaypolygonsapportionedintostreetsegmentsandintersection
areas.Itprovidesarepresentationofthestreetandeasementrightofwayareas,butitdoesnot
provideengineeringlevelsofspatialaccuracy.ThisdatacouldbeusedbyaMasterDeveloperto
identifyrightofwayandeasementlocationstohelpinformtheinitialnetworkdesign.Theright
ofwaypolygoninformationcanbefoundondataSFviathelinkbelow.
137
RightofWayPolygons:https://data.sfgov.org/CityInfrastructure/RightofWay
Polygons/a2mggwmg
Facility System of Record (FSR)

ThisdatasetprovidestheCitysbestavailablelistofpublicproperty,includingCityownedand
leased facilities, and facilities owned by nonCity public entities. This dataset would inform a
MasterDeveloperofpossiblelocationsfortheplacementofstructurestohousetheequipment
needed to distribute and operate the FTTP services. Identifying these structure locations is a
crucialpartoftheinitialdesignprocess.TheFacilitySystemofRecordlistcanbemadeavailable
byDT.
Map of Schools
Thisdatasetdetailsthelocationofallschools(Infant,PreK,andK12).Thedatapointscanbe
combined with parcel data to identify school property limits. The school location data point
informationcanbefoundondataSFviathelinkbelow.
MapofSchoolLocations:https://data.sfgov.org/EconomyandCommunity/Mapof
Schools/qb37w9se
Recreation and Parks Properties

This dataset details data points of property owned and maintained by the San Francisco
Recreation and Parks Department. The information includes parks and other types of open
spaceswithintheCity.Thedatapointscanbecombinedwithparceldatatoidentifyproperty
limits.TheRecreationandParkspropertyinformationcanbefoundondataSFviathelinkbelow.
RecreationandParksPropertyLocations:https://data.sfgov.org/CultureandRecreation/Park
andOpenSpaceMap/4udcs3pr
Parcels with Planning Department Zoning

ThisdatasetcontainseveryparcelinSanFranciscoandeachparcelszoning.Thisinformationcan
befoundondataSFviathelinkbelow.
Parcels with Zoning Information: https://data.sfgov.org/CityInfrastructure/ParcelsWith

PlanningDepartmentZoning/6b2nv87s

The available datasets should provide enough information for purposes of the initial network
design. A Master Developer will need to perform additional research as part of the detailed
designengineeringtoensurethatnewinfrastructureisplacedproperlywithinthepublicspace.
138
Potential Network Users

TheCitysgoalistobuildaubiquitouscommunicationsnetworkthatwillprovidefibertoevery
home and business within San Francisco. To achieve this, a Master Developer would need to
understandthequantityandlocationofpotentialusers.
Available Data
TheCityhasafewdatasetsthatwillprovideagoodstartingpointforaMasterDevelopertobegin
theprocessofidentifyingservicelocations.Table28identifiesdatasetsavailableforuse.
Table28:AvailableInformationonPotentialUsers
ID# Dataset
4.1.1 AddresseswithUnitsfromtheCitysEnterpriseAddressingSystem(EAS)
4.1.2 DepartmentofPlanningLandUseMap
4.1.3 TreasurerTaxCollectorRegisteredBusinessLocations

Addresses with Units from the Citys Enterprise Addressing System (EAS)
ThisdatasetlistsallactiveaddressesfromtheCity'sEnterpriseAddressingSystem,includingsub
addresses, such as individual units. The Master Developer could use this dataset to identify
servicelocationsintheCity,whichwillhelpinformtheinitialnetworkdesign.TheEAScanbe
foundondataSFviathelinkbelow.
EAS: https://data.sfgov.org/GeographicLocationsandBoundaries/AddresseswithUnits
EnterpriseAddressingSystem/dxjsvqsy
Department of Planning Land Use Map

ThisdatasetdetailsthelanduseinformationforeachparcelwithintheCityandidentifiesthe
numberofresidentialunitslocatedonthatparcel.Thedatasetdoesnotincludeunregisteredor
illegal residential units, nor does it include any information on nonresidential buildings. This
datasetwillbeusefulbecauseitprovidesthenumberofresidentialunitsforeachparcel;the
MasterDevelopermaybeabletousethesedatainconjunctionwiththeRegisteredBusiness
Locationdatasettoidentifythedensityofusersforagivenarea,whichwillhelpinformtheinitial
networkdesign.TheLandUseMapcanbefoundondataSFviathelinkbelow:
LandUseMap:https://data.sfgov.org/HousingandBuildings/LandUse/us3sfp9q
TreasurerTax Collector Registered Business Locations

This dataset lists the street address and latitudelongitude coordinates of each registered
businessintheCity.Thisdatasetwillbeusefultomapthelocationsofallbusinessesthroughout
theCity.TheMasterDevelopermaybeabletousethisdatasetinconjunctionwithresidential
informationfromtheLandUseMaptoidentifythedensityofusersforagivenarea,whichwill
139
helpinformtheinitialnetworkdesign.TheregisteredbusinesslocationscanbefoundondataSF
viathelinkbelow.
Registered Business Locations: https://data.sfgov.org/EconomyandCommunity/Registered

BusinessLocationsSanFrancisco/g8m3pdis

TheavailabledatasetsshouldbesufficientforaMasterDevelopertoidentifythepotentialusers
andservicelocationsforproposesoftheinitialnetworkdesign.Aspartofthedetaileddesign
engineering,theMasterDevelopershouldconfirmthetotalnumberofservicelocationswhile
performingthesitesurveywork.
Construction Standards and Restrictions

Asthescopeofthisprojectwillrequirecitywideconstruction,itisimportantthattheMaster
Developer understands the Citys approved construction standards and restrictions. This will
allowthemtobetterunderstandthelevelofeffortrequiredfortheprojectandhelpthemto
makeappropriateassumptionsonhowthenetworkcanbebuilt.Importantinformationwould
include approved construction methodologies, standard specifications, identification of
jurisdictionalrestrictions,andpermittingprocesses.
Available Data
Table29identifiesdatasetsthatareavailableforusebyaMasterDeveloper.
Table29:AvailableInformationonConstructionStandardsandRestrictions
ID# Dataset
5.1.1 CaliforniaPublicUtilitiesCommissionGeneralOrderNo.95(GO95)
5.1.2 NorthernCaliforniaJointPoleAssociationOperations/RoutineHandbook2016
5.1.3 San Francisco Municipal Transportation Agency Regulations for Working In San
FranciscoStreets(BlueBook)
5.1.4 DepartmentofPublicWorksStandardSpecificationsandPlans
5.1.5 DepartmentofPublicWorksExcavationRegulations
5.1.6 AT&TGeneralTermsandConditionsforConduitUse
5.1.7 PG&EConduitLicenseAgreement
5.1.8 DepartmentofPublicWorksUtilityExcavationMoratoriumStreetsList
5.1.9 DepartmentofPublicWorksPavementConditionIndex(PCI)ScoresList
5.1.10 CityProjectsforUtilityExcavationandPaving
California Public Utilities Commission General Order No. 95 (GO 95)

This document provides rules and requirements for overhead line design, construction, and
maintenancewithintheStateofCaliforniaandisapplicabletoutilitypoleswithintheCity.The
140
Master Developer could use this information to understand the effort required to construct
aeriallywithintheCity.
AlinktoeachofthesectionsofGO95canbefoundonCaliforniaPublicUtilitiesCommissions
websiteviathefollowinglink:http://www.cpuc.ca.gov/gos/GO95/go_95_section_1.html
Northern California Joint Pole Association Operations/Routine Handbook 2016

Thisdocumentprovidesacomprehensiveoverviewofthepoleattachmentprocessforalljoint
ownedpoleswithintheCity.Informationinthedocumentincludestheattachmentprocessfor
jointowners,attachmentstandards,andauthorizedcosts.TheMasterDevelopercouldusethis
information to understand the pole attachment process and the effort required to construct
aeriallywithintheCity.AsamemberoftheNCJPA,theCityhasaccesstothisdocumentandcan
provideittotheselectedMasterDeveloper.
San Francisco Municipal Transportation Agency Regulations for Working in San Francisco
Streets (Blue Book)
ThisdocumentisamanualforCityagencies(DPW,Muni,SFWD,DPT,PortofSF,etc.),utility
crews,privatecontractors,andothersdoingworkinSanFranciscosstreets.Itestablishesrules
forworkingsafelyandinawaythatwillcausetheleastpossibleinterferencewithpedestrian,
bicycle, transit, and other traffic. The document includes requirements for permitting, traffic
control procedures, and agency contact informationall which would be useful for a Master
Developer.
AlinktoeachofthesectionsofthestandardspecificationdocumentcanbefoundonSFMTAs
websiteviathefollowinglink:https://www.sfmta.com/services/streetssidewalks/construction
regulations
Department of Public Works Standard Specifications and Plans

TheStandardSpecificationsprovideinformationontheCitysrequirementsforvarioustypesof
constructionprojects,includingtypesofconstructionthatwouldoverlapwithfiberconstruction.
ThespecificationsthatwillmostlikelybeusefulforaMasterDeveloperwouldbefoundinthe
electricalwork,excavationandbackfill,andlandscapingworksectionsofthedocument.
AlinktoeachofthesectionsofthestandardspecificationdocumentcanbefoundonDPWs
website via the following link: http://sfpublicworks.org/services/standardsspecificationsand
plans
This webpage also provides links to DPWs standard plans for pullboxes as well as other
informationthatmaybeusefulfortheMasterDeveloper.
141
Department of Public Works Excavation Regulations

ThisgroupofdocumentsprovidescompetedetailontheCitysrulesandregulationsregarding
excavation in the public rightofway. This information will be useful to a Master Developer
becauseacitywideFTTPbuildwillmostlikelyrequireasubstantialamountofexcavationwithin
theCity.
AlinktoeachofthesedocumentscanbefoundundertheExcavationRegulationssectionofthe
ProjectManualandReferenceDocumentspageofDPWswebsite,whichcanbeaccessedviathe
followinglink:http://sfpublicworks.org/services/projectmanualandreferencedocuments
AT&T General Terms and Conditions for Conduit Use

ThisdocumentprovidesthegeneraltermsandconditionsforaccessanduseofAT&Tsconduit
in the City. This information will be useful to a Master Developer because leveraging existing
conduitinfrastructureintheCitymaybeaviablewaytoreducetheoverallprojectcost.TheCity
has access to this document and can provide it to the selected Master Developer. While the
MasterDeveloperwillneedtoobtainitsownagreementwithAT&T,thisdocumentwillprovide
andunderstandingofwhattheMasterDevelopercanexpect.
PG&E Conduit License Agreement

ThisdocumentoutlinestheprocessforapplyingtogainaccesstoPG&Econduit,therulesand
regulationsforuseoftheconduit,andanoverviewoftheestimatedpricingtousetheconduit.
This information will be useful to a Master Developer because leveraging existing conduit
infrastructureintheCitymaybeaviablewaytoreducetheoverallprojectcost.TheCityhas
accesstothisdocumentandcanprovideittotheselectedMasterDeveloper.WhiletheMaster
Developer will need to obtain its own agreement with PG&E, this document will provide and
understandingofwhattheMasterDevelopercanexpect.
Department of Public Works Utility Excavation Moratorium Streets List

This dataset identifies the streets within San Francisco that are under a utility excavation
moratorium. This information may be useful as it will help identify streets where a Master
DeveloperwouldneedtoavoidstreetexcavationorobtainaspecialwaiverfromDPW.AMaster
Developercouldpotentiallyusethisinformationtoinformprojectphasingandotherplanning
considerations.TheutilityexcavationmoratoriumstreetslistcanbefoundondataSFviathelink
below.
Utility Excavation Moratorium Streets List: https://data.sfgov.org/CityInfrastructure/Utility

ExcavationMoratoriumStreets/5wbpdwzt
Department of Public Works Pavement Condition Index (PCI) Scores List

Thisdatasetprovidesinformationontheconditionofthepavementformorethan12,800streets
inSanFrancisco.ThisinformationmaybeusefulasitwillhelpidentifylocationswhereDPWwill
142
bemoreamenabletoundergroundandexcavationwork.AMasterDevelopercouldpotentially
use this information to inform project phasing and other planning considerations. The PCI
spreadsheetcanbefoundondataSFviathelinkbelow.
PCI: https://data.sfgov.org/CityInfrastructure/StreetsDataPavementConditionIndexPCI
Scores/5aye4rtt
City Projects for Utility Excavation and Paving

Thisdatasetprovidesinformationonplannedexcavationandpavementprojects,includingboth
public and private capital/utility projects. This information may be useful as because it could
allowaMasterDevelopertoreduceitscostsbypotentiallycoordinatingconstructioneffortswith
otherongoingprojects.Thisinformationcouldinformthenetworkdesign,projectphasing,and
otherplanningconsiderations.The utilityexcavationandpavingprojectslistcanbefoundon
dataSFviathelinkbelow.
Utility Excavation and Paving Projects List: https://data.sfgov.org/CityInfrastructure/Envista

ProjectsforUtilityExcavationandPaving/sf936dmr

The City has a significant amount of documentation relating to construction standards and
restrictions. Compiling this information into a format that could be easily reviewed and
understood would be beneficial to a Master Developer. Accordingly, we make the following
recommendations.
Develop Standard Specifications for Fiber Construction

TheDepartmentofPublicWorkshasstandardspecificationsanddesignplansforvarioustypes
of construction, including streets and highways, structures, electrical work, excavation and
backfill, and landscaping. While some of these standards overlap with work that would be
conductedinacitywideFTTPproject,manystandardsthatarespecifictofiberconstructionare
not covered in the existing specifications. Developing a single document that clearly defines
approved construction methodologies and practices for this specific project would be very
beneficialtoaMasterDeveloper.
CTCrecommendsthatasetofstandardsbedevelopedspecificallyforfiberconstructionwithin
theCityandthatthedevelopmentofthisdocumentshouldbeacollaborationbetweenDTand
DPW.
Finalize Standards for Microtrenching

TheCityhasapprovedlegislationforanordinanceamendingthePublicWorksCodetoallowthe
useofmicrotrenchingtoinstallfiberopticcablesinthesidewalkportionofthepublicrightof
way.WhilethisdocumentwillprovidesomeguidancetoaMasterDeveloperontheproposed
143
regulationsformicrotrenchingintheCity,therearestillcertainparametersofthisconstruction
methodologythatrequireclarificationandacceptancefromDPW.
CTCrecommendsthattheCityfinalizeoutstandingrequirementsformicrotrenchinganddevelop
asetofDPWapprovedstandardsthataMasterDevelopercoulduseforguidance.Theapproved
microtrenchingstandardscouldbeincludedaspartofthestandardspecificationsdocumentfor
generalfiberconstruction.
144

Table of Available Datasets

ID# Dataset Resource

2.1.1 DepartmentofTechnologyFiberMap BrianRoberts,(415)5814061,Brian.Roberts@sfgov.org
1.1.2 DepartmentofPublicWorksUtilityUndergroundingMap JoyDiFranza,(415)5545813,joy.difranza@sfdpw.org
2.1.3 SanFranciscoPublicUtilitiesCommissionConduit,Pullboxes, SFPUCTeam,streetlights@sfwater.org
andStreetLights
2.1.4 SanFranciscoMunicipalTransportationAgencyConduit RamonZamora,(415)7015668,Ramon.Zamora@sfmta.com
2.1.5 San Francisco Municipal Transportation Agency Routes and dataSF:https://data.sfgov.org/Transportation/SFMTAroutesandstopsforMarch2012/f5c3
StopsforMarch2012 8kkj
2.1.6 San Francisco Municipal Transportation Overhead Lines TeePhang,(415)7014239,Tee.Phang@sfmta.com
TrolleyPoleDatabase150813
2.1.7 SanFranciscoMunicipalTransportationMapofTrafficSignals dataSF:https://data.sfgov.org/Transportation/MapofTrafficSignals/8xtasna8
2.1.8 DepartmentofPublicWorksMapofStreetTrees dataSF:
https://data.sfgov.org/browse?q=Street%20Tree%20Map&sortBy=relevance&utf8=%E2%9C%93
3.1.1 DepartmentofPublicWorksROWandEasementInformation dataSF:https://data.sfgov.org/CityInfrastructure/RightofWayPolygons/a2mggwmg
3.1.2 FacilitySystemofRecord(FSR) BrianRoberts,(415)5814061,Brian.Roberts@sfgov.org
3.1.3 MapofSchools dataSF:https://data.sfgov.org/EconomyandCommunity/MapofSchools/qb37w9se
3.1.4 RecreationandParksProperties dataSF:https://data.sfgov.org/CultureandRecreation/ParkandOpenSpaceMap/4udcs3pr
3.1.5 ParcelswithPlanningDepartmentZoning dataSF:https://data.sfgov.org/CityInfrastructure/ParcelsWithPlanningDepartment
Zoning/6b2nv87s
4.1.1 Addresses with Units from the Citys Enterprise Address dataSF:https://data.sfgov.org/GeographicLocationsandBoundaries/AddresseswithUnits
System(EAS) EnterpriseAddressingSystem/dxjsvqsy
4.1.2 DepartmentofPlanningLandUseMap dataSF:https://data.sfgov.org/HousingandBuildings/LandUse/us3sfp9q
4.1.3 TreasurerTaxCollectorRegisteredBusinessLocations dataSF: https://data.sfgov.org/EconomyandCommunity/RegisteredBusinessLocationsSan
Francisco/g8m3pdis
145

ID# Dataset Resource

5.1.1 California Public Utilities Commission General Order No. 95 CPUCWebsite:http://www.cpuc.ca.gov/gos/GO95/go_95_section_1.html
(GO95)
5.1.2 NorthernCaliforniaJointPoleAssociationOperations/Routine BrianRoberts,(415)5814061,Brian.Roberts@sfgov.org
Handbook2016
5.1.3 San Francisco Municipal Transportation Agency Regulations SFMTAWebsite:https://www.sfmta.com/services/streetssidewalks/constructionregulations
forWorkingInSanFranciscoStreets(BlueBook)
5.1.4 DepartmentofPublicWorksStandardSpecificationsandPlans DPWWebsite:http://sfpublicworks.org/services/standardsspecificationsandplans
5.1.5 DepartmentofPublicWorksExcavationRegulations DPWWebsite:http://sfpublicworks.org/services/projectmanualandreferencedocuments
5.1.6 AT&TGeneralTermsandConditionsforConduitUse BrianRoberts,(415)5814061,Brian.Roberts@sfgov.org
5.1.7 PG&EConduitLicenseAgreement BrianRoberts,(415)5814061,Brian.Roberts@sfgov.org
5.1.8 Department of Public Works Utility Excavation Moratorium dataSF: https://data.sfgov.org/CityInfrastructure/UtilityExcavationMoratoriumStreets/5wbp
StreetsList dwzt
5.1.9 DepartmentofPublicWorksPavementConditionIndex(PCI) dataSF: https://data.sfgov.org/CityInfrastructure/StreetsDataPavementConditionIndexPCI
ScoresList Scores/5aye4rtt
5.1.10 CityProjectsforUtilityExcavationandPaving dataSF: https://data.sfgov.org/CityInfrastructure/EnvistaProjectsforUtilityExcavationand
Paving/sf936dmr
146
ThirdParty Data Requests
Data Request to PG&E

PG&E has a large amount of infrastructure in the City including utility poles and conduit that
couldbeleveragedinaCitywidefiberbuild.Asapowercompany,PG&Edoesnotprovideservices
thatwouldcompetewiththeCitysproposedFTTPnetworkandshouldbemoreamenabletothe
CitysdatarequeststhentelecomproviderssuchasAT&T.WiththecurrentNCJPAguidelines,
theburdenfordatacollectiongenerallyfallstotheentitylookingtoattachasopposedtothe
poleowner.BecausePG&Eisnotusuallyrequiredtofacilitateaccesstotheirdata,theoddsare
low that they would be willing to freely provide information without incentive. It could be
expectedthatthemoreeffortrequiredbyPG&Etoprovidethedata,thelesslikelytheywillbe
willingtoprovidethedata.
Obtaining from a power company is not impossible, an awardee of a Broadband Technology

OpportunitiesProgram(BTOP)grant,HowardCountyMarylandwasabletosuccessfullyobtain
geospatial data containing utility pole locations for the entire footprint of their local power
company, Baltimore Gas and Electric. This allowed the County to identify aerial attachment
opportunitiesandmakeinformedengineeringdecisionsthatbenefitedtheirfiberconstruction
project.
Priortosendingthedatarequest,theCityshouldidentifytheproperpointofcontact(s)tohandle
the request. PG&E will most likely have separate contacts for utility poles and underground
infrastructure.Thedatarequestshouldincludeanoverviewoftheproposedprojectandstate
theCityswillingnesstocomplywithnondisclosurerequirements.
TheCityshouldrequestthefollowinginformation:
#1 UtilityPoleLocations
Geospatial data identifying the location of all utility poles are located within the City should be
requested.ThedatashouldbeprovidedinGISformatifpossibletoallowforeasytransferandreview
of the information. Though not preferred, paper maps of the pole locations would be better than
receivingnodata.PapermapsmayrequiremoreeffortfromPG&Etoprovidetheinformation,sothe
CitymayconsiderrequestingdataforspecificsectionsoftheCitytoincreasethelikelihoodofobtaining
thedata.

#2 ConduitLocations
Geospatial data identifying the locations of available conduit routes should be requested. The data
shouldbeprovidedinGISformatifpossible.Thoughnotpreferred,apapermapofthegenerallocation
ofavailableconduitwithintheCitywouldbebetterthanreceivingnodata,asitcouldprovideastarting
pointforfurtherinvestigation.

147
#3 UtilityPoleInformation
SpecificinformationonPG&Espolesshouldberequestedtohelpwithanalyzingthefeasibilityofaerial
construction.Usefuldatawouldincludethefollowing:

Individualpoleinformationincluding;poleclass,height,age,material,compliance
Currentattachmentsandloadinganalysisforindividualpoles
PG&Espolereplacementandmakereadyschedule

ItispossiblePG&EhassomeoftheindividualpoledatainGISformatwhichwouldbepreferred,though
theyshouldhavepaperrecordsofthisdataasthisinformationisprovidedtothemeverytimeanew
attachmentismade.

#4 ConduitInformation
SpecificinformationonPG&Esavailableconduitwouldprovideadditionalbenefit.Usefuldatawould
includethefollowing:

Conduitsize
Currentoccupancyandavailablespace
Accesspoints(i.e.handholelocations,buildingentrances)

PG&EmayhavesomeoftheconduitattributesinGISformatwhichwouldbepreferred,thoughitis
morelikelytheyhavepaperrecordsofthisinformation.

#5 AbandonedGasLines
Geospatialdataidentifyingthelocationsofabandonedgaslinecouldberequested.Whenpermitted
by the utility owner, fiber can be placed in the abandoned lines to avoid new fiber construction.
Abandonedgaslinestendtobelessusefulbecausethelackofaccesspointsandgeneralplacementof
theutilitymakeaccessingthefiberdifficult.IdeallythisinformationcouldbeprovidedinGISformat,
howeverpapermapswouldbeacceptable.

Data Request to AT&T

AT&ThasalargeamountofinfrastructureintheCityincludingutilitypolesandconduitthatcould
be leveraged in a Citywide fiber build. It is possible that AT&T could see the Citys proposed
networkasacompetitorandreluctanttoprovideinformationtotheCity.WiththecurrentNCJPA
guidelines,theburdenfordatacollectiongenerallyfallstotheentitylookingtoattachasopposed
tothepoleowner.BecauseAT&Tisnotusuallyrequiredfacilitateaccesstotheirdata,theodds
arelowthattheywouldbewillingtofreelyprovideinformationwithoutincentive.Itcouldbe
expectedthatthemoreeffortrequiredbyAT&Ttoprovidethedata,thelesslikelytheywillbe
willingtoprovidethedata.
the request. AT&T will most likely have separate contacts for utility poles and underground
148
infrastructure.Thedatarequestshouldincludeanoverviewoftheproposedprojectandstate
theCityswillingnesstocomplywithnondisclosurerequirements.
#1 UtilityPoleLocations
Geospatial data identifying the location of all utility poles are located within the City should be
requested.ThedatashouldbeprovidedinGISformatifpossibletoallowforeasytransferandreview
of the information. Though not preferred, paper maps of the pole locations would be better than
receivingnodata.PapermapsmayrequiremoreeffortfromAT&Ttoprovidetheinformation,sothe
CitymayconsiderrequestingdataforspecificsectionsoftheCitytoincreasethelikelihoodofobtaining
thedata.

#2 ConduitLocations

#3 UtilityPoleInformation
SpecificinformationonAT&Tspolesshouldberequestedtohelpwithanalyzingthefeasibilityofaerial
construction.Usefuldatawouldincludethefollowing:

Individualpoleinformationincluding;poleclass,height,age,material,compliance
Currentattachmentsandloadinganalysisforindividualpoles
AT&Tspolereplacementandmakereadyschedule

ItispossibleAT&ThassomeoftheindividualpoledatainGISformatwhichwouldbepreferred,though
theyshouldhavepaperrecordsofthisdataasthisinformationisprovidedtothemeverytimeanew
attachmentismade.

SpecificinformationonAT&Tsavailableconduitwouldprovideadditionalbenefit.Usefuldatawould
includethefollowing:

Conduitsize
Informationonthetermsandconditionstoincludeanyapplicableservicerestrictions

AT&TmayhavesomeoftheconduitattributesinGISformatwhichwouldbepreferred,thoughitis

149
Data Request to Other Providers

OtherproviderssuchasZayo,CenturyLink,andCrownCastlehaveundergroundinfrastructurein
theCitythatcouldbeleveragedinaCitywidefiberbuild.Itispossiblethattheseentitieswould
be willing to provide information of their available conduit routes, and may offer additional
benefitinreturnsuchasjointbuildopportunitiesduringthecitywidefiberconstruction.
therequest.Thedatarequestshouldincludeanoverviewoftheproposedprojectandstatethe
Cityswillingnesstocomplywithnondisclosurerequirements.
#1 ConduitLocations

Specificinformationontheprovidersavailableconduitwouldprovideadditionalbenefit.Usefuldata
wouldincludethefollowing:

Conduitsize
Informationonthetermsandconditionstoincludeanyapplicableservicerestrictions
Perfootpricingforleasedconduit

ProvidermayhavesomeoftheconduitattributesinGISformatwhichwouldbepreferred,thoughitis

150
Appendix G: Potentially Usable City Facilities

CityOwned Fiber Infrastructure
TheCitysexistingfiberinfrastructurehaspotentialtobeusedforthisproject.Theinfrastructure
comprisesconduit,handholes,andpoleattachmentswhichareallbuilttoaccommodatefiber
deployment.Asanexample,theprojectcouldleverageadditionalconduitspaceintheexisting
undergroundfiberroutestoreduceconstructioncosts.InareaswheretheCityhasaerialfiber,
new fiber can be overlashed to the City attachments instead of requiring a new attachment,
whichmayreduceoreliminatemakereadywork.
Figure17:CityFiberFacilities
151
CityOwned Property
Citypropertycanprovideaplacetolocatenetworkhubsites,fiberdistributioncabinets,orother
infrastructure.Thesampledesignestimatestheneedforapproximately10hublocationsspread
throughouttheCity.Eachhubwillrequireapproximately200feetofspace,power,secureaccess,
andenvironmentalcontrols.Thehubscanbeplacedinbuildingsthatmeettheserequirements,
orintelecomhutsbuiltonCityproperty.Fiberdistributioncabinetsaremuchsmallerthanhubs
andcanbeplacedintheROWorbemountedonutilitypoles.However,itmaybeidealtoplace
thecabinetsonCitypropertywhenfeasibletokeeptheROWclearofadditionalcongestion.
Figure18:PropertyOwnedbyCityandCounty
152
Appendix H: Digital Inclusion and Broadband Affordability

TheBroadbandforSanFranciscoInitiativeisintendedtodevelopnewmeansofexpandingnext
generation,worldclassbroadbandinfrastructureintheCity.Thefollowingisananalysisofsome
considerationsregardingdigitalinclusionintheCity.
Thissectionnotesthataffordabilityisonecriticalchallengeinensuringtheuniversaladoptionof
broadbandservices,despitetheapparentuniversalavailabilityoffixedandmobilebroadbandin
San Francisco, in part because of the uneven availability of lowerpriced products that are
ostensiblytargetedtowardlowincomeresidentsoftheCity.Affordabilityofbroadbandservices
isbynomeanstheonlyfactorimpactingadoptionrates,butaddressingthatissueisonewaythat
theCitycanimproveratesofuseofhighspeedbroadbandbylowerincomemembersofthe
community.Complementaryinitiativesfocusedondeviceownershipandinternetliteracywillbe
necessarytosupplementtheCityseffortstomakehighspeedbroadbandaffordable.
Low Broadband Adoption in San Francisco Correlates with LowerIncome Areas

RecentdatafromtheFCCsuggestsacorrelationbetweenlowerbroadbandadoptionratesand
lowerincome areas in the City. While the data do not speak to causation, the correlation is
importantwhenconsideringbarrierstobroadbandadoption.
Twiceayear,internetserviceproviders(ISPs)submitdatatotheFCCviaForm477,reporting
wherewirelinebroadbandserviceisavailableandthepercentofhouseholdssubscribingtotheir
service.105ThisinformationisthenaggregatedbytheFCCandreleasedtothepublicroughlyone
yearaftercollection.Thedataillustratethepercentageofhomeswithminimuminternetspeeds
withinagivencensustract.
Censustractsaredefinedbygeographicalboundaries(ratherthanpopulation),andcancontain
from1,200to8,000people.106Asaresult,itisimpossibletoobtainadefinitivenumberofhow
manyhouseholdssubscribetointernetservicesfromForm477data.Rather,themostrecent
availabledata(fromDecember31,2015)showthepercentofhouseholdsreceivingaserviceof
aminimum200Kbpsoneway,andaminimum10Mbps/1Mbps(upstream/downstream)within
196censustractsintheCity.
Table30providesabreakdownofcensustractsbypercentsubscriptionrate.107
105
Entitiesrequiredtosubmitthesedataincludefacilitiesbasedprovidersofbroadbandconnectionsand
providersofwiredorfixedwirelesslocalexchangetelephoneservice.See:WhoMustFileForm477?,Federal
CommunicationsCommission,https://transition.fcc.gov/form477/WhoMustFileForm477.pdf(accessedMarch
2017).
106
GeographicTermsandConceptsCensusTract,U.S.CensusBureau,
https://www.census.gov/geo/reference/gtc/gtc_ct.html(accessedMarch2017).
107
InternetAccessServiceReports,FederalCommunicationsCommission,https://www.fcc.gov/reports
research/reports/internetaccessservicesreports/internetaccessservicesreports(accessedMarch2017).
153
Table30:NumberofCensusTractsbySubscriptionRate
CensusTractswith
Percentof CensusTractswithMinimum
10MbpsUpstream/
Households 200KbpsOneWay
1MbpsDownstream
0 1 1
1to20% 0 2
21to40% 4 4
4160% 4 37
6180% 44 130
81%+ 143 22
Themajorityofcensustractshaveasubscriptionrateof80percentoraboveforminimum200
Kbpsservice,anda60to80percentsubscriptionrateforminimum10Mbps/1Mbps.Whilethese
adoptionratesareslightlyhigherthanthenationwideaverage,108itshouldbenotedthatneither
ofthesespeedsmeetthecurrentFCCdefinitionofbroadband,whichis25Mbps/3Mbps.109
WhensuperimposedoveramapoftheCity,thedataillustratethegeographicalareaswithlower
subscriptionrates.Figure19shows200Kbpsadoptionratesbycensustract,110andFigure20
shows10Mbps/1Mbpsadoptionratesbycensustract.111Inthesemaps,darkbluerepresents
areaswithhigheradoptionrates,whilelightblueandwhiterepresentareaswithloweradoption
rates.
108
"report_dec_2015_body_17NOV16.doc,,"downloadedfile,FederalCommunicationsCommission,
https://www.fcc.gov/file/11764/download(accessedMarch2017).
109
2015BroadbandProgressReportandNoticeofInquiryonImmediateActiontoAccelerateDeployment,
FederalCommunicationsCommission,January29,2015(adopted),
https://apps.fcc.gov/edocs_public/attachmatch/FCC1510A1.pdf(accessedMarch2017).
110
Form477CensusTractDataonInternetAccessServices,FederalCommunicationsCommission,
https://www.fcc.gov/general/form477censustractdatainternetaccessservices(accessedMarch2017).
111
Ibid.
154
Figure19:PercentofHouseholdswithMinimum200KbpsService
155
Figure20:PercentofHouseholdswithMinimum10Mbps/1MbpsService

Inbothfigures,themapshowsaclearcrescentshapedareaoflowersubscriptionrates,passing
fromtheareanorthofChinatown,throughtheMissiondistrict,andthenbacksoutheasttothe
Bayviewdistrict.Thereisalsoapocketoflowsubscriptionincensustractsinthesouthofthe
City.
Figure21illustratescensustractsbyyearlyhouseholdincome,withdarkbluerepresentinghigher
average household income. 112 This map illustrates the same crescent shape and southern
pocketan illustration of lowincome areas that correlates with low broadband adoption
illustratedinFigure20.Theclearcommonalitiesofthegeographicdistributioninthetwodata
112
DataUSA,https://embed.datausa.io/profile/geo/sanfranciscoca/economy/income_geo/?viz=True(accessed
March2017).
156
setsdemonstratethatareasoftheCitywithlowhouseholdincomealsohavelowersubscription
rates.
Figure21:YearlyHouseholdIncomebyCensusTract

Cost Is Among the Key Factors Driving Lower Internet Adoption Rates Nationally
With 13.2 percent of the Citys residents living below the poverty level, 113 affordability of
broadbandservicesremainsasignificantfactorintheCitysdigitalinclusionefforts.Formany
Americanhouseholdsthathavenotadoptedbroadbandservices,decisionsregardinginternet
accessareoftenachoicebetweenhavinginternetserviceandhavingfood.114Ina2015Pew
ResearchCentersurvey,33percentofrespondentswhodidnothaveinternetaccessathome
citedthehighcostofserviceasthemainreason.115
113
QuickFacts:SanFranciscoCounty,California;SanFranciscocity,California,U.S.CensusBureau,DataasofJuly
1,2016,https://www.census.gov/quickfacts/table/PST045216/06075,0667000(accessedMarch2017).
114
ColinRhinesmith,DigitalInclusionandMeaningfulBroadbandAdoptionInitiatives,BentonFoundation,January
2016,https://www.benton.org/sites/default/files/broadbandinclusion.pdf(accessedMarch2017).
115
HomeBroadband2015,PewResearchCenter,December21,2015,
http://www.pewinternet.org/2015/12/21/homebroadband2015(accessedMarch2017).
157
Althoughthedigitaldividehasnarrowedoverthecourseofthelastdecade,manylowincome
individuals remain unable to access the internet from home, hindering their ability to reach
economicandeducationalresourcescriticalforsuccessinthe21stcentury.Accordingtoarecent
PewResearchCentersurvey,47percentofAmericanswithhouseholdincomesunder$30,000
per year lack home broadband service, compared with 6 percent among individuals with
householdincomesover$100,000peryear.116
TheCityanalyzeddigitaldivideissuesin2007andagainin2015.Inthelatterstudy,theBoardof
Supervisorsbudgetandlegislativeanalystreportedonastudythatfoundthat12percentofCity
residents lack internet access at home (and another 6 percent reported using dialup
connections).Whilethestudynotedthat75percentofhouseholdswithincomesoflessthan
$25,000hadhomeinternetaccessasignificantlyhigherpercentageoflowincomeresidents
thanthenationalaverageaffordabilityisstillclearlyanissue.117The2007studyalsoindicated
thataffordabilitywasafactorinhomeinternetuse.118
Whilethesenumbersarebetterthanthenationalnumbers,thereisstill,asofthedateofthese
studies,aremarkablegapbetweenlowerandhigherincomeresidentsoftheCityintermsof
homeinternetuse.Despitethebarriers,eachyearmoreandmorelowincomeindividualsare
findingwaystoaccessonlineresources.Asof2016,79percentofadultslivingwithhousehold
incomes under $30,000 reported using the internet 119 and 64 percent reported having a
smartphone. 120 Even those without home broadband service increasingly acknowledge that
accesstotheinternetiscriticalinmanykeyareasoflife.121
116
MonicaAnderson,DigitaldividepersistsevenaslowerincomeAmericansmakegainsintechadoption,Pew
ResearchCenter,March22,2017,http://www.pewresearch.org/facttank/2017/03/22/digitaldividepersistseven
aslowerincomeamericansmakegainsintechadoption/(accessedMarch,2017).
117
FredBrousseau,PolicyAnalysisReport:DigitalDivideinSanFrancisco,CityandCountyofSanFranciscoBoard
ofSupervisors,April15,2015,http://sfbos.org/sites/default/files/FileCenter/Documents/52249
BLA.DigitalDivide.041515.pdf(accessedMarch2017).
118
ThomasBerman,MatthewCase,JoshuaDaniels,MaraLarsenFleming,andSunHaLee,DigitalInclusioninSan
Francisco,UniversityofCalifornia,Berkeley,andSanFranciscoDepartmentofTelecommunicationsand
InformationServices,June2007,
http://sfgov.org/dt/ftp/uploadedfiles/dtis/tech_connect/UCBDigitalInclusionReport060282007Final%281%29.pdf
(accessedMarch2017).
119
Internet/BroadbandFactsheet,PewResearchCenter,January12,2017,http://www.pewinternet.org/fact
sheet/internetbroadband/(accessedMarch2017).
120
ResearchCenter,March22,2017,http://www.pewresearch.org/facttank/2017/03/22/digitaldividepersists
evenaslowerincomeamericansmakegainsintechadoption/(accessedMarch2017).
121
JohnHorriganandMaeveDuggan,HomeBroadband2015,PewResearchCenter,December21,2015,
http://www.pewinternet.org/2015/12/21/homebroadband2015/pi20151021_broadband201503/(accessed
March2017).
158
For those who lack home internet, some make use of free broadband services at libraries,
communitycenters,andprivatebusinesses,whileothersrelyontheirmobiledevices.Roughly
onefifth of individuals living with household incomes under $30,000 report having a
smartphone,butnointernetathome,upfrom12percentin2013.122Withfewotheroptions,
lowincomehouseholdsdisproportionallyrelyonsmartphonestocompletetasksthatwealthier
counterpartscompleteonlaptopsanddesktops.Themajorityoflowincomesmartphoneowners
reportusingtheirmobiledevicetolookupinformationaboutajob,andathirdreportusingtheir
smartphonetosendinjobapplications.123
So as the internet becomes increasingly relevant to the lives of lowincome individuals, what
keepsthemfromsigningupforhomebroadbandservice?Whilenosinglefactorcanexplainthe
persistenthomebroadbandgap,thehighcostofmonthlysubscriptionsisundoubtedlyamajor
contributor.Whenindividualswithouthomebroadbandwereaskedtheirmostimportantreason
fornothavingahomebroadbandconnection,33percentsaidthatmonthlyhomebroadband
subscriptionsaretooexpensive,morethantwicetheresponsethatanyotherreasonreceived.124
Arecentstudyofbroadbandadoptioninitiativesfoundthat,whilealllowincomeindividualsand
familieswhoparticipatedinthisstudyunderstoodthevalueofbroadbandconnectivity,many
wereoftenforcedtochoosebetweenhavinginternetandhavingenoughfoodforthemonth.125
In a study of individuals who have dropped home broadband service, all individuals with
householdincomeupto$40,000peryearcitedtheservicebeingtooexpensiveastheprimary
reasonwhytheydroppedservice.126
Broadband Service Costs in San Francisco Are Unevenly Distributed, Unpredictable, and,
for Some, Unaffordable
CurrentdataproductsavailableintheCityreflectunpredictablelayersofcost.
122
ResearchCenter,March22,2017,http://www.pewresearch.org/facttank/2017/03/22/digitaldividepersistseven
aslowerincomeamericansmakegainsintechadoption/(accessedMarch2017).
123
AaronSmith,U.S.SmartphoneUsein2015,PewResearchCenter,April1,2015,
http://www.pewinternet.org/2015/04/01/chaptertwousageandattitudestowardsmartphones/#job
percent20seeking(accessedMarch2017).
124
JohnHorriganandMaeveDuggan,HomeBroadband2015,PewResearchCenter,December21,2015,
http://www.pewinternet.org/2015/12/21/homebroadband2015/pi20151021_broadband201512/(accessed
March2017).
125
ColinRhinesmith,DigitalInclusionandMeaningfulBroadbandAdoptionInitiatives,BentonFoundation,
January2016,https://www.benton.org/sites/default/files/broadbandinclusion.pdf(accessedMarch2017).
126
BrianWhitacreandColinRhinesmith,UnderstandingBroadbandUnadopters,BentonFoundation,January
12,2016,https://www.benton.org/blog/understandingbroadbandunadopters(accessedMarch2017).
159
Wireline Data Products

Comcastadvertisesnocontractresidentialinternetserviceswith25Mbpsmaximumdownload
speedfor$49.99permonth.Whileitsadvertisedpricealonemaybeprohibitivetolowincome
citizens,theplancomeswithfurtherhiddencharges:
Activationfeeupto$500(totalpricenotdiscloseduntilaftersignup)
Promotionalpriceincreasesto$59.99$64.99permonthafterfirstyear(totalprice
notdiscloseduntilafterfirstyear)
Plansubjecttoearlyterminationfee(pricenotdisclosed)
Plan price does not include equipment charges, which can be either a substantial
upfrontcost,ormonthlynominalfees
Planallows1TBofdatausepermonth,afterwhicheach50GBincrementwillcost
$10permonth(nottoexceed$200monthly)
Allfees,otherthanthecontractedmonthlysubscriptionfee,aresubjecttochangeat
anytime127
In the first month of service alone, pricing stipulations of at least $500 in costs additional to
servicefeearelikelytoserveasobviousbarrierstoadoptionbylowincomeresidents.
Mobile Wireless Data Products

Forsubscriberswhodonotpurchasewirelinedataservicesathome,dataonlymobilebroadband
servicesprovidedbymajorwirelesscarriersareapotentialalternative.However,thehighcost
of connectivity and adequate capacity can prove to be deterrents to even moderateincome
subscribers.Similartotheirwiredcounterparts,mobilebroadbandplansarefrequentlyrifewith
upchargesandunexpectedcosts.Unlikemostwiredplans,manymobileserviceschargebythe
amount of data transmitted, limiting their feasibility for highcapacity applications such as
transmission of large files, videoconferencing, and extended content and mediarich web
browsing.
ThecostofanentrylevelmobiledataproductfromVerizonbeginsat$20permonthfor2GBof
data.Pricesincrease$10permonthper2GBadditionaldata.Furtherchargesinclude:
$5permonthperdeviceconnectedtotheplan
$10permonthpertabletconnectedtotheplan
127
Selectthepackagethatsperfectforyou,ComcastXfinity,https://www.xfinity.com/locations/internet
service/california/sanfrancisco.html(accessedMarch2017).
160
$20permonthperhotspotconnectedtotheplan
$15pereachadditional1GBofdata
Costofdevice,tablet,orhotspot
Subscriberswhodonotwanttobechargedoveragefeescanchoosetobethrottledonceplan
dataallowanceshavebeenused,slowingconnectionspeedsto128Kbpswellbelowbroadband
speeds,andlimitingfunctionality.128
AT&Talsoofferswirelessinternetplans,characterizedbyhighpricepointsandalimitedamount
ofhighspeeddata.Forthoseinterestedinadataonlyplan,AT&Thasa250MBplanfor$14.99
monthly, which includes unlimited access to AT&T WiFi hotspots distributed around major
metropolitan areas. This plan does not allow customers to tether another device, so all
browsing must occur on the mobile device on the plan. If the customer exceeds the 250 MB
allowance,eachadditional250MBusedcostsanadditional$14.99.Highercapacity3GBand5
GB plans can be purchased for $30 and $50 per month, respectively, and come with similar
overagefees.Forcustomerswhowishtotetheranotherdevice(suchasahomecomputer,an
additional tablet, or phone other than that on the plan), customers must purchase the $50
plan.129
Recently,AT&Tannouncedanunlimitedplanthatincludesunlimitedtalk,text,anddata.For
$90permonth,theplanallows22GBofdataat4GLTEspeeds,afterwhichAT&Tthrottlesthe
userto128Kbps.Forthoseinterestedintetheringtheirhomecomputeroranotherdeviceto
thisplan,only10GBoftheinitial22GBwillbe4GLTEspeedsfortetheringormobilehotspot
use. Like the above plans, the advertised price is lower than the amount customers will be
charged in the first few months. To receive $90 monthly pricing, customers must enroll in
autopay(withacreditcard),paperlessbilling,andallow13monthsofbillingcyclesbefore
discountswillbeapplied.130Giventheseconstraints,thisproductmaynotbeusefulforlower
incomeSanFranciscans.
TMobilealsooffersprepaidmobile,dataonlyplansforcustomerswholacksufficientcreditto
qualifyformonthlyplans,orwhodonotwishtosignupformultiyearcontracts.TMobilespay
asyougoplancosts$3permonth,butcustomersthenmustregularlypurchasedatapassesata
rateof$5perdayinclusiveof500MBof4Gdata,or$10perweekfor1GBof4Gdata.131The
128
DataOnlyPlan,VerizonWireless,https://www.verizonwireless.com/plans/dataonlyplan(accessedMarch
2017).
129
SelectaPlan,AT&T,https://www.att.com/shop/wireless/plans/planconfigurator.html(accessedMarch2017).
130
UnlimitedData.UnrealDeal,AT&T,https://www.att.com/plans/unlimiteddataplans.html(accessedMarch
2017).
131
PayAsYouGo,TMobile,https://prepaidphones.tmobile.com/payasyougo(accessedApril2017).
161
companyalsooffersmonthlyprepaidplansatarateof$45permonthfor4GBof4Gdata,$55
permonthfor6GBof4Gdataor$75forunlimited4Gdata.132However,theunlimitedplans
fineprintcontainsanumberofrestrictions,includinganadditional$15tetheringfeeforevery5
GBof4Gdatausedontethereddevices.133
Sprinthasprepaid3G/4G(dependingonavailability)mobiledataplansaswell,rangingfrom$15
perdayinclusiveof150MBofdata,$30perweekinclusiveof500MBofdata,and$50permonth
inclusiveof1.5GBofdata.134Theyalsohaveapostpaidunlimitedplanfor$60permonth,butit
requiresacreditcheck,andonceagain,theunlimitedplanscomewithanumberofrestrictions
inthefineprint,includinga10GBlimitontethereddatause.135
AT&T,TMobile,andSprintalsohavetheirownprepaidbrandsthatmarketthemselvestolower
incomecustomers.AT&TsCricketWirelessoffersmonthlyplansatarateof$30for1GB,$40
for3GB,$50for8GB,and$60forunlimiteddata(throttledafter22GB);however,theseplans
are restricted to 8 Mbps when a user is on AT&Ts 4G LTE network, and to 4 Mbps when on
AT&TsHSPA+network.136
TMobiles MetroPCS offers monthly plans at a rate of $30 for 1 GB, $40 for 3 GB, $50 for
unlimitedLTEdataonasinglehandset,and$60forunlimitedLTEdataonasinglehandsetplus
8GBoftethereddata.137
MetroPCScustomersenjoyspeedsclosetothespeedsofTMobilecustomers.Inarecenttest,
MetroPCSboastedanaveragedownloadspeedof22Mbpsand uploadspeedof 16Mbps.138
Sprinthastwoprepaidbrands,VirginMobileUSAandBoostMobile.VirginMobilehasmonthly
planspricedat$35for5GBof4GLTEdataand$45for10GB,witheachadditionalGBofdata
costing$5.BoostMobileoffersa$35monthlyplanwith2GBofdata,witheachadditionalGBof
132
PrepaidPhones,TMobile,https://prepaidphones.tmobile.com/prepaidplans
individual?icid=WEB__Q117PREPAI_YYP087QVHNJ7801#gettingstarted(accessedApril2017).
133
JacobKastrenakes,TMobileandSprint'snewunlimitedplanshavesomenastyfineprint,TheVerge,August
18,2016,http://www.theverge.com/2016/8/18/12533308/tmobilesprintunlimiteddataplanslimitvideo
(accessedApril2017).
134
Services,Sprint,
http://shop.sprint.com/mysprint/services_solutions/details.jsp?detId=mbb_passes&catId=service_mbb&catName
=Mobile%20Broadband&detName=Mobile%20Broadband%20Passes&specialCat=(accessedApril2017).
135
Unlimited,Sprint,https://www.sprint.com/landings/unlimitedcellphoneplans/(accessedApril2017).
136
Plans,Cricket,https://www.cricketwireless.com/support/plansandfeatures/cricketplansand
features/customer/plans.html(accessedApril2017).
137
Plans,MetroPCS,https://www.metropcs.com/shop/plans(accessedApril2017).
138
PhilipMichaels,MetroPCSIsFastestDiscountCarrier,StraightTalkSlowest,TomsGuide,March28,2017,
http://www.tomsguide.com/us/discountcarriersperformance,news24762.html(accessedApril2017).
162
datacosting$5.Italsohasanunlimitedplanfor$50permonth,butagain,tetheringislimitedto
8GBofdata.139
Inadditiontothefourmajorcarriers,thereareanumberofmobilevirtualnetworkoperators
(MVNOs)thatofferlowcost,nocontract,dataplans.GooglesProjectFioffersa$20permonth
plan,wherecustomerspay$10perGBofdatatheyuse.140TingMobileoffersplansunderwhich
consumerspayonlyforasmanyvoiceminutes,texts,andGBofdataastheyuseeachmonth,
with a $6 per line base rate and each GB of data used costing $10. 141 Republic Wireless has
monthlyplansthatofferunlimitedvoiceminutesandtextsatarateof$20for1GB,$30for2
GB, $45 for 4 GB, and $60 for 6 GB of cellular data. 142 These MVNOs rely on partnership
agreements with major carriers, and therefore speeds and availability are dependent on the
strengthoftheirpartnercarriersnetworkinagivenarea.Giventhatmuchofthenewinvestment
inhighspeedmobilenetworksisconcentratedinhighincomeareas,individualslivinginlower
incomeneighborhoodshaveaccesstoslowerthanaverageservicespeeds.
Mobile Broadband Services Represent an Inadequate Alternative to Fixed Services

While mobile data plans may be helping to bridge the divide for now, they have distinct
disadvantagescomparedtowiredbroadbandsubscriptions.Lowcostmobiledataplansareoften
metered,withstrictdatacaps.ThriftymobileuserslearntowaituntiltheyareconnectedtoWi
Fi to download large files, or use dataintensive applications, such as videostreaming or
teleconferencing.Thosewholivewith,orincloseproximitytoavailableWiFinetworkscanpay
substantiallylessformobiledatathansomeonewhousesthesameamountofdata,butrelies
exclusivelyontheirmobilecarriersnetwork.
Thedemandformobiledataisskyrocketing,143andinmanyareas,carriersarestrugglingtokeep
up.144Ifthefederalgovernmentfailstofreeupsubstantialamountsofnewspectruminthenext
few years, the U.S. could face a serious shortage of spectrum available for commercial data
139
ShopPlans,BoostMobile,https://www.boostmobile.com/#!/shop/plans/monthlyphoneplans/(accessed
April2017).
140
ProjectFi:Asimpleplanforeveryone,Google,https://fi.google.com/about/plan/(accessedApril2017.)
141
Onlypayforthelevelsyoureach,Ting,https://ting.com/rates(accessedApril2017).
142
Smartphoneserviceshouldnthavetocostalot.Nowitdoesnt.RepublicWireless,
https://republicwireless.com/cellphoneplans/?gclid=CjwKEAjwq5LHBRCN0YLf9
GyywYSJAAhOw6mWSuShk9jtYZlI_se8gHqgH3j82ex7MXUonHIli0wOBoCvZbw_wcB(accessedApril2017).
143
PhillipTracy,Movinginbuildingforwardwithsharedinfrastructureandneutralhost,RCRWirelessNews,
November1,2016,http://www.rcrwireless.com/20161101/carriers/inbuildingneutralhosttag31
144
KeithCarls,NextGenerationCellTowerComingtoAreaNearYou;cellularcompaniesscramblingtokeepup
withdemandfrommobiledevices,KEYT,May25,2016http://www.keyt.com/news/santabarbarascounty/next
generationcelltowerscomingtoareanearyou/87615564(accessedMarch2017).
163
services, which would lead to slow downs and price spikes.145Even if an individual purchases
highspeedserviceoranunlimitedplan,manyofthemajorcarrierstermsofservicecontainfine
printallowingthecarrierstothrottlespeedsduringperiodsofnetworkcongestion,oraftera
customerreachesacertainusagemilestoneeachmonth.Mostoftheunlimitedplanshavestrict
limits on how much data tethered devices can use, making them lessthan ideal for use as a
householdsprimaryinternetconnection.146
LowerCost Broadband Services Are Not Universally Available and Tend to Offer Lower
Speeds
Comcast Internet Essentials

For some lowincome and senior households that meet very specific requirements, Comcast
offersitsInternetEssentialsservicefor$9.95permonth.Theserviceincludesupto10Mbps
downloadspeeds,andadvertisesnocontractorinstallationfeesnecessary.Tobeeligibleforthe
service,allapplicantsmustmeetallthefollowingeligibilityrequirements:
LiveinanareacurrentlyservedbyComcast,
NothavebeenaComcastinternetsubscriberwithinthepast90days,and
HavenooutstandingdebttoComcastfromwithinthepastyear147
All families with students enrolled in schools with more than 40 percent of the student body
participating in the National School Lunch Programs free or reduced lunch program are
automaticallyapprovedfortheprogramanddonotneedtosubmitindividualdocumentation
beyondcertifyingthatachildisenrolledinapreapprovedschool.148
145
GeorgeFord,Havewegotitallwrong?Forecastingmobiledatauseandspectrumexhaust,PhoenixCenterfor
AdvancedLegal&EconomicPublicPolicyStudies,October21,2014,http://www.phoenix
center.org/perspectives/Perspective1406Final.pdf(accessedMarch2017).
146
TimMoynihan,VerizonsUnlimitedPlanisBack.Hereshowitcomparestoothercarriers,Wired,February13,
2017,https://www.wired.com/2017/02/verizonsunlimiteddataplanbackherescomparescarriers/;Christina
Warren,IsVerizonsresurrectedUnlimiteddataplanagooddeal?Gizmodo,February13,2017,
http://gizmodo.com/isverizonsresurrectedunlimiteddataplanagooddeal1792294971;JonBrodkin,TMobile
nowthrottlingmobilehotspotswhennetworkiscongested,ArsTechnica,October4,2016,
https://arstechnica.com/informationtechnology/2016/10/tmobilenowthrottlingmobilehotspotswhen
networkiscongested/;AdamIsmail,CricketsUnlimiteddatacustomerscouldsoonseespeedcutsduringpeak
hours,DigitalTrends,February27,2017,http://www.digitaltrends.com/mobile/cricketreducesunlimiteddata
speeds/(accessedMarch2017).
147
AffordableInternetathomeforyourfamily,InternetEssentialsfromComcast,
https://www.internetessentials.com/apply(accessedMarch2017).
148
InternetEssentialsfromComcast,FactSheet,Comcast,August2016,
https://www.internetessentials.com/~/media/InternetEssential/IE%20Fact%20Sheet%209.9.2016.pdf(accessed
April2017).
164
Householdsthatdonotmeetthepreapprovedcriteriamustdemonstratefinancialeligibilityby
submittingproofthat:
AtleastonechildiseligiblefortheFederalSchoolLunchProgram,
At least one member of the household is eligible for the Supplemental Nutrition
AssistanceProgram(SNAP)or
AtleastonememberofthehouseholdiseligibleforHousingandUrbanDevelopment
(HUD)assistance.149
Inaddition,seniorsinselectpilotareas,includingSanFrancisco,areeligiblefortheprogram,
provided they are older than 62, and are receiving state or federal assistance.150Community
collegestudentsenrolledinselectpublicpostsecondaryeducationalinstitutionsinColoradoand
Illinois that are Pell Grant recipients are also eligible for the program. 151 Comcast has not
announcedplanstoexpandthesepilotprogramsatthistime.
ItshouldbenotedthatwhiletheInternetEssentialsprogramisanindisputablestepintheright
direction,itslimitedscopeinassistingonlythoseonfederalassistanceprogramsmaybeleaving
behind those in the lower income brackets who do not qualify for federal help. For many
Americanswhodonotqualifyforfederalassistance,thechoicebetweenfoodandinternetmay
beanobvious,thoughdifficultdecision.
AnddespitetheavailabilityofInternetEssentials,theCitysBudgetandLegislativeAnalystfound
in 2015 that although 70 percent of public school students in the Citys public schools
(approximately11,000studentfamilies)areeligiblefortheprogram,onlyapproximately1,500
students(13.6percent)dosubscribe.152
AT&T Access
AT&Talsooffersalowcostproductforlowerincomeconsumers.Asoneoftheconditionsof
AT&TsmergerwithDirecTV,AT&Tagreedtoofferlowcostwirelinehomeinternetservicefor
lowincome individuals and households. The companys Access service is available to
householdslivinginAT&Tsexistingserviceareainwhichatleastoneresidentparticipatesinthe
Supplemental Nutrition Assistance Program (SNAP) program. The program is also open to
149
FrequentlyAskedQuestions:HousingAssistance,InternetEssentialsfromComcast,
https://www.internetessentials.com/GettingHelp/HousingAssistance(accessedMarch2017).
150
FrequentlyAskedQuestions:Seniors,InternetEssentialsfromComcast,
https://www.internetessentials.com/GettingHelp/Seniors(accessedMarch2017).
151
Accesstoschoolandjobopportunities,nowfromhome,InternetEssentialsfromComcast,
https://www.internetessentials.com/college(accessedApril2017).
152
CityandCountyofSanFranciscoBoardofSupervisorsBudgetandLegislativeAnalyst,PolicyAnalysisReport
regardingDigitalDivideinSanFrancisco,April15,2015.
165
Californians receiving Supplemental Security Income (SSI) benefits. Further requirements

include:
HomeaddressinacurrentlyservedareawithinAT&Ts21statefootprint
NooutstandingdebtforstandardAT&Tinternetservicewithinthepast6months
NooutstandingdebtforAccessservice
AT&Tprovidesqualifyingindividualswithspeedsrangingfrom768Kbpsto10Mbps,depending
onavailability.Servicelevelsbetween768Kbpsand3Mbpscost$5permonth,whilethe5to10
Mbpsservicecosts$10permonth.Forqualifyingindividuals,theservicecontainsnostartupor
installation fees, and includes a WiFi router for the home. Customers also receive access to
AT&Tsnetworkofhotspots.Likeotherwiredplans,Accessplansaresubjecttoadatacapof150
GBfor$5plansand1TBfor$10plans,andoverageswillbecharged$10per50GBoverplan
limits.153
TheAccessservicewillonlyremainavailableuntil2020.154
Initially,AT&TusedaloopholeinthemergerconditiontoofferthediscountedAccessservice
onlyinlocationswhereitsnetworkcouldsupport3Mbpsservice,regardlessofeligibility.155After
somecontroversy,thecompanyhassinceexpandedtheprogramtothoselivinginareaswithat
least768Kbpsserviceavailable.Individualslivinginareaswhere768Kbpsisnotavailablestill
cannotqualifyforservicethroughAccess.
Federal Lifeline Subsidy

Lowincome residents may be able to obtain a service subsidy through the federal Lifeline
program,whichoffersbenefitsforeithertelephoneorinternetservicetoqualifyinglowincome
householdsnationwide,effectivelyprovidinglowornocosttelephoneorinternetservice.Fora
householdtobeeligiblefortheprogram,amembermustbeenrolledinoneofninefederalaid
programs, including the Supplemental Nutrition Assistance Program (SNAP), Supplemental
Security Income (SSI), and Medicaid; or have a household income below the Federal Poverty
153
AccessfromAT&T,https://www.att.com/shop/internet/access/#/(accessedMarch2017).
154
AccessfromAT&T,https://www.att.com/shop/internet/access/#/(accessedMarch2017).
155
JonBrodkin,AT&Trefusestoofferlowincomediscountsforsub3MbpsInternet,ArsTechnica,September6,
2016,https://arstechnica.com/informationtechnology/2016/09/attrefusestoofferlowincomediscountsfor
sub3mbpsinternet/(accessedMarch2017).
166
guideline.Householdsareeligibleforonebenefiteithertelephoneorinternetserviceusually
totalinga$9.25creditontheirmonthlybill.156
Toapply,newcustomerspetitionqualifiedLifelineproviders(ofwhichthereare36inCalifornia,
includingAT&Tforhomephoneservice,butnotanyoftheCitysotherproviders157)andprovide
documentation of eligibility for the program. The providers then check applicants eligibility
against the registries of federal programs, and if accepted, receive funds for each eligible
subscriberfromtheUniversalServiceFund,managedbytheUniversalServicesAdministrative
Company.158Oncesubscribed,Lifelinecustomersmaychangeprovidersonceevery60daysfor
telephoneservice,oronceevery12monthsforinternetservice.159
Recentpoliticalchangeshavecreatedthreatstotheprogram.InFebruary,FCCChairmanAjitPai
announced that nine potential Lifeline providers would not be allowed to participate in the
Lifeline program,160despite their earlier certification by the FCC under former Chairman Tom
Wheeler.WheelersformeradvisorGigiSohncriticizedPaisactionsandnotedthattheywill
makethemarketforLifelinebroadbandserviceslesscompetitive,limitingchoiceandkeeping
priceshigh.Asaresult,fewerlowincomeAmericanswillbeabletoaffordbroadband.161
ChairmanPaireleasedaclarifyingstatementinlateMarch,notingIwanttomakeitclearthat
broadband will remain in the Lifeline program so long as I have the privilege of serving as
ChairmanButasweimplementtheLifelineprogramaswithanyprogramweadministerwe
mustfollowthelaw.Andthelawhereisclear:Congressgavestategovernments,nottheFCC,
theprimaryresponsibilityforapprovingwhichcompaniescanparticipateintheLifelineprogram
underSection214oftheCommunicationsAct.162
156
GetConnected:BecomeaLifelineCustomer,UniversalServiceAdministrativeCompany,
http://www.lifelinesupport.org/ls/(accessedMarch2017).
157
CompaniesNearMe:California,Lifeline,UniversalServiceAdministrativeCompany,
http://www.lifelinesupport.org/ls/companies/CompanyListing.aspx?state=CA&stateName=California(accessed
March2017).
158
LifelineProgramforLowIncomeConsumers,FederalCommunicationsCommission,
https://www.fcc.gov/general/lifelineprogramlowincomeconsumers(accessedMarch2017).
159
ChangeMyCompany,Lifeline,UniversalServiceAdministrativeCompany,
http://www.lifelinesupport.org/ls/changemycompany.aspx(accessedMarch2017).
160
BrianFung,TheFCCisstopping9companiesfromprovidingfederallysubsidizedInternettothepoor,
WashingtonPost,February3,2017,https://www.washingtonpost.com/news/theswitch/wp/2017/02/03/thefcc
isstopping9companiesfromprovidingsubsidizedbroadbandtothepoor/?utm_term=.2a7ab9593e96(accessed
March2017).
161
GigiSohn,DefendingtheIndefensible:ChairmanPaisLifelineReversalWillWidentheDigitalDivide,Blog
Posting,BentonFoundation,https://www.benton.org/blog/DefendingtheIndefensible(accessedMarch2017).
162
StatementofFCCChairmanAjitPai:OntheFutureofBroadbandintheLifelineProgram,NewsRelease,
FederalCommunicationsCommission,March29,2017,https://www.fcc.gov/document/chairmanpaifuture
broadbandlifelineprogram(accessedMarch2017).
167
LowincomeresidentscanalsoobtainfreedevicesthatwillconnecttoWiFinetworksprovided
bytheCity,communitycenters,andbusinessesthroughthefederalLifelineAssistanceProgram.
KnowncolloquiallyasObamaphones,thesedevicesareWiFienabledandprovidedatnocost
tolowincomeAmericanswhoqualify.Thereismuchanecdotalevidenceofthedevicesusein
lowincome and homeless communities both in San Francisco and nationwide. While many
Lifelinewirelesscarriersofferlimitedamountsofmobiledata(50to500MBpermonth),163,164,165
there isnooptiontoincreasetheamountofdataontheplans.Forhomelessandlowincome
families,thedevicesinternetfunctionalityisdependentontheavailabilityofWiFiserviceinthe
City.
Costs for Items Other Than Broadband Services Also Impact Broadband Adoption Rates
It should be noted that monthly service costs are not the only prohibitive factor for non
broadbandadopters.InthePewResearchpoll,10percentofrespondentscitedthehighcostof
purchasingacomputerasthereasontheydonothavebroadbandaccess.Whileinvestigating
barrierstobroadbandadoption,Daileyet.alfoundhardwareandsoftwarecosts,installation
costs and deposits, equipment maintenance fees, transaction costs for disconnecting, and
changes to subscription pricing all introduce additionaland often unpredictablelayers of
cost. 166 For respondents to Daileys surveys, these unanticipated costs were often cited as
reasonsfordroppingbroadbandathome.167
Factors Other Than Cost Also Impact Broadband Adoption Rates

Alackofdigitalandinternetliteracyandcomputerskillsmakethetaskofgetting onlinefeel
daunting,andcontributetolowerbroadbandadoptionrates.ManyAmericansofallageslack
the technical skills needed to feel comfortable and confident accessing online resources and
services,butthedigitalreadinessgapisespeciallypronouncedamongolderadults,whogrewup
inanerabeforetheinternetbecameintegratedintoeverydaylife.A2016PewResearchCenter
study found that 52 percent of adults were relatively hesitant to access online learning
services.168A2014studyalsofoundthatnearlyonethirdofAmericanshadlowlevelsofdigital
163
LifelineProgramDescription,AssuranceWireless,
http://www.assurancewireless.com/Public/MorePrograms.aspx(accessedMarch2017).
164
Plans,LifeWireless,https://www.lifewireless.com/main/plans(accessedMarch2017).
165
FrequentlyAskedLifelineQuestions,enTouchWireless,
https://www.entouchwireless.com/pages/free_phone_more_info#collapseTwentyFour(accessedMarch2017).
166
BroadbandAdoptioninLowIncomeCommunities,SocialScienceResearchCouncil,March2010,
http://webarchive.ssrc.org/broadband_adoption.pdf(accessedMarch2017).
167
Ibid.
168
JohnHorrigan,DigitalReadinessGaps,PewResearchCenter,September20,2016,
http://www.pewinternet.org/files/2016/09/PI_2016.09.20_DigitalReadinessGaps_FINAL.pdf(accessedMarch
2017).
168
skills, and this lack of skills have the largest impact on determining what online services an
individualwillaccess.169
Evensomeuserswhohavethetechnicalcapabilitiestousetheinternettoaccesscertainservices
or resources still may not feel comfortable accessing other services due to a lack of trust or
understanding of privacy issues. According to a 2015 NTIA survey, 84 percent of online
households named at least one concern they had about online privacy and security risks,
includingidentitytheft,creditcardfraud,datacollection,lossofcontroloverprivatedata,data
collectionbygovernmentandthreatstopersonalsafety.170Thesefearsleadsomeindividualsto
stayofflinealtogether,whileotherslimittheirinternetusetoactivitiestheyfeelcomfortable
with.
Manyindividualswillnotwantinternetenoughtosignupforhomebroadbandserviceuntilthey
gain the skills they need to feel confident and comfortable online. Many organizations have
developedsuccessfulprogramsthathelplowincomeandelderlypopulationsdevelopthedigital
literacy skills they need to take advantage of online resources and services. 171 The most
successfulprogramsuserelevantcontent,andestablishtrustbetweentheindividualsandthe
organizations offering the service. 172 These efforts help build demand for the internet, and,
togetherwithlowcostaccesstobroadbandandhardware,leadtohigherratesofbroadband
adoption.
Digital Inclusion Can Be Expanded Through LowCost, HighBandwidth Services

Makingreliable,highbandwidth,homebroadbandserviceavailableatanaffordablepricewill
notsolvealldigitalinclusionproblems,butitwillmakeasubstantialdifferenceonthesupplyside
ofthechallenge.ComcastsInternetEssentialsprogramhasmadegoodprogressinthisdirection,
butitisnotavailabletoalllowincomeindividuals,andwithadvertisedspeedsupto10Mbps,
itwillnotsupportmanyofthedatahungryapplicationsthatarequicklybeingintegratedinto
schoolsandworkplaces.
169
JohnHorrigan,DigitalReadiness:NearlyonethirdofAmericanslacktheskillstousenextgenerationInternet
ofthingsapplications,June2014,
http://jbhorrigan.weebly.com/uploads/3/0/8/0/30809311/digital_readiness.horrigan.june2014.pdf(accessed
March2017).
170
RafiGoldberg,LackofTrustinInternetPrivacyandSecurityMayDeterEconomicandOtherOnlineActivities,
BlogPosting,NationalTelecommunicationsandInformationAdministration,May13,2016,
https://www.ntia.doc.gov/blog/2016/lacktrustinternetprivacyandsecuritymaydetereconomicandother
onlineactivities(accessedMarch2017).
171
StrategiesandRecommendationsforPromotingDigitalInclusion,ConsumerandGovernmentalAffairs
Bureau,FederalCommunicationsCommission,January11,2017,
https://apps.fcc.gov/edocs_public/attachmatch/DOC342993A1.pdf(accessedMarch2017).
172
Libraries'IncreasingRoleinBroadbandAdoption:Barriers:DigitalLiteracy,Relevancy&Cost,Benton
Foundation,January2016,https://www.benton.org/initiatives/librariesbroadbandadoption/barriers(accessed
March2017).
169
Inordertomovebeyondthedigitaldivide,lowincomeindividualsandfamiliesneedtohave
access to a similar online experience as their higherincome counterparts. If affordable home
broadband services only provide a secondclass experience, they will continue to hinder low
income individuals ability to fully partake in economic, educational, and civic life in the 21st
century. Instead of building demand for home broadband, a slow service could generate
frustrationandcausesomeuserstogiveuponhomebroadbandservicealtogether.
Abetterapproach,intermsofmeetingtheCityspolicygoals,wouldbetomakethesameservice
thatisavailableontheopenmarket(i.e.,thestandard,offtheshelfservicethatconsumerscan
buy)availabletolowincomeconsumersatareducedrate.
Creating a special, deliberately inferior product (i.e., lower speed) at a reduced rate for low
incomeconsumersistheapproachtakenbyComcastforitsInternetEssentialsprogramandby
AT&T for its Access service. These products are low cost only in terms of the monthly fee in
comparisontootherbroadbandservicesonthemarket;theyarenotlowfeeonapermegabit
basis.
This means that the industrys lowcost digital inclusion services are very basic. They do not
enablemanyoftheapplicationsthatarefundamentaltotheinternettoday,suchasstreaming
educationalvideos.
Thealternativeapproachtocreatingadigitalinclusionproductistomakeavailableaninternet
productthathasthesamelevelofqualitythatwealthypeopleareabletobuy,buttomakeit
availabletolowincomeconsumersatlowercost.Thiswasthemodelinitiallyadoptedforthe
Lifelineprogramforvoicecommunications.
Giveneverythingdiscussedhereincludingtheimportanceofbandwidth,andthefactthatwhile
manylowincomeresidentsalreadyhaveaccesstoComcastandAT&Tsproducts,thoseproducts
arenotequivalenttowhatwealthypeopleareabletobuy(andarenottransformativeforlow
incomeusers)theBroadbandforSanFranciscoInitiativeshouldconsiderprioritizingthesecond
model,notthefirst.
That said, for individuals struggling to make ends meet, even a $10 a month service can be
difficulttoafford.Successfuldigital inclusioneffortsoftengoastepfurtherintryingtomeet
peoplewheretheyare,providingasmuchflexibilityinpaymentoptionsaspossible.
Forexample,PCsforPeopleinSt.Paul,Minnesota,setsupprepaidplansfortheirclients,and
allowsthemtopaycashforonemonth,threemonth,sixmonth,or12monthperiodsofservice
atatime.Theorganizationoffersthefirstthreemonthsfree,tohelpoffsetthecostoftherouter
andallowtheirclientstosaveupbeforeanotherpaymentwasdue.Iftheydonothaveenough
savedupbythetimethenextpaymentisdue,theirserviceisjustputonholduntiltheymake
170
theirnextpayment.Andinsteadofhavingtointerfacewithalarge,corporatecustomersupport
system,userscanwalkintoalocalfacilitystaffedwithfriendlyindividualsthatwanttohelpthem
getonline.173
InWilson,NorthCarolina,apartnershipbetweenthemunicipalGreenlightnetworkandthelocal
housingauthorityallowsresidentstosignupfor50Mbpssymmetricalserviceat$10permonth.
Thehousingauthorityprovidesresidentswithafreerouter,thuseliminatingadditionalstartup
costs,andthenaddsthemonthlyfeetousersexistingmonthlybills,simplifyingthepayment
process.174
Tomaximizetheimpactofanaffordable,highbandwidthservice,itwillneedtobepairedwith
digitalliteracytraining,lowcosthardwareinitiatives,andpublicaccesscomputercentersthat
offer support. These complementary initiatives will help build demand for highspeed home
broadband.
However,inmanycases,thedemandforhomebroadbandisalreadythere.Thereissimplyno
oneofferingtosupplyadequateserviceatapricethatlowincomeindividualscanafford.
173
AngelaSiefer,InnovatorsinDigitalInclusion:PCsforPeople,BentonFoundation,September28,2016,
https://www.benton.org/blog/innovatorsdigitalinclusionpcspeople(accessedMarch2017).
174
ChristopherMitchell,WilsonGreenlight,PublicHousingAuthoritySolveAccessGapCommunityBroadband
BitsEpisode236,CommunityNetworks,January17,2017,https://muninetworks.org/content/transcript
communitybroadbandbitsepisode236(accessedMarch2017).
171
Appendix I: Potential Technical Framework for the Lit Fiber, Multi

Provider, OpenAccess Network
Thefollowingdescribes,atahighlevel,ourrecommendationsfortheLitFibernetworkthatwill
be administered by the Lit Fiber Concessionaire: An openaccess FTTP network capable of
simultaneoususebymultipleRetailServiceProviders(RSP).
An openaccess FTTP network serves as a platform for multiple service providers to reach
subscribersoverasharedinfrastructure.Thenetworkmodelprovidesacentralizedplacewhere
RSPscanconnecttothenetworkanddelivertheirservicestotheirsubscribers.
Thissharedlastmilefiberwillreducethecostfornewserviceproviderstoenterthemarket,
reduceredundantandcostlybuildstoreachsubscribers,andgivesubscribersmorechoicesfor
serviceproviderstomeettheirconnectivityneeds.
Network Architecture
This openaccess network model takes a layered approach to dividing the responsibilities of
runningafullyfunctionalFTTPnetwork.Wedescribethethreerolesasfollows:
1. Dark Fiber Concessionaire: Responsible for the physical fiber plant. This includes
construction,theinitialnetwork,respondingtoandrepairingfiberbreaks,constructing
newfiber,splicing,andanyongoingmaintenancetasks.
2. LitFiberConcessionaire:Responsibleforlitservicesandthenetworkelectronicsrequired
to deliver those services. This includes provisioning services between RSPs and their
subscribersandactingasacentralpointofcontactforthenetwork:providingserviceto
RSPsanddispatchingtheDarkFiberConcessionairewhenfiberworkisrequired.TheLit
FiberConcessionairewillalsoprovideNOCservicesandoperateaCivicNetworkandself
serviceportal.
3. Retail Service Provider (RSP): Responsible for providing network services, such as an
internetconnection,tosubscribersovertheFTTPnetwork.RSPsobtainaccesstotheFTTP
networkthroughtheLitFiberConcessionaireandselltheirservicestosubscribersonthe
network.
172
Figure22illustratestheresponsibilitiesofaDarkFiberConcessionaire,aLitFiberConcessionaire,
andaRetailServiceProviderinalayeredservicemodel.
Figure22:ResponsibilitiesinaLayeredServiceModel

The primary goal of the FTTP network is to provide a fast, secure, and reliable connection
betweenanRSPanditssubscribers.TheFTTPnetworkwillallowanRSPtoconnectviaaprovider
accessportandofferretailservicetoanysubscriberonthenetwork.
TheLitFiberConcessionairewillcreateanEVC(Ethernetvirtualconnection)thatassociatesan
RSPaccessportwithoneormoresubscriberconnections,ineffect,creatingapipebetweenan
RSPandtheirsubscribers.ManyEVCscanbecreatedtocarryvariousservicesfromdifferentRSPs
acrossthenetworkinasecureway.
TheLitFiberConcessionairemayemployvariousnetworkingtechnologiesandcreateanetwork
thatisascomplexorassimpleasnecessaryandstillpresenttheEVCasasimpleconnection
betweenanRSPanditssubscribers.
173
Figure23illustrateshowmultipleRSPscanconnecttotheFTTPnetworkandprovidevarious
servicestoconnectedsubscribers.
Figure23:MultipleServicesProvidedovertheSameFTTPNetwork
Figure24illustratesanEVCprovisionedbetweenRSPAsportandaportonasubscribersONT.
AllthreeserviceprovidersshownhavetheirownEVConthenetwork,butonlyRSPAsserviceis
provisionedtothisparticularcustomer.
Figure24:ServicefromRSPAIsProvisionedtotheCustomerONTviaanEVC
174
TheFTTPnetworkshouldallowformultipletypesofEVCservicesincluding:
1. Pointtopointservicesusedtoconnecttwosubscriberlocations(Figure25)
2. Multipoint(LAN)servicesusedtointerconnectmorethantwolocations(Figure26)
3. Teeservicesusedtoconnectmultipleleaflocationsbacktoarootlocationwhich
allowstheleaflocationstocommunicatewiththeroot,butnotwithoneanother(Figure
27)
Atreeservicemaybedesirableforcertainapplicationssuchasresidentialinternetserviceora
corporateWANwherebranchlocationsshouldnotbeabletoseeeachotherstraffic.
ThenetworkalsomustallowforperformancemonitoringorindividualEVCsandqualityofservice
classificationssothatSLAscanbemonitoredandenforcedorsothatcertaintraffic,suchaslive
videoorVoIPtraffic,canbeprioritizedwhencrossingtheFTTPnetwork.
Figure25:APointtoPointTopology Figure26:AMultipoint,orLAN,Topology
Figure27:ATreeTopology

Network Operational Model
WedividetheoperationalresponsibilitiesfortheFTTPnetworkintothreelayeredcategories:
fibermaintenance,transportservice,andretailservice(Figure28).TheDarkFiberConcessionaire
175
will be responsible for constructing and maintaining the physical fiber plant, the Lit Fiber
Concessionairewillberesponsibleforoperatingtransportserviceoverthephysicalfiberplant,
andmultipleRSPswillconnecttothenetworkandpurchasetransportservicesinordertoprovide
retailnetworkservicestosubscribers.Thismodelestablishesclearseparationofresponsibilities
andlinesofcommunication.
Fibermaintenance:Responsibilityforthephysicalfiberplant.Thisincludesrespondingto
andrepairingfiberbreaks,constructingnewfiberasrequired,splicing,andanyongoing
maintenancetasks.
Transport service: Responsibility for lit services and network electronics (routers and
switches)andatthefiberaccesslayer(OLTandONT).Thisincludesprovisioningservices
betweensubscribersandservicesprovidersandprovidingnetworkmonitoringservices.
Retailnetworkservice:Responsibilityforprovidingnetworkservices,suchasaninternet
connection,tosubscribersovertheFTTPnetwork.
Figure28:LayeredOperationalResponsibilitiesfortheOpenAccessFTTPNetwork

TheLitFiberConcessionaireservesacentralroleinthisoperationalmodel.Theyareresponsible
notonlyforoperatingtransportservices,butalsoforprovidingNOC(networkoperationscenter)
serviceswhichincludemonitoring,troubleshooting,andfixingthetransportnetwork;providing
supportandprovisioningservicesforRSPs;andengagingtheDarkFiberConcessionairewhena
fixornewfiberinstallationisrequired.TheNOCshouldserveasasinglepointofcontactforRSPs
andtheDarkFiberConcessionairewhilethesubscriberrelationshipshouldbemanagedbythe
RSPs.
TheNOCshouldactivelymonitorthenetworkanditsservicesandproactivelyfindandresolve
issuesonthenetwork.TheLitFiberConcessionaireshouldalsoprovidetheRSPswithsomeview
intotheirmonitoringsolution.RSPsshouldalsobeabletoescalatetroublereportstotheLitFiber
Concessionaire (Figure 29). The Lit Fiber Concessionaire will assist in troubleshooting and
176
determinewhethertheissueiswiththetransportnetwork,thephysicalfiberplant,oranissue
withtheRSPorsubscribernetworkanddirecttheissuetothecorrectgroup.
Figure29:TroubleReportWorkflow

New Subscriber Provisioning
TheLitFiberConcessionairewillberesponsibleforacceptingrequestsfromRSPsfornewservices
tobeprovisioned.Mostservicerequestswillfallintooneofthreecategories:
AnRSPwishestoofferserviceinanewarea.Thiswillrequireinstallationofanewfiber
distributioncabinet(FDC)andnewfibertoconnecttheFDCtothenetworkandtothe
subscriberpremiseswithinthenewservicearea.
AnRSPwishestoofferservicetoanewsubscriberpremisesinanexistingservicearea.
Thiswillrequireeitherinstallationofaservicedropbetweenthesubscriberpremisesand
existingFDCorconstructingalateralfiberconnectionfromanexistingfibercableaswell
asconfigurationandinstallationofnewelectronicsandprovisioningservicetothenew
subscriber.
AnRSPwishestooffernewserviceorchangeexistingservicesatanexistingsubscriber
premises.ThiswillrequireprovisioninganRSPsEVCtoanexistingsubscriberport.
177
Services requiring new fiber: If a service requires new fiber, the Lit Fiber Concessionaire will
engagetheservicesoftheDarkFiberConcessionaireandprovideanestimatedcompletiondate.
Dependingonthemodelbeingused,theDarkFiberConcessionairemayinstallandconfigurean
ONTorswitchatthesubscriberpremisesaspartoftheconstructionprocess.
Servicesrequiringnewelectronics:WeenvisionascenariowheretheLitFiberConcessionaire
will preconfigure and preprovision a new subscriber ONT and provide it to the RSP to be
installed along with the service providers CPE. This process would limit the number of visits
requiredtothesubscriberpremisesandallowtheretailserviceprovidertohavemorecontrol
over the subscriber relationship. The Lit Fiber Concessionaire would need to be immediately
availabletoprovideassistanceiftheretailserviceproviderisunabletobringupthesubscriber
connection.
Alternatively, the ONT may be installed by the Dark Fiber Concessionaire during the fiber
constructionprocess.Thiswouldallowthetransportservicetobefullytestedbeforetheretail
serviceproviderarrivesandmaybemoredesirableincertainscenarios,suchascaseswherethe
ONT is being installed on the outside of the subscriber building, where the fiber installation
processnecessitatesentryintothesubscriberpremise,orwhereconfigurationandprovisioning
oftheONTrequiresthattheONTbeinstalledandactive.
Servicesrequiringprovisioningtoanexistingsubscriberport:TheLitFiberConcessionairewill
implementaselfserviceportalthatallowssubscriberstoselectstandardservicesfromavailable
retailproviders.(ThisselfserviceportalisdescribedingreaterdetailinSection7.4.5.)Formore
advanced services, the Lit Fiber Concessionaire NOC should be prepared to provision the
appropriate subscriber ports within a time frame that will be specified in the providers SLA
(Figure30).
178
Figure30:ServiceProvisioningWorkflow
179
Appendix J: Financial Analysis of Purely Public Model

Tofullyunderstandthepotentialviabilityofthepurelypublicmodelforthisinitiative,theProject
Teamdevelopedafullfinancialmodelandanalysisoftheeconomicsofthetraditionalmunicipal
ownershipandoperationsmodel.
Framework
ThefinancialanalysisinthissectionassumestheCityowns,operates,andprovidesretailvoice,
video,anddataservicestoresidentsandbusinessesinthecommunity.Theanalysisisbasedon
thelowcostmodeldevelopedbytheengineers175andtheassumptionsdescribedbelow.
Themodelisdesignedtobecashflowpositiveinyearone;thisisaccomplishedthroughbond
andloanfinancing.Overtime,giventhelowercostestimatetoconstruct,maintain,andoperate
theFTTPnetwork,themodelindicatesthata47percenttakerateofhouseholdsandbusinesses
passedwillberequiredtobreakeven.
TheanalysisassumesthattheCityoffersthreedataservices:
1Gbpsresidentialserviceat$80permonth
1Gbpssmallcommercialserviceat$90permonth
1Gbpsmediumcommercialservice(withservicelevelagreements)at$240permonth176
TheseservicepricesareabovethoseofGoogleFiberinothermarkets($70for1Gbps)butare
lowerthansomeotherproviderspricing.Forexample,Comcastsrecentlyannouncedresidential
1Gbpsserviceislikelytobepricedinexcessof$150permonth.
The analysis assumes that 25 percent of business customers will purchase the higherlevel
service.ItalsoassumesthattheCitywilloffervideoanddataservicesthroughaservicepartner
orcontractor.Forvideo,weassumethatcustomerswillpayanaveragepackagepriceof$85per
month,andthattheCitywillreceive$12.75.Fortelephone,weassumethattwoservicesare
offeredaresidentialserviceat$20permonthandabusinesslineat$40permonthandthat
theCitywillreceive$5permonthand$10permonth,respectively.
Table31representsasummaryofthismodel.
175
Theimplicationsofthehighercostestimatearedescribedasascenariobelow.
176
Mediumcommercialservicereceivesaloweroversubscriptionrate,thatis,lesscustomerssharingthe
connection,decreasingtheinstancesofnetworkcongestionreducingoverallspeeds.
180
Table31:MunicipalRetailModelFinancialSummary
IncomeStatement Year1 Year5 Year10 Year15 Year20

TotalRevenues $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
TotalCashExpenses (14,641,640) (95,009,580) (91,973,400) (91,973,400) (91,973,400)
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)
InterestExpense (16,457,000) (57,012,630) (44,572,300) (28,991,300) (9,308,020)

Taxes

NetIncome $(34,566,455) $(8,637,520) $7,402,050 $24,623,430 $44,306,710

CashFlowStatement Year1 Year5 Year10 Year15 Year20
UnrestrictedCash $498,455 $9,553,265 $22,393,535 $38,247,385 $54,219,245
Balance
DepreciationReserve 61,421,730 71,081,400 10,923,510 16,297,160
InterestReserve 15,057,000
DebtServiceReserve 16,730,000 64,420,000 64,420,000 64,420,000 64,420,000
TotalCashBalance $32,285,455 $135,394,995 $157,894,935 $113,590,895 $134,936,405
Themodelassumesthatsubscribershipfordataserviceswillrampupoveryearsonethrough
four,andthenremainsteady.Themodelassumesthatsubscribershipforvoiceandvideoservices
will ramp up over the same period, but then decline as adoption of overthetop (OTT)
alternativesincreases.177
Financing Costs
ThisfinancialanalysisassumesthattheCitywillseekaseriesofbondsandaonetimeloan.It
assumesthatadebtservicereserveof5percentandaninterestreserveaccountisrequiredfor
thebonds.
Theanalysisestimatestotalfinancingrequirementstobealmost$1.29billioninbondsanda$20
millionloan.Thebondsareissuedata4.5percentfinancerateandprincipalpaymentsstartin
thethirdyearafterissuance.
Wealsoassumea20year,$20millionloanisobtainedininyearone.Theloanisissuedata7
percentfinancerateandprincipalpaymentsstartinsixthyearafterissuance.
Themodelassumesastraightlinedepreciationofassets,andthattheoutsideplantwillhavea
20yearlifespanwhilethenetworkequipmentwillneedtobereplacedafter10years.Customer
PremisesEquipment(CPE)aswellasothermiscellaneousequipment(testequipment,vehicles,
177
Theanalysisusesaflatmodel(i.e.,doesnotincludeinflationandsalarycostincreasesbecausethoseoperating
cost increases will be offset and passed on to subscribers in the form of increased prices). Models that add an
inflationfactortobothrevenuesandexpensestypicallygreatlyoverstatefuturecashflowbecausenetrevenuesare
unlikelytoincreaseasquicklyasinflation.Atbest,theproviderwillbeabletomatchexpensesincreaseswitha
dollarfordollarrateincrease,whichiswhattheflatmodelrepresents.
181
computers)willneedtobereplacedeveryfiveyears.Networkequipment,includinglastmileand
CPEswillbereplacedorupgradedat80percentoforiginalcost,miscellaneousimplementation
costsat100percent.Themodelplansforadepreciationreserveaccount,startinginyearthree,
tofundfutureelectronicsreplacementsandupgrades.
Table32showstheincomestatementforyearsone,five,10,15,and20.TheCitysnetincomeis
aboutnegative$34.6millioninyearone,andnegative$8.6millioninyearfive.Byyear10,the
operationwillgenerateanetincomejustunder$7.4million,growingtojustover$24.6million
inyear15,andjustover$44.3millioninyear20.
Table33showsthecashflowstatementforyearsone,five,10,15,and20.Theunrestrictedcash
balance is approximately $498,000 in year one and $22.4 million in year 10. By year 15, the
unrestrictedcashbalanceisjustover$38.2million;itisjustover$54.2millionbyyear20.
182
Table 32: Income Statement
Revenues
Video $5,595,720 $39,615,780 $39,615,780 $39,615,780 $39,615,780
InternetResidential 8,004,480 160,089,600 160,089,600 160,089,600 160,089,600
InternetBusiness 672,120 13,428,360 13,428,360 13,428,360 13,428,360
Enterprise
Voice 1,125,600 13,692,000 9,688,800 9,688,800 9,688,800
ConnectionFee(net) 658,275

AncillaryRevenues

Total $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
ContentFees
Video $4,759,830 $33,674,280 $33,674,280 $33,674,280 $33,674,280
Internet 124,350 2,418,490 2,418,490 2,418,490 2,418,490
Enterprise

Voice 844,200
10,269,000 7,266,600 7,266,600 7,266,600
Total $5,728,380 $46,361,770 $43,359,370 $43,359,370 $43,359,370

OperatingCosts
OperationCosts $4,923,260 $28,186,810 $28,153,030 $28,153,030 $28,153,030

LaborCosts
3,990,000 20,461,000 20,461,000 20,461,000 20,461,000
Total $8,913,260 $48,647,810 $48,614,030 $48,614,030 $48,614,030

EBITDA $1,414,555 $131,816,160 $130,849,140 $130,849,140 $130,849,140

Depreciation 19,524,010 83,441,050 78,874,790 77,234,410 77,234,410

OperatingIncome(EBITDAless $(18,109,455) $48,375,110 $51,974,350 $53,614,730 $53,614,730
Depreciation)

NonOperatingIncome
InterestIncome $ $314,600 $338,750 $188,360 $201,790
InterestExpense(10YearBond)
InterestExpense(20YearBond) (15,057,000) (55,927,230) (43,758,410) (28,446,980) (9,366,150)

InterestExpense(Loan) (1,152,640)
(1,400,000) (1,400,000) (732,680) (143,660)
Total $(16,457,000) $(57,012,630) $(44,572,300) $(28,991,300) $(9,308,020)
NetIncome(beforetaxes) $(34,566,455) $(8,637,520) $7,402,050 $24,623,430 $44,306,710
FacilityTaxes $ $ $ $ $
NetIncome $(34,566,455) $(8,637,520) $7,402,050 $24,623,430 $44,306,710
183
Table 33: Cash Flow Statement

a.NetIncome $(34,566,455) $(8,637,520) $7,402,050 $24,623,430 $44,306,710
b.CashOutflows
DebtServiceReserve $(16,700,000) $ $ $ $
InterestReserve (30,114,000)
DepreciationReserve (21,277,470) (20,113,070) (19,694,770) (19,694,770)
Financing (3,346,000)
CapitalExpenditures (303,926,100)

Total
$(354,116,100) $(21,277,470) $(20,113,070) $(19,694,770) $(19,694,770)
c.CashInflows
InterestReserve $15,057,000 $3,307,500 $ $ $
DepreciationReserve
InvestmentCapital
10YearBond/LoanProceeds
20YearBondProceeds 334,600,000
LoanProceeds 20,000,000
Total $369,657,000 $3,307,500 $ $ $

d.TotalCashOutflowsandInflows $15,540,900
$(17,969,970) $(20,113,070) $(19,694,770) $(19,694,770)

e.NonCashExpensesDepreciation $19,524,010 $83,441,050 $78,874,790 $77,234,410 $77,234,410

f.Adjustments
ProceedsfromAdditionalCashFlows(10Year
$ $ $ $ $
Bond)
ProceedsfromAdditionalCashFlows(20Year
$ $ $ $
Bond) $(334,600,000)
ProceedsfromAdditionalCashFlows(Loan) $(20,000,000) $ $ $ $

g.AdjustedAvailableNetRevenue $56,833,560 $66,163,770 $82,163,070 $101,846,350
$(354,101,545)

h.PrincipalPaymentsonDebt
10YearBondPrincipal $ $ $ $ $
20YearBondPrincipal 47,289,880 62,195,610 77,507,040 96,587,870

LoanPrincipal 1,043,250 1,463,220 2,052,240

Total $ $47,289,880 $63,238,860 $78,970,260 $98,640,110

i.NetCash $498,455 $9,543,680 $2,924,910 $3,192,810 $3,206,240

j.CashBalance
UnrestrictedCashBalance $498,455 $9,553,265 $22,393,535 $38,247,385 $54,219,245
184
Capital Additions
Significant network expenses, known as capital additions, are incurred in the first few years
during the construction phase of the network. These represent the equipment, material and
constructionlaborassociatedwithbuilding,implementing,andlightingafibernetwork.Table34
showsthecapitaladditionscostsinyearsonethroughfour,assuminga48percenttakerate,or
about175,500customers.
Thisanalysisprojectsthatcapitaladditionsinyearonewilltotalapproximately$303.9million.
Thesecostswilltotalapproximately$486.3millioninyeartwo,justunder$342.6millioninyear
three,and$106.2millioninyearfour.
185
Table 34: Capital Additions

CapitalAdditions Year1 Year2 Year3 Year4
NetworkEquipment
Core&GPONEquipment $65,615,200 $16,403,800 $ $

AdditionalAnnualCapital

Total $65,615,200 $16,403,800 $ $

OutsidePlantandFacilities

TotalBackboneandFTTP $
$214,675,200 $357,792,000
$143,116,800


$
Total $214,675,200 $357,792,000 $143,116,800

LastMileandCustomerPremises
Equipment
CPE(residentialandsmallcommercial) $5,762,600 $28,654,300
$51,033,800 $27,170,000
CPE(mediumcommercial) 77,000 384,300 690,900 383,600
CPE(enterprise)
AverageDropCost 16,656,100 82,810,000
147,476,700 78,530,100

AdditionalAnnualReplacementCapital

$22,495,700 $106,083,700
Total $111,848,600 $199,201,400

MiscellaneousImplementationCosts
OSS $400,000 $ $ $
Vehicles 70,000 140,000 105,000
EmergencyRestorationKit 100,000
WorkStation,Computers,andSoftware 70,000 138,000 170,000 82,000
FiberOTDRandOtherTools 100,000
BillingSoftware 250,000
FiberManagementSoftware 150,000


Total $1,140,000 $278,000 $275,000 $82,000

TotalCapitalAdditions $106,165,700
$303,926,100 $486,322,400 $342,593,200
Operating and Maintenance Expenses

Undertheassumptionsinthemodel,theCitysoperatingandmaintenanceexpenses,including
principalandinterestpayments,willtotalover$31millioninyearone,growingto$199.3million
inyearfive,$199.7millioninyear10,and$199.9millioninyears15and20.
O&Mexpensesincludestaffingforsalesandmarketing,networkoperations,andotherfunctions
new to the City. The addition of new staff will require new office space. Similarly, network
inventoryrequirementswillrequirewarehousingspace.ThemodelbuildsincostsfortheCityto:
186
Expandofficefacilitiesformanagement,technical,andclericalstaff
Openaretailstorefronttofacilitatecustomercontactandenhancetheirexperiencedoing
businesswiththeFTTPenterprise
Providewarehousingforreceiptandstorageofcableandhardwarefortheinstallation
andongoingmaintenanceofthebroadbandinfrastructure
Establish a location to house servers, switches, routers, and other core network
equipment
TrainingnewandexistingstaffwillberequiredtofullyrealizetheeconomiesofstartingtheFTTP
network.Marketingandsaleswillalsoberequired,andwillrepresentanewactivityfortheCity.
Staffingwithskillsinthefollowingdisciplineswillberequired:
Sales/Promotion Finance
Internetandrelatedtechnologies VendorNegotiations
StaffManagement Networking(addressing,segmentation)
StrategicPlanning Marketing
Theexpandedbusinessandincreasedresponsibilitieswillrequiretheadditionofnewstaff.The
initialadditionalpositions,staffinglevels,andbasesalariesareshowninTable35.Thesenumbers
assume two shifts of customer service representative support and 1.5 shifts of customer
technicians. Changing to full 24x7 staffing will increase costs. Similarly, reducing the support
hours(e.g.,to7a.m.to8p.m.)willdecreasetherequiredstaffing.
Note that Table 35 lists only new employeesthe model assumes no existing staff will be
allocatedtotheenterprise.Thelistedsalariesdonotincludeoverhead.Toaccountforoverhead,
themodeladds40percenttothebasesalaries.
187

Table 35: Labor Expenses
NewEmployees Year1 Year2 Year3 Year4 Year5+ Year1Salary
BusinessManager 1.00 1.00 1.00 1.00 1.00 $150,000
Marketing&SalesManager 1.00 1.00 1.00 1.00 1.00 $125,000
GIS&Recordkeeping 2.00 3.00 3.00 3.00 3.00 $90,000
NetworkEngineer 2.00 3.00 3.00 3.00 3.00 $120,000
NetworkTechnicians 4.00 6.00 6.00 6.00 6.00 $80,000
Sales&MarketingRepresentatives 10.00 20.00 25.00 25.00 25.00 $65,000
CustomerServiceRepresentatives 4.00 24.00 56.00 74.00 74.00 $65,000
ServiceTechnicians/Installers&ITSupport 4.00 23.00 55.00 73.00 73.00 $90,000
HR,Administration&Support 1.00 3.00 3.00 3.00 3.00 $75,000
CallCenterSupport 2.00 8.00 18.00 23.00 23.00 $65,000
FiberPlantO&MTechnicians 4.00 12.00 18.00 18.00 18.00 $90,000
Total 35.00 104.00 189.00 230.00 230.00
Additionalkeyoperatingandmaintenanceassumptionsinclude:
Directinternetaccesscostsareestimatedat$0.50perMbpspermonthforthefirst3,000
Mbps,and$0.40perMbpspermonthbeyondthat
Insuranceisestimatedtobezeroinyearoneand$400,000fromyeartwoon
Utilitiesareestimatedtobe$100,000inyearoneand$200,000fromyeartwoon
Officeexpensesareestimatedtobe$200,000inyearoneand$400,000fromyeartwo
on
Facilityleasefeesareexpectedtobe$240,000inyearoneand$480,000fromyeartwo
on
Locatesandticketprocessingareestimatedtostartinyearoneat$71,300,increaseto
$356,400inyeartwo,andincreaseto$712,900fromyearthreeon
Peeringexpensesareestimatedat$42,100inyearone,$252,800inyeartwo,$631,900
inyearthree,and$842,600inyearfouron
Contingencyisestimatedtobe$200,000inyearoneand$400,000fromyeartwoon
Billingandmaintenancecontractfeesareestimatedat$75,000fromyeartwoon
Legalfeesareestimatedtobe$200,000inyearsoneandtwo,and$100,000fromyear
threeon
Consultingfeesareestimatedat$150,000inyearoneand$75,000fromyeartwoon
188
Marketing and promotional expenses are estimated to be $1 million in year one and
$750,000fromyeartwoon
Vendormaintenancecontractfeesareexpectedtostartat$8.2millioninyeartwoandremain
steady from year two on (based upon 10 percent of accrued investment). Annual variable
operatingexpensesnotincludingdirectinternetaccessinclude:
Educationandtrainingarecalculatedas2percentofdirectpayrollexpense
Allowanceforbaddebtsiscomputedas0.5percentofrevenues
Churn is anticipated to be 6 percent annually, which initiates a $175 per subscriber

acquisitioncost.
TheestimatedcostofelectronicbillingforthenewFTTPenterpriseis$0.50perbill.Thisisin
addition to the $250,000 cost for the billing software, which will require updating every five
years,178and$75,000inannualsoftwaresupportfees.
Fibernetworkmaintenancecostsarecalculatedat1percentofthetotalconstructioncost,per
year.Thisisestimatedbasedonatypicalrateofoccurrenceinanurbanenvironment,andthe
costofindividualrepairs.Thesecostswilltotal$2.3millioninyearone,$6.7millioninyeartwo,
$9.6millioninyearthree,androughly$10.4millioninyearfouron.Thisisinadditiontostaffing
coststomaintainthefiber.
Table36showsoperatingexpensesforyearsone,five,10,15,and20.Asseen,someexpenses
willremainconstantwhileotherswillincreaseasthenetworkexpandsandthecustomerbase
increases.
178
Replacementvalueestimatedat80percentoftheoriginalcost.
189
Table 36: Operating Expenses and P&I Payments
OperatingExpenses Year1 Year5 Year10 Year15 Year20

SupportServices $ $ $ $ $
Insurance 400,000 400,000 400,000 400,000
Utilities 100,000 200,000 200,000 200,000 200,000
OfficeExpenses 200,000 400,000 400,000 400,000 400,000
FacilityLease 240,000 480,000 480,000 480,000 480,000
Locates&TicketProcessing 71,300 712,900 712,900 712,900 712,900
Peering 42,100 842,600 842,600 842,600 842,600
Contingency 200,000 400,000 400,000 400,000 400,000
BillingMaintenanceContract 75,000 75,000 75,000 75,000
Fiber&NetworkMaintenance 2,313,310 10,410,570 10,410,570 10,410,570 10,410,570
VendorMaintenanceContracts 8,201,800 8,201,800 8,201,800 8,201,800
Legal 200,000 100,000 100,000 100,000 100,000
Planning
Consulting 150,000 75,000 75,000 75,000 75,000
Marketing 1,000,000 750,000 750,000 750,000 750,000
EducationandTraining 79,800 409,220 409,220 409,220 409,220
CustomerHandholding
CustomerBilling(Unit) 55,640 1,080,940 1,075,940 1,075,940 1,075,940
AllowanceforBadDebts 80,280 1,134,130 1,114,110 1,114,110 1,114,110
Churn(acquisitioncosts) 97,380 1,891,650 1,882,890 1,882,890 1,882,890
PoleAttachmentExpense 93,450 623,000 623,000 623,000 623,000
Video 4,759,830 33,674,280 33,674,280 33,674,280 33,674,280
Internet 124,350 2,418,490 2,418,490 2,418,490 2,418,490
Enterprise

Voice 844,200 10,269,000
7,266,600 7,266,600 7,266,600
SubTotal $10,651,640 $74,548,580 $71,512,400 $71,512,400 $71,512,400

LaborExpenses
$3,990,000 $20,461,000 $20,461,000 $20,461,000 $20,461,000
SubTotal $3,990,000 $20,461,000 $20,461,000 $20,461,000 $20,461,000

TotalExpenses $14,641,640
$95,009,580 $91,973,400 $91,973,400 $91,973,400
PrincipalandInterest $16,457,000 $104,302,510 $107,811,160 $107,961,560 $107,948,130

FacilityTaxes

SubTotal $16,457,000
$104,302,510 $107,811,160 $107,961,560 $107,948,130
TotalExpenses,P&I,andTaxes $31,098,640 $199,312,090 $199,784,560 $199,934,960 $199,921,530
190
Sensitivity Scenarios
To understand the sensitivity of the financial model to a range of possible outcomes, we
developedsensitivityscenarios.Manyofthesescenariosmaynotberealisticallyattainable,but
aremeanttodemonstratethesensitivityofthefinancialprojectionstotheseassumptions.
Base Case
Aswepreviouslynoted,thebasecaseshowsthata47percenttakerateisrequiredtobreak
even,assumingthelowercostconstructionestimate.179Table37showsafinancialsummaryfor
thisscenario.
Table 37: Base Case Scenario

TotalRevenues $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)
Taxes
NetIncome $(34,566,455) $(8,637,520) $7,402,050 $24,623,430 $44,306,710

Take Rate Drops by 10 Percentage Points

Ifthetakerateweretodropby10percentagepoints,theenterprisewouldquicklyturnfroman
unrestrictedcashbalanceofabout$406,000inyearonetoanegativecashbalance.Byyearfive,
theunrestrictedcashbalancedeficitwouldbeover$61.9million,growingtoadeficitofover
$324millionbyyear20.
179
Indicatestakerateinyear4.Thetakeratedeclinesslightlyovertimeduetothereductionofvideoandvoice
subscribers.
191
Table 38: Take Rate Drops by 10 Percent

TotalRevenues $12,644,430 $178,564,620 $175,415,820
$175,415,820 $175,415,820
TotalCashExpenses (13,108,080) (79,018,980) (76,630,810)
(76,630,810) (76,630,810)
Depreciation (17,910,490) (73,623,070) (70,028,340)
(68,684,740) (68,684,740)
InterestExpense (15,669,500) (52,030,410) (40,592,890)
(26,281,980) (8,223,660)

Taxes

NetIncome $(34,043,640) $(26,107,840) $(11,836,220) $3,818,290 $21,876,610

UnrestrictedCashBalance $406,050 $(61,936,285) $(151,077,335) $(237,684,675) $(324,075,705)
TotalCashBalance $30,530,550 $51,217,495 $(20,509,365) $(146,818,695) $(219,236,235)
Voice and Video Revenue Drops by 50 Percent

Whilethebasecaseassumesthatvoiceandvideorevenuewilldropovertimeascustomersturn
to overthetop (OTT) alternatives, we also assumed a revenue share with a voice and video
provider.Asthetablebelowillustrates,reducingvoiceandvideorevenueby50percentfromthe
basecaseturnstheunrestrictedcashbalancenegativebyyear5,withthedeficitincreasingto
justunder$7millionbyyear10.Thisdeficitwouldbereducedovertime,totaling$4.9millionby
year20.
Table 39: Voice and Video Revenue Share Drops by 50 Percent

TotalRevenues $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)

Taxes

NetIncome $(34,986,769) $(11,611,110) $4,428,460 $21,649,840 $41,333,120

UnrestrictedCashBalance $78,141 $(4,964,808) $(6,992,490) $(6,006,592) $(4,902,684)
Residential Internet Fee Decreases or Increases by $5 per Month

IfcompetitivepressuresforcetheCitytoreducetheresidentialinternetservicefeeby$5per
monthfromthelevelsassumedinthebasecase,theenterprisewouldnotbefinanciallyfeasible.
192
Byyear10,theunrestrictedcashbalancedeficitwouldbemorethan$58.2million,growingtoa
deficitofjustunder$126millionbytheendofyear20(seethetablebelow).
Table 40: Residential Internet Fee Decreases by $5 per Month

TotalRevenues $15,555,915 $216,820,140 $212,816,940 $212,816,940 $212,816,940
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)

Taxes

NetIncome $(35,064,235) $(18,593,090) $(2,553,520) $14,667,860 $34,351,140

UnrestrictedCashBalance $675 $(21,309,005) $(58,246,585) $(92,170,585) $(125,976,575)
TotalCashBalance $31,787,675 $104,532,725 $77,254,815 $(16,827,075) $(45,259,415)
With a $5 per month increase in residential internet service fees, with all other assumptions
remainingthesame,theunrestrictedcashbalancewouldballoon,comparedtothebasecase,
andhit$234.4millionbyyear20(seetablebelow).
Table 41: Residential Internet Fee Increases by $5 per Month

TotalRevenues $16,556,475 $236,831,340 $232,828,140 $232,828,140 $232,828,140
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)

Taxes

NetIncome $(34,068,675) $1,318,050 $17,357,620 $34,579,000 $54,262,280

Reducing or Increasing Operating and Staffing Expenses by 25 Percent

Assumingallotherfactorsadheretothebasecase,reducingtheoperatingandstaffingexpenses
by25percentwouldhaveasubstantialpositiveimpactoncashflow.Asthetablebelowshows,
unrestrictedcashbalancewouldsoartoover$354.7millionbyyear20inthisscenario.
193
Table 42: Reduce Operating and Staffing Expenses by 25 Percent (Content Fees Remain the Same)
TotalRevenues $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)

Taxes

NetIncome $(31,590,015) $7,360,870 $23,391,990 $40,613,370 $60,296,650

UnrestrictedCashBalance $3,474,895 $70,196,065 $162,990,265 $258,793,815 $354,715,375

Ontheothersideofthatequation,iftheCitysoperatingandstaffingexpensesweretoriseby
25percentoverthebase,theenterprisewouldfacelargedeficitsovertime.Asthetablebelow
shows,theunrestrictedcashbalancedeficitinthisscenariowouldberoughly$140.5millionby
year10andover$294.2millionbyyear20.
Table 43: Increase Operating and Staffing Expenses by 25 Percent (Content Fees Remain the Same)
TotalRevenues $16,056,195 $226,825,740 $222,822,540 $222,822,540 $222,822,540
Depreciation (19,524,010) (83,441,050) (78,874,790) (77,234,410) (77,234,410)

Taxes

NetIncome $(38,041,645) $(27,193,540) $(11,145,520) $6,075,860 $25,759,140

UnrestrictedCashBalance $(2,976,735) $(60,685,675) $(140,587,475) $(217,471,475) $(294,237,465)
TotalCashBalance $28,810,265 $65,156,055 $(5,086,075) $(142,127,965) $(213,520,305)
High Network Implementation Costs

If implementation costs approach the high cost estimate, breaking even becomes even more
challenging. The total bonding in the first four years increases to roughly $1.43 billion (from
almost$1.29billion),whiletheinitialloanremainsat$20million.Theincreasedbondingisfrom
acombinationofthehigherconstructioncostsandtherequiredincreaseinsubscribers.Thetake
raterequiredtobreakevenincreasesto53percentinyearfour(from47percentinthelowFTTP
costmodel).Thecashflowandincomestatementforecastforthismodelareshowninthetable
below.
194
Table 44: Use FTTP High Cost Estimate

TotalRevenues $18,109,770 $255,787,560 $251,273,160 $251,273,160 $251,273,160
Depreciation (20,768,800) (93,031,680) (87,882,480) (86,242,100) (86,242,100)

Taxes

NetIncome $(36,241,550) $(5,314,860) $12,529,090 $31,447,130 $53,240,890

195

San Francisco Municipal Fiber Network: CTC Report - 10.11.17

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

San Francisco Municipal Fiber Network: CTC Report - 10.11.17

Uploaded by

Copyright:

Available Formats

ThePotentialforUbiquitous,OpenFiberto

1.1 FibertothePremises Design and Cost Estimate

CostComponent LowCostScenario HighCostScenario

CostComponent LowCostScenario HighCostScenario

1.2 Delivery Mechanism Analysis

Private broadband investment is resulting in new broadband competition in unevenly

1.2.2 Application of P3 Delivery to Broadband Involves Novel Questions and

In contrast, in the broadband environment, application of a P3 involves an opposite set of

1.2.3 The Classic P3 Model Applied to Broadband Creates Some Challenges

1.3.2 Financial Analysis and Modeling of Dual P3 Model

Output BaseCase Range

Output BaseCase Range

Output BaseCase Range

1.3.3 Potential P3 Model: Single Dark Fiber P3 Only

1.3.6 Financial Analysis and Modeling of Single Dark Fiber P3

1.3.7 Upfront securing revenues

1.3.8 Innovative, LowerCost Financing Structure

1.3.9 Subsidy of Connection Fee for Lower Income San Franciscans

1.4 Procurement Strategy and Recommendations

2 FTTP Design and Cost Estimate

2.1 Summary of Cost Estimates

2.1.1 Lit FTTP Cost Estimate

CostComponent LowCostScenario HighCostScenario

2.1.2 Dark FTTP Cost Estimate

CostComponent LowCostScenario HighCostScenario

2.2 Approach to Technical Analysis

2.2.1 Field Survey

2.2.2 Subsurface Investigations

2.2.2.1 Directional Boring

2.2.2.3 Street Restoration

2.2.2.4 Sidewalk Restoration

2.3 FTTP Network Design

2.3.1 Network Design

2.3.2 Network Core and Hub Sites

2.3.3 Distribution and Access Network Design

2.3.3.1 Access Network Technologies

2.3.3.2 Expanding the Access Network Bandwidth

2.3.3.3 Customer Premises Equipment (CPE) and Subscriber Services

2.4 FTTP Cost Estimate

CostComponent LowCostScenario HighCostScenario

2.4.2 Components of Construction Cost Estimate

Special Crossings consists of specialized engineering, permitting, and incremental

2.4.3 Outside Plant Construction Costs

2.4.4 Central Network Electronics Costs

NetworkSegment Passings CostperPassing Subtotal

2.4.4.1 Core Electronics

2.4.4.2 Distribution Electronics

2.4.4.3 Access Cabinet Electronics

2.4.5 Subscriber Electronics and Service Drop Installation Costs (PerSubscriber

2.4.5.1 Customer Premises Equipment and Related Subscriber Electronics

2.4.5.2 Drop Installation

HighDensity $11,000,000 26,000 300,000 98 $700

3 The Purely Public Broadband Option Involves High Risk

3.1.1 Track Record of Public Broadband Efforts in FTTP

3.1.2 Track Record of Public Broadband in California

3.1.3 Advantages of Having a Public Electric Utility

3.4 Financial Analysis

DSL represents a relatively lowbandwidth form of broadbanda network of roads, not

4.3 Etheric Networks

Webpassoffersitscustomersspeedsrangingfrom100Mbpsto1Gbpsfor $60 permonth.72

4.8 Factors That Can Impact Private Deployment Patterns

4.9 Benefits and Challenges of the Private Investment Model

4.10 Risk of Consolidation and Reduction in Competition