ManagementOperationsReviewTeam(MORT) ReportoftheDataCenterTigerTeam December23,2009 CommitteeCharge ExecutiveVicePresident(EVP)DiazandProvostWoodsoncreatedaManagementOperationsReview Team(MORT)toseekandimplementimprovementstoouroperationalservices.TheMORTteam consistsofdeanakridge(agriculture),deanjamieson(engineering),deanweiser(liberalarts),vice PresidentAlmond(BusinessServices),VicePresidentBuckius(Research),VicePresidentMcMains (PhysicalFacilities),VicePresidentMcCartney(InformationTechnology),andManagingDirectorKen Sandel(OfficeofEVPandTreasurer). TheMORTgroupselectedinformationtechnology(IT)ontheWestLafayettecampusasthefirstfocus areatoexaminepotentialsynergiesandcostssavings.sixcommittees,calledtigerteams,wereformed toexaminethefollowingareas:datacenters,campusitorganizationalstructure,emailservices, OnePurdue,desktopcomputingservices,andthecomputerlabs.Eachteamwastaskedtoprovide recommendationstoimprovethemanagement,efficiency,andaccountabilityofcampus wide informationtechnologyoperations. ExecutiveSummary Adatacenterisafacilityusedtohousecomputersystemswiththeirassociatedcomponents (networking,storage,etc.).astechnologyhasbecomeinfusedineveryaspectofcampus,data center facilitieshavebeencreatedinanad hocmannertomeettheimmediateneedsofdepartmentsand colleges.thesefacilitiesweredevelopedbyadaptingexistingrooms,placedinthecorneroflaboratories andoffices,orincludedinnewbuildingprojects.theresultingpatchworkofsmalldatacentershas increasedthecomplexityofprovidingadequateutilitiestobuildings,managingtheinstitutionalutility costs,andmeetingtheongoinggrowthofcyberinfrastructure. ThisdatacenterTigerTeamproposesaseriesofstepsthatwillultimatelyleadtobetterfacilitiesand moreefficientutilizationofuniversityresources.thesestepsinclude: Identifyingexistingdatacenters. Identifyinganticipateddatacenterneeds. Establishingstandardsandprocesses. Developingandimplementingaplanforcreationofnewand/orrenovateddata centerspace. Developingandimplementingaplanforconsolidatingexistingdata centerspaces. Developingandimplementingadata centerenergyplan Whilethecommittee schargewastoreviewinformation technologyoperationstomakesuggestionson themanagement,efficiency,andaccountabilityofthedatacenters,therecommendationsinthisreport arenotlikelytohaveanimmediateimpact.thedatacenterissueshaveevolvedoverdecadesandwill likelytakeyearstoevolvetoamoresustainablemodel.iftheinstitutionimplementedthese recommendations,asignificantsavingscouldberealized totalingmorethan$10millionoverthe nextdecadeafterfullimplementation. MORT DataCenterReport Page1
Longer termbenefitsoftheserecommendationsincludecostsavingsfrom: Adoptionofenergy efficientcomputers:adoptingmoreenergy efficientcomputertechnologies couldsaveapproximately40wperserver(4,400servers)and15wperdesktop/laptop(26,000 desktops/laptops).usingindustry standardpractices,eachcomputerwattsavedequatestoa 2.8wattsavingsintheentireutilitychain(includingcooling).Byadoptingmoreenergy efficient chips,theuniversitycouldsaveapproximately$700,000peryearwhenfullyimplementedover 5years.Theadditionalhardwarecostwouldbemitigatedthroughconsolidatedpurchases. Consolidationofadministrativeserverstoincreaseoverallutilization:Accordingtothe2008 2009DataDigest,morethan1,200serversarelocatedacrosscampusforadministrativeuse.If thoseserverscanbeconsolidatedtovirtualmachinesatasimilarrateasthepilotproject(30:1), consolidationcouldresultinasmuchas$1.2millioninannualhardwaresavings.theresulting consolidationwouldalsoresultin80percentlesspowerconsumption(assuming1,200servers at85watts) savingapproximately$400,000inutilitycosts(usingthe2.8powermultiplierfor completesupply chaincosts).theresultmaybehighersincemanyserversmaynotbereported. Consolidationofresearchclusterstoincreaseoverallutilization:Typicalresearchclustersare utilized30 35percentofthetimewhilethecommunityclustersaretypicallyutilizedmorethan 95percent.Usinga2:1utilizationratio,approximately50percentormorecanbesavedon utilitiesbysharingcomputationalresources.for1,000computerclusternodesdistributedon campus,approximately$200,000peryearinutilitiescanbesavedbyprovidingthesame numberofcomputationalhoursinasharedcluster($560,000peryearutilizingthe2.8utility multiplier). Consolidationofdatacenterstoreduceconstructionandmaintenancecostsassociatedwiththe facilities:duringthepastyear,theuniversityhasspentmorethan$100,000onpowerand coolingupgradestosmallerspaces.additionalsavingscouldbeanticipatedthroughreductionin maintenance. Eliminationofenergy intensiverooms(power/cooling)withinbuildingstoallowforbetter building wideenergysavings(suchasreducingcoolingcostsinsummermonthsandoccupancydrivenenergyreductions):withapotential$100,000perbuildingutilitysavings,approximately $4millioncouldbesavedacrossthe40buildingswithidentifieddatacenters. Leveragingcampus widepurchasestomaximizevendordiscounts:typicalsavingsfor consolidatedpurchasesrangefrom$300 $500permachine.Assumingafive yearlifecyclefor the4,400serversreportedinthe2008 2009DataDigest,totalsavingscouldrangebetween $264,000and$440,000.Iftheprogramwasextendedtothe26,000desktopsandlaptops(fiveyearlifecycle),anadditionalsavingscouldbebetween$1.5and$2.6million. Maintainingsecurityforcomputersystems(physical,data,andsystemsecurity):Accordingto industrystandards,anaveragecomputerincidentcostsapproximately$200.ifbettersystem administrationcouldreducetheincidentrateby50percent(to600incidents),then approximately$120,000couldbesavedperyear. Utilizingidlecomputationalcyclestosupportteachingandresearchratherthanpurchase additionalhardware:iftheinstitutioncouldharnessthecomputationalpowerofthe30,000 machinesreportedinthe2008 2009DataDigestfor12hoursaday,researcherswouldhave accesstonearly360,000cpuhoursperday ortheequivalentofnearly2,000researchnodes costingmorethan$4million(or$800,000peryear)overthelifetimeoftheequipment.the savingscouldbedoubledasnewerdual coremachinesarebeingpurchased.note:notall researchcanbeaccommodatedthroughthisresource. MORT DataCenterReport Page2
Background Definition Adoptingasimple,clear,easy to understanddefinitionofadatacenteristhefirstessentialstepto managingtheinstitutional facilityneeds.thecommitteeproposesthatadatacenterisaroomdevoted tocomputersthatsatisfiesatleastthreeofthefollowingcharacteristics: Theareaisatleast300squarefeet. Theroomcontainsatleast20computers/servers. Theroomcontainsatleast3racksofcomputers/servers. Thepowerusedbythemachinesisatleast50kilowatts. Theaveragepowerusedbythemachinespersquarefootofareaoftheroomisatleast1.5 kilowattspersquarefoot. Theroomusesatleast10percentoftheelectricalpowerforitsbuilding. Thetotalweightofthecomputersintheroomisatleast500pounds. Theroomisdesignedwithraisedfloors. Whilethedefinitionofadatacenterwillevolveovertime,thecommittee sinitialdefinitionwillallowa dialogueaboutspacededicatedtothisfunction.thedefinitionwillbecomeabasistoreviewthe57 spacesidentifiedwithinthespace managementdatabaseasserverrooms(codedas710 CEN COMPUTERinthesystem),aswellasallowcollegesanddepartmentstoidentifyotherspaces. Byidentifyingdata centerspaces,theinstitutionwillbebetterableto: Providebetternotificationofpotentialelectricalandcoolingmaintenance. Implementbuilding wideenergy savingstrategieswithoutimpactingoperations(e.g.,reducing heatingandcoolingwhenunoccupied). Provideincentivesforrelocatingequipment(e.g.,availabilityofredundantcooling). Theclassificationofdatacentersisnotintendedtobeusedfortypicaloffices,studentlabs,or laboratories althoughsomeofthesespaceshavebeenutilizedtohousecomputerequipment.the intentistoidentifyspacewithsignificantcomputerequipmentand,overthelongerterm,toaddressthe institutionaltechnology,electrical,andcoolingneeds. TheChallenge:Access,Capacity,andEfficiency Thecurrentdistributionofdatacentershasevolvedoverdecadesandhasresultedinapatchworkof datacenters.mostofthesespacesaresmallroomsdesignedtomeetimmediateneedswithout considerationoflong termsustainabilityorimpactoncentralutilityservices.withoutalong term, strategicviewofthesespaces,theinstitutionwillremainunabletomanagetheimpactonutilitycostsor optimizethesespacestomeetinstitutionalneeds.thechallengesofaccess,capacity,andefficiencyof datacentersarethekeyissuestosuccess.thissectionwillrevieweachoftheseissues. MORT DataCenterReport Page3
Access Accesstothelargerdata centerspaceswastypicallylimitedtothoseapplicationssupportedbythe centraladministrativeand/orteachingunits.asaresult, eachdepartmentandcollegewasresponsiblefor providingitsowndatacenterspacestomeetits programmaticneeds.forunitswithrecentbuilding projects,data centerspaceswerefrequentlyincludedas partofvariouscampusprojects.duringthepastfive years,datacenterswerebuiltwithinsevennew academicbuildings addingapproximately8percentto totaldata centerspace.theinclusionofadditionaldatacenterspacesreflectstheimportanceofsuchspacesto includewithinnew buildingprogramdocuments. Forunitswithmoreimmediateneeds,equipmentwasinstalledinanyavailableassignedspace.These locationsoftenincludedresearchlaboratories,teachingspaces,remotebasement storageareas,and convertedclosets whateverlocationmightprovideenoughspace,power,andcoolingforthe moment(seepicturesbelow.). Serversinstalledinanoldstoragearea (Note:Thermometeratnearly90 o ) Clusterinstalledinalab MORT DataCenterReport Page4
Serversinstalledaspartofanoffice Clusterinstalledinabasementoncampus Theoveralldecentralizationofthedecision makingprocesshasresultedinmorethan50 declared datacentersandprobablydozensofotherspacesbeingusedtohousecomputerequipment.ofthe identified38,597squarefeetofdata centerspace,nearly21,000squarefeetareoperatedbyitap.of the21,000squarefeet,approximately8,000isutilizedbyresearch,7,000byadministration/teaching, 4,000bytelecommunications,andtheremaining2,000bysupportingdepartmentalequipment(colocationfacility).Amajorityofthecampusdata centerspacesarerelativelysmall withonlyninesites occupyingmorethan1,000squarefeetandmorethanhalfoccupyinglessthan300squarefeet. Withoutaccesstoproperlyconfigureddata centerspace,departmentswillcontinuetocreateadditional spaces distributingthechallengeofprovidingadequateandcost effectivepowerandcooling solutionstobuildings.anyconsolidationofdata centerspacesmustaddressaccesstodepartmental equipmentandwilllikelyincludeasetofdistributeddatacentersratherthanonesingledatacenter. Capacity Whilepastgrowthhasbeenaccommodatedthrough modestincreasesinnewconstruction,aswellas consolidationofequipment,projectedgrowth requirementsexceedtheuniversity sdata center capacity.thetable(right)illustratestheprojected growthindata centerspace rangingfromlittlegrowth intelecommunicationstomodest(2percent)growthin departmentalandadministrative/teachingtodoubling theresearchdata centerspace.demandforadditionalresearchspacehasbeenpropelledbygrowthof researchfunding,aswellasdiscipline basedshiftstoutilizingmorecyberinfrastructure.aspurdue continuestomakesignificantstridesinexternalresearchfunding,thedata centercapacityproblemsare alsolikelytogrow. EYP MCF,adata centerconsultingfirm,particularly notesthat Purdue sextensivecapabilitytoshare researchcomputingsystemsthroughoutthepurdue Universityenvironmentisthebestimplementationof researchsystemsutilizationthateyp MCFhasseento date. Thisimplementationofresearchsystemshas allowedtheinstitutiontogrowfrom80activeprincipal investigators(pis)tonearly200(150percentgrowth) since2000.theimpactofthisefficientuseofour capacitycanbemeasuredbytheongoinggrowthof computationalcyclesusedandthenumberofactive principalinvestigators.whilethechartdemonstrates centrallysupportedresearchgrowth,thetrendsandneedforadditionaldata centercapacityarelikely tobetypicalofmanydepartments. MORT DataCenterReport Page5
Efficiency ThelargerUniversitychallengeisoptimizingtheresourcesallocatedtosupportingthevariousdata centersandtheequipmentwithinthem.efficiencycanbeexaminedinseveralways: Serverutilization:Serverutilizationcanbeincreasedthroughtwoprimarymethods virtualization andsharing.throughvirtualizationtechnologies,multiplevirtualserverscanbeconsolidatedona singlephysicalmachine.virtualizationreduceshardware,staffing,andutilitycosts.attheendof 2009,ITaPhas297virtualserversrunningonninephysicalmachines.Inadditiontosaving$350,000 peryearinhardwarecosts,thevirtualserversuse80percentlesspowerandcooling,andoccupy lessthan1/10 th asmuchspaceastheywouldrunningonphysicalhardware.throughthe community clusterprogram,researchersshareidlecomputationalcycleswithfellowresearchers improvingcpuutilizationfromatypical30 35percenttomorethan95percent.Thesharingallows productiveusageofthepowerconsumed. Institutionallifecyclecost:Thetotallifecyclecost includinghardware,support,andutilities shouldbeconsideredtoimproveoverallefficiency.throughwarrantynegotiationsandequipment retirement,theinstitutioncanrecognizesignificantsavingsinutilitycostsandmaintenance agreements.astandard$1,800server(300watts)willutilize$130inpower($.05/kwh)peryearor $650duringthelifetime.Whenutilizingthe2.8supplychainmultiplier,thetotalcostsare$1820 overthefive yearlife. Purchasingleverage:Bypoolingamajorityofequipmentintoacommonpurchase,theinstitution canreduceoverallhardwareequipmentcosts.forexample,thecommunity clusterprogramwas abletosaveapproximately20percent($500)oneachserver savingmorethan$1millionduring thepasttwoyears. Powerandcoolinginfrastructureandmaintenance:Theimplementationandmaintenanceofpower andcoolingisnotalinearfunctionoftheamountofequipment.forexample,thecosttoinstall powerandcoolingfor2,000machinesmaycostonly20percentmorethanthecostofinstallingthe infrastructuretosupport1,000machines.inaddition,thecosttomaintaindistributedinfrastructure willbesignificantlymoreexpensivethanfewer,largerunits. Powerandcoolingutilization:Utilizingpowerandcoolingstandards,theUniversitycanmonitor overallutilizationandadjusttomeetindustrystandards.theinstitutioncurrentlyspends approximately$60,000perdayonutilities(nearly$21millionperyear).implementingbuilding wide energy efficiencyplansaregreatlycomplicatedbythedistributionofthedata centerinfrastructure andthelackofbuilding independentsystems. MORT DataCenterReport Page6
Recommendations TheteamrecommendsthattheUniversitytakethefollowingactions: IdentifyExistingDataCenters Collaboratingwiththedepartmentsandcollegestoidentifycurrentdata centerspacesisthefirststep toestablishadata centerplan.thecommitteerecommends: Utilizingtheinitialdefinitionofdatacenters,asktwotothreecolleges/departmentstoreport spacesthatmeetthecharacteristics. Basedontheinformationcollected,refinedefinitionofdatacenter. Collectremaininglistofdata centerspacesfromremainingcolleges(includingthoseinnewor renovationplans).collectpower,cooling,floor space,andrack spaceinformation.calculate relevantindustry efficiencyratingssuchaspower usageeffectiveness(pue). Classifydata centerusagebasedonteaching,research,andadministrativeusage. Updatethespace managementdatabase. IdentifyAnticipatedData CenterNeeds Basedonthedataavailable,thecommitteeestimatedthattheacademicneedsfordata centerspace arelikelytoincreaseby2percentandtheresearchneedsarelikelytogrowbymorethan10percent eachyear.foramoreaccurateprojection,thecommitteerecommends: Surveyingthecollegesanddepartmentsforanticipateddata centerneedsduringthenextfive to10years. Identifyingpotentialprojectsthatmayhavesignificantshort termimpactondata centerspace (e.g.,proposedmajor researchcenters). Determiningtheneedsforspecializeddata centerspaces,suchasthosemeetingitaror DepartmentofDefensestandards. EstablishStandardsandProcesses Thecreationofdata centerspacesiscurrentlyanad hocprocessdeterminedbytheneedsofindividual units.asaresult,theinstitutionisspendingmorefundingthannecessarysincelargerspacesare significantlymorecost effective.thecommitteerecommends: Establishingdata centerstandardsintermsofenergyefficiency,security,andmanageability. Creatingareviewpanelthatwouldexaminethebuilding/remodelingofdata centerspacesto findsynergieswithexistingprojects. Communicatingdata centerissuesandstandardstodepartmentheadsandinformationtechnologyleaders. MonitoringChangesinTechnologyandServices Technologyandthesupportinginfrastructurecontinuetochangerapidly.Forexample,theintroduction ofcloudservicesisfrequentlymentionedasalong termsolutiontoresearchneeds.whencomparing currentcloudofferingswiththecommunitycluster(e.g.,steele),thecostsarenotjustified($2.7million forcloudservicesversus$1.8millionforsteele).however,theservicesareexpectedtocontinually evolveandpricingisanticipatedtodecline.thecommitteerecommends: Establishingacost reviewcommitteetomonitorprogressofemergingtechnologies(e.g., clouds)andmakeadditionalrecommendations. MORT DataCenterReport Page7
DevelopandImplementPlanforCreationofNewand/orRenovatedData CenterSpace Thegrowthofcampususagewilllikelyrequiretheexpansionand/orrenovationofexistingdata center space.developingalonger termplanwillsavetheinstitutionsignificantfundingwhilemeetingthe departmentalneeds.forexample,thecostofimplementingadatacentertohandle1,000servers withinanewfacilitycostsapproximately$750,000,whilethecostofrenovatinganexistingbuildingis typically$1,250,000.similarsavingscanbefoundbetweenrenovatingseveralsmalldatacentersversus renovatingonelargerfacility.thecommitteerecommends: Creatinglarger community data centerspacesthatprovideappropriate,accessible,secure, anddistributedspacesfordepartmentaluse. Designingdata centerspacesthatadapttothegrowthofinstitutionalneedsincludingthe potentialofdevelopacontainerizeddata centerfarm(datacentersbuiltwithincargo container units). Distributingdata centerspacesthatbalancedepartmentalaccesswithcampusspaceandutility needs. Developincentivesfordepartmentstoutilize community data centerspaces. DevelopandImplementaPlanforConsolidatingExistingData CenterSpaces Thenumberofsmalldatacenterspresentsanumberofchallengestotheinstitution,includingdifficulty inmanagingbuilding powerand coolingloads(e.g.,buildingtemperaturemustbemaintainedfor entirefacility),adoptingenergy efficientstrategies,maintainingproperphysical/systemsecurity,and providinghigherlevelofservice.thecommitteerecommends: Providingincentivestoconsolidateserversanddatacentersintofewerlocations.Incentivesmay include: o Offering cashforclunkers tomigrateoffoldsystemsorconsolidateserversthrough virtualization. o TotheextenttheUniversitymovestodistributeutilitycost,providing freeelectrical power tounitsthatutilizetheconsolidateddatacenters. o Implementingadditionalchargesfordepartmentsretaininginefficientdatacenters o (suchasthecostforcoolingtheentirebuildingwhenunoccupied). Providingsystem administrationstafftomaintaintheoperationandsecurityofthe equipment. Recognizingdepartmentalneedsforaccesstoequipment. Developingagovernancestructurethatallowsfacilitiestobemanagedbyinformationtechnologygroupslocatedcentrallyorwithincolleges. Returningexistingdata centerspacestodepartments. Developingagovernanceprocesstomakedecisionswhenconflictingdemandsarise. DevelopandImplementaData CenterEnergyPlan Astheinstitutioncontinuestogrowcyberinfrastructuretomeettheacademicandresearchmissions, thecostforutilitieswillbeasignificantissue.adata centerenergystrategywillbenecessaryto maintainahighlevelofservicewhilemoderatingtheimpactofutilitycosts.astrategymayinclude: Adoptingenergy efficiencystandardsforcomputerpurchases. Creatinglifecycleplansforequipment,includingmandatoryequipmentretirement. Developingstrategiesforutilizingwasteheatfromdatacenters. Locatingdatacentersnearalternativeenergysources. Creatingincentivesfordepartmentaladoptionofenergystandards. MORT DataCenterReport Page8
AppendixA:TeamMembership ThedatacenterTigerTeamconsistedofthefollowingmembers: JeffBolin CollegeofScience jtb@purdue.edu 44922 VinceBralts CollegeofEngineering bralts@purdue.edu 45345 JohnCampbell ITaP john campbell@purdue.edu 41289 MichaelFosmire LibraryScience fosmire@purdue.edu 42859 BobOlson OfficeofUniversityArchitect rolsonjr@purdue.edu 63786 KenSandel OfficeofEVP&Treasurer sandel@purdue.edu 40990 ErickVanMeter PhysicalFacilities eevanmeter@purdue.edu 48797 SamWagstaff(Chair) ComputerScience&theCenterfor ssw@cs.purdue.edu 46022 EducationandResearchin InformationAssuranceandSecurity (CERIAS) BobWild Chemistry wild@purdue.edu 45205 MORT DataCenterReport Page9
AppendixB:Data CenterSpaces Bldg. Abbv. Room Data CenterSpaces(coded710):AsReportedbySMASon12/22/09 Dept. Code Dept.Abbv. Area Pcs.Abbv. Description ARMS B191 1282 AERO&ASTRO 965 ACADINSTR ServerRoom BIND B025A 1212 BINDBIOCTR 194 RESCHCENTER ServerRoom BRNG 6145 1440 EDUCATIONAD 67 ACADADMIN Servers BRWN 1151 1393 CHEMISTRY 84 ACADINSTR ComputerServer BSG 115 1661 BLDGSERV 357 CUSTODIAL EnergyManagementSystem BTV 100 1744 UNIVRESDIR 616 AUXSTUDENT BoilerTelevision CIVL 4252 1397 EARTH&ATMS 380 ACADINSTR ComputerMachineRoom CIVL B146 1301 ENGCPTNTWK 630 COMPUTING ECNComputers EE 007B 1285 ELEC+CPTENG 484 ACADINSTR ComputerRoom ENAD 301C 1528 INTLPRGMS 97 ANCILLARY ServerRoom Computers ENAD 403 1128 SPACEMGMT 164 GENADMSERV ComputerServers FREH G002 1061 ITSYS&OPS 4,993 ADMCOMPUTNG ComputerRoom Comp.AidedProc.Engr.Mach. FRNY 1025 1283 CHEMENGR 543 ACADINSTR Room FS 3216A 1160 FOODSCIENCE 100 ACADINSTR FileServers GRIS B001 1301 ENGCPTNTWK 657 COMPUTING ECNNeverVacated HAAS 234 1061 ITSYS&OPS 1,985 ADMCOMPUTNG MachineRoom HAAS 279 1398 COMPUTERSCI 301 ACADINSTR MachineRoom HOCK 203 1392 BIOLOGY 309 ACADINSTR Server KNOY 204 1410 TECHADMIN 427 ACADINSTR TCNCentralInfrastructure KRAN 761 1310 MANAGEMENT 411 ACADINSTR ComputerEquipment KRAN 761 1145 AGECON 202 ACADINSTR ComputerEquipment KRAN 761A 1310 MANAGEMENT 128 ACADINSTR ComputerServers LYNN B205 1240 VETMEDADMN 484 ACADADMIN ServerRoom MANN 112 1079 ADVCOMPUTNG 480 RESCHCENTER MachineRoom MATH 503 1399 STATISTICS 161 ACADINSTR ComputerEquipment MATH 937 1390 SCIENCEADMN 135 ACADADMIN MachineRoom MATH B060 1061 ITSYS&OPS 1,957 ADMCOMPUTNG ComputerRoom MATH G109 1061 ITSYS&OPS 3,093 ADMCOMPUTNG ComputerRoom MATH G190 1061 ITSYS&OPS 3,890 ADMCOMPUTNG Computer MJIS B067 1296 BIOMEDENGR 338 ACADINSTR ServerRoom MSEE 130H 1301 ENGCPTNTWK 1,103 COMPUTING TapeDrives PAO B127 1352 VIS+PERFART 310 ACADINSTR Server/Maintenance PHYS 140 1396 PHYSICS 235 ACADINSTR PCNComputers PHYS 394A 1396 PHYSICS 1,020 ACADINSTR PCNMachineRoom PIT1 100 1077 ITNTWK&SCTY 187 UTILITIES SatelliteTransmission+BDF PMU 021A 1740 PMU 46 AUXSTUDENT Office,ServerRoom MORT DataCenterReport Page10
POTR B016 1301 ENGCPTNTWK 713 COMPUTING ECNMachineRoom REC 413 1394 CERIAS 158 RESCHCENTER ComputerEquipment SAT 100 1744 UNIVRESDIR 261 AUXSTUDENT BoilerTV+Telecomm BDF SCCE 24 1131 PURMKTGCOM 101 PR/DEVELOP ServerRoom SERV 6 1678 PFCOMPUTING 648 PHYSPLNTAD SERV 007D 1678 PFCOMPUTING 702 PHYSPLNTAD SMTH 113 1144 AGINFOTECH 577 COOPEXTENS ComputerServer STEW B001K 1079 ADVCOMPUTNG 187 ADMCOMPUTNG ServerRoom STEW B001M 1079 ADVCOMPUTNG 276 ADMCOMPUTNG Server+ClusterRoom STEW 35 1593 CTRINSTREX 1,013 ACADINSTR Computer+DataProcessing STEW 61 1593 CTRINSTREX 611 ACADINSTR Computer+DataProcessing STON B004 1260 CFSADMIN 171 ACADADMIN ServerRoom STON B006B 1260 CFSADMIN 165 ACADADMIN ComputerMachineRoom TEL 210 1077 ITNTWK&SCTY 3,514 UTILITIES SwitchgearRoom UPOF 220 1678 PFCOMPUTING 635 PHYSPLNTAD PFComputingServerRoom VTCH B105D 1148 AGSTAT 230 COOPEXTENS ComputerEquipmentRoom WTHR 182 1393 CHEMISTRY 321 ACADINSTR TheoreticalChemistry Computer WTHR 182B 1393 CHEMISTRY 140 ACADINSTR TheoreticalChemistry Computer WTHR 260A 1393 CHEMISTRY 104 ACADINSTR ComputerEquipment YONG 442 1410 TECHADMIN 142 ACADADMIN TCNEquipment YONG 547 1408 ITTEACH&LRN 395 COMPUTING TestLab MORT DataCenterReport Page11