The Role of ICT in Energy Consumption and Energy Efficiency

EMPA20090243

Project Number 224017 ICT-ENSURE: European ICT Environmental Sustainability Research Call identifier: FP7-ICT-2007-2 Funding scheme: Coordination and support action

The Role of ICT in Energy Consumption and Energy Efficiency Lorenz M. Hilty, Vlad Coroama, Margarita Ossés de Eicker, Thomas F. Ruddy, Esther Müller Technology and Society Lab Empa, Swiss Federal Laboratories for Materials Testing and Research St.Gallen, Switzerland

Materials Science & Technology

The Role of ICT in Energy Consumption and Energy Efficiency

Executive Summar y DespitethefactthatInformationandCommunicationTechnologies(ICTs)areresponsibleforonlyasmallpart ofworldwidegreenhousegasemissions–currentestimationsattributearound2%ofmanmadeemissionsto ICT–thissectoristheonewiththefastestgrowingemissions.Asaresult,thereisanincreasingconcernabout the environmental impact of ICT, especially the climate change potential induced by ICTrelated energy consumption. At the same time, there is a growing perception that ICT can also substantially reduce the environmental impactsofothersectors,inparticularbyincreasingtheirenergyefficiency.ICTcanhelpalleconomicsectorsto becomemoreenergyefficient–sinceICTallowsexistingprocessestobeoptimizedorenablesentirelynew, moreenergyefficientprocesses.TheenergythatcouldbesavedbyICTinducedenergyefficiencyisestimated to be several times larger than the overall energy consumption of ICT itself. The European Commission recognizes this potential and hopes that Europe will go a long way toward achieving its target of 20% greenhousegasreductionby2020bydeployingICTforenergyefficiency. The present study looks at the field spawned by these two main issues at the intersection between ICT and energy:ICT’sownenergyconsumptionandICT’spotentialtoinduceenergyefficiencyacrosstheeconomy.Inits approachtotheseissues,thestudylooksbothattoday’ssituation,aswellasfutureopportunitiesandrisks. Thestudydiscussesthefollowingresearchquestions: a)estimatesofthecurrentenergyconsumptionofICT, b)prospectivefuturedevelopmentsinthisenergyconsumption,and c)futureenergyefficiencypotentialsinducedbyICTinvariouseconomicsectors. ICTrelated energy efficiency potentials already realized are not covered by this study, because it is virtually impossibletoretrospectivelyallocateadvancesinenergyefficiencytothevariouschangesthatcreatedthem. Twomethodologieshavebeenusedforthestudy:literaturereviewandexpertinterviews.Fortheformer,we havereviewed: i) recent quantitative studies (since 2005) with a focus on “ICT energy consumption”, “ICT for energy efficiency”,“GreenI(C)T”,“ICTandclimatechange”or“ICTandsustainability”ingeneral; ii) other documents describing projects, programs or initiatives aimed at reducing ICTrelated energy consumptionorincreasingICTrelatedenergyefficiency; iii)LifeCycleAssessment(LCA)studiesonICTproductsandservices,and iv)studiesonthepotentialofsmartpowernetworks(smartgrids). Wedecidedtoincludethemorespecifictopicsiiiandivbecausetheyarenotsufficientlycoveredbyitemsi andii. These sources were then evaluated and the relevant content structured according to relevance (Chapter 2), stateoftheart(Chapter3),andresearchprogrammes(Chapter5)inordertogivethereaderinsightintothe motivationdrivingthisresearcharea,currentknowledge,andthefocusofongoingresearch,respectively. Chapter4 presents the results of expert interviews based on a questionnaire which we developed to fill the knowledgegapsidentifiedintheliteraturereview.Theaimoftheinterviewswastocollectideasbeyondthe currentstateofquantitativeknowledgeandtoidentifyresearchquestionsnottodoarepresentativesurvey. Theexpertswereonlyaskedaboutfuturedevelopments(researchquestionsbandc).ForICT’sfutureenergy consumption (b), the experts were asked to estimate for different categories of technologies (such as “data centres”,or“embeddedICT”)howtheirrespectiveglobalenergyconsumptiontotalswouldevolveinbotha businessasusualscenario(alongsidetheforeseeabletechnological,political,andmarketdevelopments)andin

© ICT-ENSURE consortium: all rights reserved

page

iii


an “energyoptimistic” scenario, in which energyreducing measures would be rigorously applied. As for question c above (ICT’s potential for energy efficiency), the experts were presented with the possible applicationareasinwhichthedeploymentofICTwasexpectedtoleadtobetterenergyefficiency,andwere asked to estimate their relative importance. Furthermore, the experts were asked to determine which ICT categorieswererelevantforinducingenergyefficiencyinothereconomicsectors. Inouranalysisofthecurrentsituationandthefuturepotentialofboth“ICTenergyconsumption”and“ICTfor energyefficiency”,threemainresultsbecomeevident: x

A thorough overview of the stateoftheart literature for all three questions considered, with emphasisontheLCAmethodologyandincludingaformaldefinitionofICTrelatedenergyefficiencyas wellasaconceptualframeworkoftheeffectsofICTonenergyefficiency.

x

Anoverviewofexistingresearchprogrammes,projectclusters,andinstitutionsinvolvedinthem,both intheEUandbeyond.

x

The results ofexpert interviews regarding future ICT energyconsumption and future applications of ICTforenergyefficiency.Inadditiontocomparingbusinessasusualwithenergyoptimisticscenarios, thus revealing where the largest energysaving potentials for ICT lie, the experts have also – as a novelty–relatedtheconsumptionofindividualtechnologiestotheirrespectivepotentialsforinducing energyefficiency.

Atamoredetailedlevel,theresultsshowthatsomeapplicationfields(suchas“TVsandsettopboxes”)are expectedtodrasticallyincreasetheirenergyconsumptionwithoutcontributingtoenergyefficiencyinanyway, while others (such as “embedded ICT”), although increasing their collective energy consumption as well, are expectedtoplayacrucialroleinenergyefficiencyacrosstheeconomy.Wehopethatthisfreshthinkingwill helptointroduceamoredifferentiatedviewofICTintothepublicdiscourseandpoliticaldecisionmaking.


page

iv


Contents Executive Summary ........................................................................................................................... iii 1 Basic Definitions and Scope of the Study ......................................................................................1 1.1

1.2

1.3

1.4

Information and Communication Technologies (ICT) ...............................................................1 1.1.1

Existing Definitions .................................................................................................................... 1

1.1.2

Reasons for Definitional Difficulties – Rapid ICT Development ................................................. 2

1.1.3

Types of ICT Considered in this Study ...................................................................................... 3

Energy Consumption.................................................................................................................3 1.2.1

General Definition and Considerations ...................................................................................... 3

1.2.2

ICT-Related Energy Consumption ............................................................................................. 4

Energy Efficiency ......................................................................................................................5 1.3.1

General Definition and Considerations ...................................................................................... 5

1.3.2

ICT-Related Energy Efficiency................................................................................................... 6

Scope and Methodology of the Study .......................................................................................6 1.4.1

Selection of Literature ................................................................................................................ 6

1.4.2

Expert Interviews ....................................................................................................................... 7

2 Relevance of ICT-Related Energy Consumption and Energy Efficiency .....................................8 2.1

2.2

2.3

2.4

Importance of Sustainability Research in this Field ..................................................................8 2.1.1

Outstanding Opportunities and Risks ........................................................................................ 8

2.1.2

Relative Importance of ICT Regarding Energy Consumption .................................................... 9

Scientific Interest in this Field .................................................................................................12 2.2.1

Emergence of “ICT and Sustainability” as an Interdisciplinary Research Field........................ 12

2.2.2

Issues of Scientific Methodology ............................................................................................. 12

Interest of the ICT Sector in the Field .....................................................................................14 2.3.1

Private-Sector Initiatives .......................................................................................................... 14

2.3.2

Industry-Supported Studies ..................................................................................................... 15

Political Relevance of the Field ...............................................................................................16 2.4.1

EU Initiatives: Emergence of the Issue “ICT for Energy Efficiency” ......................................... 16

2.4.2

OECD ...................................................................................................................................... 19

2.4.3

UN Initiatives ........................................................................................................................... 19

3 State of the Art .................................................................................................................................20 3.1

Conceptual Framework ...........................................................................................................20

3.2

EU-Funded Projects and Studies Contributing to the Field ....................................................23

3.3

3.2.1

EU-Funded Projects ................................................................................................................ 23

3.2.2

EU-Funded Studies ................................................................................................................. 25

Measuring ICT-Related Energy Consumption ........................................................................26 3.3.1

Approaches Based on Final Energy Consumption .................................................................. 26

3.3.2

The Link between Energy Consumption and CO2 Emissions .................................................. 29


page

v


3.4

3.3.3

LCA-Based Methods and Results ............................................................................................ 30

3.3.4

From the Micro- to the Macro-Level ......................................................................................... 32

Estimating ICT-Related Energy Efficiency ..............................................................................33 3.4.1

Macro-Level Estimates of the Impact of ICT on Energy Efficiency .......................................... 33

3.4.2

Example: Direct Comparison of Virtual and Physical Meetings ............................................... 34

3.4.3

Example: Direct Comparison of Electronic and Print Media .................................................... 36

3.4.4

Example: ICT-Related Energy Efficiency Potentials in Power Generation and Distribution Including Demand-Side Management ..................................................................................... 37

4 Future Potential – Results of Expert Interviews...........................................................................40 4.1

Which Technologies Belong to the ICT Sector? .....................................................................40

4.2

Reducing ICT-Related Energy Consumption ..........................................................................42

4.3

4.4

4.2.1

Business as Usual ................................................................................................................... 42

4.2.2

Potential of Technical and Organisational Measures ............................................................... 42

Unleashing ICT-Related Energy Efficiency Potentials ............................................................43 4.3.1

Survey Results ........................................................................................................................ 43

4.3.2

Future “Killer Application” ........................................................................................................ 44

Future Research Needs and the Relative Importance of Specific Research Fields ...............45 4.4.1

ICT and Energy Efficiency Research ....................................................................................... 45

4.4.2

Quality of Data and Research Needed for Improvement ......................................................... 45

5 National and International Research Programmes ......................................................................47 5.1

EU-funded Programmes .........................................................................................................47

5.2

Selected National Programmes ..............................................................................................49 5.2.1

EU Member States .................................................................................................................. 49

5.2.2

Outside the European Union ................................................................................................... 52

6 Conclusion .......................................................................................................................................53 Annex 1: Involved Organisations and Research Institutes..............................................................55 Annex 2: List of Experts ......................................................................................................................57 Annex 3: Interview Outline ..................................................................................................................60 Annex 4: References ............................................................................................................................67


page

vi


1 Basic Definitions and Scope of the Study

1.1 Information and Communication Technologies (ICT) 1.1.1

Existing Definitions

Providingaformaldefinitionfor“InformationandCommunicationTechnologies(ICT)”ismoredifficultthanit mightseematfirstglance.Whileeveryoneseemstohaveanintuitiveunderstandingoftheterm’smeaning, definition attempts are surprisingly sparse. It is characteristic, in this context, that EU’s own Seventh Framework Programme, for example, does not formally define ICT, although it foresees 9.1 billion Euro for fundingit. 1 Moreover,bothprintedandonlineencyclopaediashaveratherpoordefinitionsoftheterm. 2 ForevaluatingtheroleofICTinenergyconsumptionandenergyefficiency,adefinitionofwhatconstitutesICT (and what not) is nevertheless needed – a need that has already been faced by national and international organizationswhenmeasuringthesizeoftheICTsectorinsideeconomies. 3 Itthuscomestolittlesurprisethat theterm’sfewexistingdefinitionscomefromtheseorganisations. ISIC The “International Standard of Industrial Classification of All Economic Activities” (ISIC), developed by the UnitedNation’sStatisticDivision,isalargelyusedstandardforclassifyingeconomicactivities.Whileitslatest revision(Rev.4) 4 containsforthefirsttimeatoplevelcategory“Informationandcommunication”(cat.J),the ISICclassification,evenonitsmostdetailedlevel,consistsofquitegeneralsubcategories.Furthermore,ICTis definedinawidesense,includingsocalled“contentindustries”enabledbyICT.Whethercategoriessuchas “591 – Motion picture, video and television program activities”, or “602 – Television programming and broadcastingactivities”belongtoICTornotisdebatable–mostexpertsdonotseethemaspartofICT(Plepys 2004; OECD 2005). Other products though, such as digital cameras, digital music players, etc. – more likely belonging to ICT and continuously gaining importance, both in economic terms and with regard to energy consumption–donotappearintheISICclassification. OECD Acknowledgingthesameneed,theOrganisationforEconomicCoOperationandDevelopment(OECD)started in 1997 its efforts towards a definition of the ICT sector, under the guidance of the newlyformed “Working PartyonIndicatorsfortheInformationSociety”(WPIIS)(OECD2005).

1

Seehttp://cordis.europa.eu/fp7/ict/(forthebudget),andftp://ftp.cordis.europa.eu/pub/fp7/ict/docs/ictwp200910_en.pdf.

2

This is also true for the wellknown onlineencyclopaedia “wikipedia,” where the term “ICT” is, as of January 2009, still very poorly defined.Seehttp://en.wikipedia.org/wiki/Information_communication_technology. 3

As put by http://www.oecd.org/dataoecd/5/61/22343094.pdf, “The main reason to have a classification of information and communication technology (ICT) goods is to facilitate the construction of internationally comparable indicators on ICT trade and ICT production. Such a classification would also provide a basis to develop internationally comparableindicators for ICT consumption and investment.” 4

Seehttp://unstats.un.org/unsd/cr/registry/regcst.asp?Cl=27.


page

1


In defining the ICT sector, the OECD started from the same ISIC categories of goods and services. It went, however,beyondasimpleenumerationofcategories.TheOECDexpertstriedinnumerousmeetingstoanswer boththequestionaboutwhatrepresentstheessenceofanICTproduct–i.e.,ofanICTgoodorservice–andof how the relevant parameters are best measured and expressed (OECD 2005), annex 1b. After years of discussions, compromises, and incremental modifications, nowadays’ OECD definition of the ICT sector is composedofthreemainpillars:ICTmanufacturing,trade,andserviceindustries(OECD2005).Manufacturing industries include the production of goods such as circuit boards, computers, or magnetic and optic media; trade industries comprise the wholesale of computers and other electronic equipment; services include computerprogramming,webhosting,orICTconsultancy. On a conceptual level, the OECD states that ICT goods “must either be intended to fulfil the function of information processing and communication by electronic means, including transmission and display, or use electronicprocessingtodetect,measureand/orrecordphysicalphenomena,ortocontrolaphysicalprocess” (OECD 2002). ICT services, on the other hand, “must be intended to enable the function of information processingandcommunicationbyelectronicmeans”(OECD2002). The components of the three main categories (ICT manufacturing, trade, and services) are presented as ISIC subcategoriesaswell.Fordoingso,theOECDclassificationcombinessome(butnotall)componentsofISIC’s “Information and communication” section with subcategories from other ISIC sections (such as section C “manufacturing”). Furthermore, and for ICT goods only, a further taxonomy is used as well – the World Customs Organization's “Harmonized System” (HS). The 6digit Harmonized System is not only much more detailedthanISIC’srelativelygeneralcategorization,butalsooffersauniqueadvantage:whiletheISICsystem canbeusedfornationalstatistics,internationaltradeismeasured(ifatall)onlythroughtheHS:“TheHSisthe only commodity classification system used on a sufficiently wide basis to support international data comparison.Alargenumberofcountriesuseittoclassifyexportandimportofgoods,andmanycountriesuse it(oraclassificationderivedfromit)tocategorisedomesticoutputs”(OECD2003). As a last remark, and as might alreadyhave become clear from the summary above, very early in the OECD processthedecisionhasbeentakentoexcludethesocalled“content”industriesfromthedefinitionoftheICT sector.ContentindustriesofferservicessuchasTVproductionorTVandradiobroadcast.IntheviewofOECD, abroader“informationeconomy”sectorexists,whichencompassestheICTsectortogetherwiththecontent industry: “In the view of the members of the Panel, the ‘information economy’ consists of the economic activitiesofthoseindustriesthatproducecontent,andoftheICTindustriesthatmoveanddisplaythecontent” (OECD2005).

1.1.2

Reasons for Definitional Difficulties – Rapid ICT Development

The pace of progress for information and communication technologies – best represented by the socalled Moore’s Law (Moore 1965) – does not only imply exponential growth of storage and computing capacity as well as bandwidth per size and price. It also means that through these rapid advances, formerly non computerizedentities–fromindividualgoodstoentireeconomicprocesses–becomeincreasinglyICTbased. Thiscontinuousshifttowardsthe“digitalization”oflifehasthreeeffects,whichallhinderaprecisedefinition oftheICTsector: x

Any enumeration or categorization of the sector will quickly become outdated. New (digital) technology appears at a quick pace. Roughly 15 years after their first appearance, for example, digital cameras are muchmorepopularthantheiranaloguecounterparts.Inevenlesstime,roughlyonedecade,audioplayers (“mp3players”)havealmostentirelyreplacedportableanaloguemusicdevices.Thefactthatthisunusual dynamicityoftheICTsectorcombinedwithslowlychangingclassificationsofgoodsandservicesmakesan inventory of ICT components rather challenging, as has been noticed by OECD’s Working Party on Indicators for the Information Society: “The difficulties in establishing a list of ICT products have been recognised by WPIIS since 1998. These difficulties were related to the rapidly changing character of ICT goodsandservices,andthedatednatureofcurrentstandardclassifications”(OECD2003).


page

2


x

Thegeneraltendencyofgoodstobeenhancedwith(andofservicestomakeuseof)ICTtechnology.This trend, called “Ubiquitous Computing” (Mattern 2005b) or “Pervasive Computing” (Hilty, Behrendt et al. 2005), observes that due to the dramatically sinking costs and size, and the equally improving performance, ICT components start to be included in more and more objects and products: in cars, for example, they enable safety features such as ABS (antilock braking system), EPS (electronic stability control),andnavigationsystems;inprintercartridgestheycountthenumberofpagesalreadyprintedand incoffeemachinesthenumberofcoffeesbrewed;andincludedattheendnodesofpowerlinestheyallow Internetconnectivitythroughthepowergrid.Inanotsodistantfuture,washingmachinescouldexchange data with the shirts (receiving thus automatically washing instructions and freeing the user from programmingthem).Moregenerallyeverydayobjectscouldknowtheirlocationandhistoryofusage,as wellasinspecttheirownstatus(allowingasheerendlessamountofpossibleapplications).Inthissense, ICTcouldbeincludedinmostobjectsandservices,makingadistinctionbetweenICTand“nonICT”ever moredifficult.

x

SeveralservicesthatnowadaysundoubtedlybelongtoICThavealwaysbeenrelatedto“information”in thewidesenseoftheword.While,forexample,digitalphotography(togetherwiththeenablingdevices, digitalcameras)isatthecoreofICTconsumerelectronics,theaimofanaloguephotographyhasalways beentopreserveinformation.DoesthedigitalcameraindustrythusrepresentanewmemberoftheICT family or has the (analogue) camera industry always been part of the ICT sector? Even more strikingly: analogueTVandprintednewspapershavealwayshadthepurposetoinformthepublic.Donoveldelivery technologiessuchasdigitalTVbroadcast,TVovertheInternet,orelectronicnewspapersmakeasemantic difference in terms of classifying the service within or outside the ICT sector? More generally: Does ICT implicitlyrelateexclusivelytothe“digitalrevolution”,i.e.,todigitallystoredortransmittedbits?Ordoes it,attheotherendofthescale,coveranygoodorservicerelatedtoinformationandcommunicationin thewidesense?

1.1.3

Types of ICT Considered in this Study

Followingthetypicalcategoriesfoundinotherstudies(e.g.,(BioIntelligenceService2008)),weconsiderthree types of ICT: servers, enduser devices, and the network infrastructure typically used to communicate either between two or more enduser devices, or among enduser devices and servers. We have not followed the oftenencountered (and increasingly artificial) separate clustering of communication devices (such as cellular phones) and computation devices, such as personal computers. For the same reason, we do not try to discriminatebetween“communicationinfrastructure”and“computinginfrastructure.” Ascanbeseenfromtheexpertanswerspresentedinsection4.1,theyallconsiderentertainmenttechnologies such as TV sets and settop boxes as belonging to ICT and significantly contributing to the overall energy consumptionofthesector.TheyarethusdefinitelyconsideredaspartofICT.Thereislessconsensuswhether otherentertainmenttechnologiessuchasdigitalmusicplayersordigitalcamerasbelongtoICTornot. Furthermore,embeddedICTcomponentsdodefinitelybelongtotheICTsector,ascanbeseenfromthesame answers.Forcontentindustries,ontheotherhand,wefollowtheconventionsmentionedaboveanddonot considerthemaspartofICT.

1.2 1.2.1

Energy Consumption General Definition and Considerations

In general, energy consumption is the transformation of energy from a usable form into an unusable form. Finalenergyconsumptionreferstotheamountofenergytransformedatthepointofuse(e.g.inanelectronic


page

3


device),whereasotherindicatorsofenergyconsumption(suchasCumulatedEnergyDemand,CED)areinuse tocoverallnecessaryenergytransformationsinasystemthatprovidesenergy(oranyothertypeofservice). Energyconsumptioncanrefertoanyenergycarrier,suchaselectricity,naturalgas,fuels,biomass,hydrogen, solarpower,etc.However,inmostofthestudiesonenergyconsumptionbyICTtheterm“energy”implicitly focuses on (final) electricity (IBM 2006; EPA 2007; Koomey 2007; DEFRA 2008; Fichter, Clausen et al. 2008; Fraunhofer IZMISI 2008). The study by (BioIntelligenceService 2008) explicitly focuses on electricity. Fewer studiesincludee.g.energyintheformoffossilfuelsthataredirectlyconsumedbytheICTindustry. Furthermore,mostofthestudiesassessingtheenergyconsumptionofICTconsideronlytheusephaseofICT products.TheconsumptionofelectricityorotherformsofenergyduringotherlifecyclephasesofICTproducts (in particular hardware production and disposal) is considered only in few studies (Malmodin 2007; Mingay 2007b;BioIntelligenceService2008;GeSI2008b). 5 ThetypesofICTproductsunderstudyareoftendefinedwithdifferentscopesandinvestigatedwithdifferent methodologies,whichmakesitdifficulttocompareresultsacrossstudies.Furthermore,theinconsistentuseof the term “cumulated energy” could lead to misunderstandings. The reason is that this term is similar to the term “Cumulated Energy Demand (CED)”, an environmental impact indicator commonly used in Life Cycle Assessmentwhichincludesallprimaryenergyneededbyaproductrelatedsystem.Finally,whenconsidering theinfluenceofICTontheenergyefficiencyinothersectors,itisinevitabletoaddressnotonlyelectricity,but allrelevantformsofenergy.

1.2.2

ICT-Related Energy Consumption

Forthepurposeofthisstudy,wedefineICTRelatedenergyconsumptionasfollows: ICTRelatedenergyconsumption(orICTenergyconsumptionforshort)istheamountofenergyconsumedbya givenICTsysteminagivenperiodoftime. This definition has several parameters which have to be set depending of the context. The most obvious parameteristhetimeperiod,whichinstatisticalcontextsisusuallysettooneyear.Giventhatweuseenergy consumptionasrelatedtoatimeperiodasanindicator,thisindicatorcouldmoredirectlybeexpressedinunits ofpower(Watt)andnotenergy(Joule,Wattseconds,Kilowatthours,etc.).However,itiscommontoexplicitly write down “energy per time” such as kWh/a, although this fraction could in principle be reduced to kW/(365*24),yieldingtheaveragepower. A less obvious parameter of the definition is represented by the broad spectrum between final energy consumption (only the energy that is transformed within the ICTSystem under study) and the cumulative energydemand(asdefinedinLCAmethodology). Thethirdparameteristhe"ICTsystem"itself.Theborderlinesofthissystemmaybedefinedaccordingtothe typeofICT,geographicboundaries,ownership,theservicetheICTsystemprovidesoracombinationofthese criteria.Itis,forexample,meaningfultoaskforthetotalannualenergyconsumptionofallPCsintheworld,of allICTinEU27,ofonespecificdatacentre,orofaspecificserviceGoogleprovidestoInternetusers. AssoonaswefocusonservicestodrawtheborderlinebetweentheICTsystemunderstudyandtherestofthe world, several methodological issues creep into the delineation task. The first issue is where to cut off the system.Forexample,isseemsnaturaltoincludetheenergyusedforcoolingintheenergyconsumptionofthe data center, simply because cooling is a necessary part of producing the service. However, will lighting also

5

(BioIntelligenceService2008)focusontheusephasebutincludesananalysisoftheproductionandtheendoflifetreatment;(Mingay 2007b)takesintoaccountthedesign,manufactureanddistributionanduse;(Malmodin2007)considersfuelchain,powerplants,thegrid, manufactureanduseand(GeSi2008a)consideredinadditiontotheusephasetheproductionandendoflifetreatment,whensuchdata wasavailable.


page

4


havetobeincluded,then?Whataboutthefacilitymanagementserviceandtheirvehicles?Thesecondissueis allocation. If the data center produces more than one type of output, i.e. we are interested in the energy consumptionofWebhosting,butthedatacenteralsoprovidesdatabackups,whichpartoftheconsumption hastobeallocatedtowhichoutput? ItisimportanttounderstandthattheseissuescannotbesolvedbyprovidingaformaldefinitionofICTenergy consumption, but only by understanding the objectives and context of each study. However, there is some methodologicalsupportfromtheLCAfield.Wewillcomebacktotheseissuesinsection3.3.3.

1.3 1.3.1

Energy Efficiency General Definition and Considerations

The energy efficiency of a system A is the ratio of the useful output of services from A to the energy consumptionbyA.Usually,boththeoutputandtheenergyconsumptionarerelatedtoaperiodoftime,which obviouslyleadstotheeliminationoftimeandyieldsthedimension"servicesperenergy",suchaskm/kJfora vehicle,kg/kWhforarecyclingprocess,orkByte/WsforanInternetservice. LetSbeameasureoftheserviceoutputofasystemandPtheenergyconsumptionofthesystem(asdefinedin 1.2.1.above).Thentheenergyefficiencycansimplybedefinedas =S/P Please note that according to this definition energy efficiency is different from energy conversion efficiency (sometimesalsocalled“energyefficiency”forshort),whichphysicsdefinesas =Pout/Pin where Pout refers to the useful energy output of the system (which is an energy conversion machine in this case)andPintotheenergyinput.Obviously,energyconversionefficiencyisaspecialcaseofenergyefficiency : in this special case, the service consists in providing energy in a specific form and energy consumption is measuredasdirectenergyinput.ForICTsystemswecangenerallyassume = 0, becauseallenergyisfinally converted to waste heat, while the purpose of an ICT system is not to provide heat. However, for some componentsofICTsystemssuchaspowersuppliesormotors,thereisofcoursea > 0. Very often, the expression "energy efficient" is used as a unary predicate, as in "A is energy efficient"; this presumesanimplicitcomparisonofAwithasystemB,meaninginfactthattheenergyefficiencyofAbehigher thantheenergyefficiencyofB.Itisthereforepreferabletoexplicitlystatethat"Aismoreenergyefficientthan B"insuchacase,i.e.touseabinarypredicate.Inordertomakequantitativecomparisons,wedefinerelative energyefficiencyastheratiobetweenthetwoenergyefficiencies: A,B = A / B = SA PB / PA SB ThisistherelativeenergyefficiencyofAwithregardtoB.Inmanypracticalsituations(andascommoninLCA studies),weassumethattheservicesAandBarefunctionallyequivalent.Inthiscase,relativeenergyefficiency ofAwithregardtoBreducesto A,B = PB / PA (with SA = SB) i.e.itexpressesbywhichfactorenergyconsumptioncanbedecreasedbysubstitutingAforB,allotherthings beingequal.Inmanystudiesonenergyefficiency,implicitreferencetothistypeofrelativeenergyefficiencyis made,usuallybystatingenergysavingspotentialsinpercent(1–1/A,B) ,oftenwithoutexplicitlyaddressing systemB. We have assumed so far that the system under study only produces one useful service output. If there are moreusefuloutputs,measuringtheenergyefficiencywithregardtoonespecificoutputrequiresdefiningan


page

5


allocation scheme which assigns a part of the total energy consumed (P) to the output of interest. The allocationproblemisageneralmethodologicalissueinmodellingmultioutputprocesses(e.g.inLCAstudies) andnotadefinitionalproblemofenergyefficiency.

1.3.2

ICT-Related Energy Efficiency

GiventwosystemsAandBproducingafunctionallyequivalentoutputofservicesSA = SB.LetusassumethatA containsasubsystemCwhichisanICTsystemandBdoesnotcontainthatsubsystem,allotherthingsbeing equal. 6 Wecanthendefine: TheICTrelatedenergyefficiencyofAwithregardtoCistherelativeenergyefficiencyofAwithregardtoB (withB=A\C). In other words, ICTrelated energy efficiency is the factor by which the energy consumption of a system decreasesifanICTsystemisaddedtoitandallotherthingsarekeptequal(inparticulartheserviceoutput). Veryoften,thisisexpressedinpercentofenergysavedinsteadofareductionfactor. ICTrelatedenergyefficiencyissspecialcaseofrelativeenergyefficiencyA,B (asdefinedabove)withSA = SB and B=A\C,CbeinganICTsystem.

PA

C

A

PB / PA PB

SA SB = SB

B

SB

Figure1.SystemmodelforthedefinitionICTrelatedenergyefficiency

1.4 1.4.1

Scope and Methodology of the Study Selection of Literature

Theliteraturesearchforthisstudywasfocusedonfourareas: 1.

Recent quantitative studies (since 2005) with a focus on "ICT energy consumption" "ICT for energy efficiency","GreenI(C)T","ICTandclimatechange"or“ICTandsustainability”ingeneral;

2.

Other documents describing projects, programs or initiatives aimed at reducing ICTrelated energy consumptionorincreasingICTrelatedenergyefficiency;

3.

LifeCycleAssessment(LCA)studiesonICTproductsandservices.

4.

Studiesonthepotentialofsmartpowernetworks(smartgrids).

6

Inpractice,Bmayoftenbeahypotheticalsystem.


page

6


Wedecidedtoincludethemorespecifictopics(3)and(4)becausethesearenotsufficientlycoveredbythe generalstudieswefoundaccordingtocriteria(1)and(2). ThesesourceswerethenevaluatedandtherelevantcontentstructuredaccordingtoChapters2(Relevance),3 (Stateoftheart),and5(Researchprogrammes)inordertogivethereaderasurveyonthemotivationbehind thisresearcharea,thecurrentknowledge,andthefocusofongoingresearch,respectively. In parallel, the literature survey was used to create a questionnaire for expert interviews (see next section) withtheaimtoenrichtheestablishedknowledgewithestimatesoffuturepotentialsandnewideas.

1.4.2

Expert Interviews

The aim of the expert interviews was to collect ideas beyond the current quantitative knowledge and to identify research questions, not to do a representative survey. We also wanted to give the authors or commissionersofthemainstudiesweevaluatedtheopportunitytoprovideuswithanupdateoftheirview, whichmayhavechangedsincethestudywenttopress. We therefore primarily invited experts involved in the studies we used as literature sources for our study. Although this was an international and unbiased selection, we mainly got answers from Germanspeaking countries(Germany,Austria,Switzerland). ForthequestionnaireseeAnnex3:InterviewOutline.


page

7


2 Relevance of ICT-Related Energy Consumption and Energy Efficiency

2.1

Importance of Sustainability Research in this Field

Informationandcommunicationtechnologies(ICTs)arecontinuouslymakingastoundingprogressintechnical efficiency. The time, space, material and energy needed to provide a unit of ICT service have decreased by threeordersofmagnitude(afactorof1000)sincethefirstPCwassold.However,itseemstobedifficultfor societytotranslatethisefficiencyprogressintoprogressintermsofsustainabledevelopment. Basically, the idea of an information society has a huge potential to solve the dilemma of sustainable development,whichis:Providingqualityoflifetoallpeoplewithoutoverusingtheecosystem.Thisdilemmacan only be solved if society manages to create value with much less material and energy input. As has been discussedfordecadesnow,a‘dematerialization’oftheeconomicsystembyafactorof410isaprecondition for sustainability. Creating an information society which makes use of ICTs to provide immaterial services where previously material goods were produced, transported and disposed of, could be a key to economic dematerialization(Hilty2008). TheEuropeanCommissionrecentlydefinedtheroleofICTsinasimilarwaybystatingthat“…thecontinued growthoftheEuropeaneconomy[…]needstobedecoupledfromenergyconsumption[…]Indeed,ifnothing were to change, final energy consumption in the EU is predicted to increase up to 25 % by 2012, with a substantial rise in greenhouse gas emissions. Information and Communication Technologies (ICTs) have an importantroletoplayinreducingtheenergyintensityandincreasingtheenergyefficiencyoftheeconomy“ (EuropeanCommission2008c),p.2. The Global Information Infrastructure Commission (GIIC) recently stated in their Tokyo Declaration: “ICT has historically been viewed as a tool to advance productivity. We found and confirmed that the use of ICT can changethebehaviourofbusinessandconsumers,andthroughthesechanges,ICTcanhelptheenvironment withoutsacrificingeconomicoutput“(GIIC2008),p.2.

2.1.1

Outstanding Opportunities and Risks

ThewellknownprinciplecalledMoore’sLaw(Moore1965),accordingtowhichthenumberoftransistorsper microchipdoublesevery1824months,hasyettobedisproved.Asasideeffect,processorperformanceper energyinputgrowsexponentiallytoo(Figure2).

Figure2.Moore'sLawandtheincreaseincomputingpowerperunitelectricalpower(Source:(Mattern2005a))


page

8


However, this astounding progress does not reveal much about the influence of ICT on overall energy consumptionfortworeasons: 1.

Despite the increasing energy efficiency of ICT hardware, the total energy demand of the installed hardwarebaseisgrowing.ThisisbecausethedemandforICTservicesisincreasingevenfasterthan the energy efficiency of ICT devices. More and more powerful devices are used by more and more people.

2.

ICT is an enabler of energy efficiency in sectors that use much more energy than the ICT sector. If these efficiency potentials are used systematically, ICT can make a substantial contribution to the reductionofenergydemandandthereforetoalowcarboneconomy.

This implies that sustainability research on energetic aspects of ICT must look far beyond the efficiency of individualICTdevices.ThecomplexdynamicsofICTimpacts–inthepositiveornegativesense–onenergyuse mustbeunderstoodbymeansofdynamicmodels. Asimulationstudyonthe“FutureimpactofICTonenvironmentalsustainability”inEU15demonstratedthat the impact of ICT on environmental indicators (including total energy demand) can be either positive or negative below the line, depending on the framework conditions assumed in the scenarios that were simulated.Theseframeworkconditionsincludedenergyprices,anexternalvariablethathadastronginfluence ontherealisationofefficiencypotentialsandontheoccurrenceofreboundeffects.Anotherimportantresult was that effects on the energy demand of specific sectors or areas of activity were usually much higher (in positive or negative directions) than the aggregated effects. In each scenario, some of the ICT effects counterbalancedeachotherwhentakentogether(Erdmann,Hiltyetal.2004;Hilty,Wägeretal.2004). WecandeducefromthisthatawelldesignedpolicythatsystematicallyreinforcesthepositiveeffectsofICTon energyefficiencyandcounteractsICTrelatedrisksisnecessary. Thechallengeforpolicymakersseemstobethatthereisnogeneralstrategyforunleashingthepotentialsof ICTtoincreasetheenergyefficiencyoftheeconomy.Instead,ICTeffectsmustbeanalyzedandprospectively assessed at some detail. A recent review of research on the environmental impact of ICT confirms that a modelbasedapproachisneededtoallow“thatpositiveeffectscanbepromotedandnegativeonesalleviated proactively”(YiandThomas2007).

2.1.2

Relative Importance of ICT Regarding Energy Consumption

Compared to the total amounts of energy consumed by the industry, residential or transport sectors, ICT related final energy consumption does not seem to be very relevant at a first glance. If we take the EU27 energy baseline scenario by DG TREN (European Commission 2008e) and distribute the ICTrelated energy consumption of the BAU scenario used in the European Commission's recent Impact Assessment on ICT for energyefficiency(EuropeanCommission2009b),thedistributionshowninFigure3results. In2005,4.5%(120TWh/a)oftheelectricalpowerinEU27wereconsumedbyconsumerelectronics(mainly TVs and HiFis) and 3.5 % (97 TWh/a) for ICT in a narrower sense (PCs, telephones and the communication infrastructureincludingdatacentres)(EuropeanCommission2009b).Weallocatedallconsumerelectronicsto residential energy demand and split the rest (ICT in a narrower sense) equally between residential and industrialenergydemand(Figure3a). The 2020 projection is based on the baseline scenario of DG TREN and the BAU scenario of the impact assessment. According to the latter, ICTrelated energy consumption will rise to at least 400 TWh, mainly drivenbytheexpecteddiffusionoflargerscreenTVs,higherspeedbroadbandaccessorhighercapacitydata centres. Since these drivers are partly residential and partly industrial, we assumed the same distribution amongsectorsfor2020asfor2005(Figure3b).TheICTrelatedenergydemandinthetransportsectorisnot known,becausetheelectricityconsumedbyonboardICTofvehiclesisusuallynotmeasured.


page

9


Final Energy Demand EU27 by Sector (2005) 16'000

>217.0

14'000

12'000

10'000

TWh

ICT 8'000

6'000

excl. ICT

13'354

48.5 ? 168.5

4'000

5'745 2'000

3'402

4'207

0 Industry, Services, Agriculture

Residential

Transport

Total

a)

Final Energy Demand EU27 by Sector (2020) 16'000

>400 14'000

12'000

TWh

10'000

ICT

8'000

15'275

excl. ICT

89.4 6'000

? 310.6

4'000

6'576 5'101 2'000

3'597

0 Industry, Services, Agriculture

b)

Residential

Transport

Total

Figure3.FinalenergydemandinEU27bysectorandICT/nonICTrelatedconsumption,a)in2005andb)projectedin2020. No data on ICT in the transport sector, because the electricity consumed by onboard ICT of vehicles is usually not measured.(Source:OwnCalculationsbasedonDGTREN(EuropeanCommission2008e)and(EuropeanCommission2009b))


page

10


An interesting question is the ICTrelated energy consumption of the traffic sector. Although vehicles are heavilyequippedwithICT,thistypeofICTisusuallynotincludedintheestimates.Wethereforeallocatedall documentedICTconsumptiontotheindustrialandresidentialsectors.Theremaybearelevant,butstatistically neglectedICTrelatedenergyconsumptioninthetrafficsector. Overall,onecouldconcludethatICTrelatedenergyconsumptionislessrelevantthantheenergydemandof other types of technologies, such as industrial machines, nonICT household appliances (heaters, ovens, fridges),andvehicles.Afterall,ICTwillaccountfor2.55%ofEU27finalenergydemandin2020accordingto theseestimates. However,thereareseveralissuesthatdeserveattention: 1.

Final energy demand is different from primary energy demand (see also Sections 1.2.2 and 2.2.2). Dependingontheenergysupplychainsused,theshareofelectricityconsumingdevices(includingICT) inprimary(orincumulated)energydemandcanbeoverproportionalascomparedtofinalconsum ption.ThesameholdsfortheemissionsofCO2orCO2equivalents.

2.

The final energy demand of ICT alone is growing faster than the overall final energy demand. From 2005 to 2020 (Figure 3), the latter increases by 15.5 %, but the former by 84.3 %. For specific subsectorssomeauthorspredictmuchfastergrowth.Forexample,theBAUscenariofordatacentres in Germany calculated by Klaus Fichter for the German Environmental Agency (see Figure 4) extrapolateduntil2020wouldleadtoanincreaseof396%(Fichter,Beuckeretal.2009).

3.

Finally,mostoftheinnovativetechnologiesneededintheindustrial,residentialandtransportsectors toincreasetheirenergyefficiencyarepartlybasedonICT;fosteringtheuseofICTforenergyefficiency beyond baseline or BAU scenarios will therefore lead to a faster increase of ICTrelated energy consumption – although this will be overcompensated for by the resulting ICTrelated energy efficiency.

Figure4.DevelopmentoftheenergyconsumptionofserversandatacentresinGermanyincludingthreefuturescenarios startingfrom2008:(1)Businessasusual,(2)halfofdatacentresadoptbestpracticesinefficientenergyuse(3)90%ofdata centresadoptbestpracticesinefficientenergyuse(Source:(Fichter,Beuckeretal.2009))


page

11


2.2

Scientific Interest in this Field

Asafieldofinvestigation,ICTrelatedenergyconsumptionandenergyefficiency(or“ICTandEnergy”forshort) can be viewed as a part of the interdisciplinary research field “ICT and Sustainability” which analyses the opportunitiesandrisksofICTforsustainabledevelopment.Placingthe“ICTandEnergy”topicinthe“ICTand Sustainability” context has the advantage that systemic interactions between energy efficiency and other environmental,economicandsocialphenomenacometothefore.

2.2.1

Emergence of “ICT and Sustainability” as an Interdisciplinary Research Field

Thisinterdisciplinaryresearchfieldhasatleastthreeroots: 1.

ResearchdonebyeconomistsinterestedinICTrelatedinnovationasadriverofstructuralchange.One of the first studies was “Dematerialisation: The Potential of ICT and Services” commissioned by the FinnishMinistryoftheEnvironment(Heiskanen,Halmeetal.2001).

2.

Environmental Informatics, a field of Applied Computer Science in which principles and systems for the processing of environmental information and environmental modelling are developed. The 15th EnviroInfo conference, held at ETH Zurich in 2001, has already been devoted to the topic “SustainabilityintheInformationSociety”(HiltyandGilgen2001).

3.

Technologyassessment(TA)andLifeCycleAssessment(LCA)studiesonICThardwarewithatendency to broaden the perspective beyond hardware issues. Early work was done at IZT Berlin (Behrendt, Pfitzneretal.1998)andbyEricWilliams(Williams,Ayresetal.2002).

“ICTandSustainability”todayisafieldofresearchwhichsystematicallyinvestigatestheactualandpotential consequencesofICTforsustainabledevelopment.ThiscoversboththeICTrelatedpotentialtoenablelargely dematerialized or “deenergized” production and consumption, the role of ICT in emerging economies, rebound effects from efficiency increases, as well as the (minimization of) environmental impacts of ICT production,useanddisposal(Hilty2008;StreicherPorte,Marthaleretal.2009). The topic of this study, “ICT and energy“, can be considered a special case of “ICT and sustainability“, since both reducing the energy consumption of ICT and realizing the energy efficiency potentials induced by ICT systemsareessentialpartsofstrategiestowardssustainabledevelopment.

2.2.2

Issues of Scientific Methodology

Ourliteraturestudyrevealedtwomethodologicalissuesincurrentresearchon“ICTandenergy”,definingthe boundariesoftheenergyconsumingsystemanddefiningthebaselineor"BusinessasUsual(BAU)"scenario for measuring relative energy efficiency, which corresponds to “system B” in our definition of ICTrelated energyefficiency(Section1.3.2). Defining the boundaries of the energy-consuming system The most striking methodological issue is the problem of defining the boundaries of the energyconsuming system. In the case of ICTrelated energy consumption, this is the ICT system itself. In the case of “ICT for energy efficiency”, it is the energyconsuming system that is intended to be improved (regarding its energy efficiency)bytheuseofICT. Boundaries of the energyconsuming system can be drawn as narrow as possible, which means considering onlythefinalenergyconsumption(e.g.theelectricenergytakenfromthepowersocketwhichentersadevice). This narrow perspective is useful to compare several energyconsuming systems which use energy from the same source, e.g. to compare different PCs powered by the same grid. However, such comparisons are not meaningful across different energy sources, e.g. when the energy consumed bydata centres is compared to


page

12


theenergyconsumedbyroadtraffic.Inthiscase,thesystemsshouldbeenlarged(inupstreamdirectionwith regardtotheenergysupplychain)untilmatchingenergysourcesarereached,e.g.fossilfuelsusedtogenerate powerandfossilfuelsconvertedtogasoline.Ifthisisdone,alllossesoccurringinthetwoenergysupplychains (viewedfromtheircommonstartingpointindownstreamdirection)areaccountedfor. InLCAmethodology,thesystemisalwaysviewedfromitsusefuloutput:theunitofservice(calledfunctional unit) it is intended to produce. All energy and material transformations that are caused by producing this functionalunithavetobeincludedinthesystem.Eveniftheobjectiveofastudyisonlytomeasureenergy consumption and not the depletion of material resources, emissions and other environmental impacts, the materialflowsareimportantbecauseprovidingmaterialsconsumesenergy,too(sometimescalled“embodied energy”). Energy and material flows are thus closely intertwined. There is good practice in LCA studies concerningthedefinitionofsystemboundariesbasedonthegoalandscopedefinitionofastudy.Evenpurely energyrelated studies that are not interested in a “full LCA” should learn from the LCA practice in defining systemboundaries. Thereisatradeoffbetweentheaccuracyofastudyandtheaimtolimitcomplexity;thelargerthesystem,the more accurate the statements about the energy consumption caused for producing one functional unit; however,thisimplieshighercomplexityandhighereffortsfordatacollection. Thereissomeconfusioninthe“ICTandenvironment”discourseregardingthemeaningoftheterms“direct” and“indirect”,becausethesetermsareuseddifferentlyinthevariouscontextsthatmergeinthisdiscourse. SeeTable1foraclarification.(Thelistmaynotbeexhaustive.) Table1.Differentusesoftheterms“direct”and“indirect”inthe“ICTandenvironment”discourse.

Context

Expression

Meaning

LifeCycle Assessment

“directimpact”

Environmentalimpactofthesocalledforegroundsystem, e.g.emissionsofachipfactory

“indirectimpact”

Environmentalimpactofthesocalledbackgroundsystem, e.g.emissionsofthepowerplantgeneratingelectricityfor thechipfactory

“directflow”

Amaterialflowthatentersthesystemofthedomestic economy

“indirectflow”

Amaterialflowthatiscausedabroadbythefactthata goodisimported,whereasthematerialitselfdoesnot enterthedomesticeconomy

“directenergyconsumption”

Theamountofenergyconsumedinadeviceduringtheuse phase(finalenergyconsumptionduringuse).

“indirectenergyconsumption”

Theamountofenergyconsumedtoprovidethedevice (“greyenergy”,“embodiedenergy”)and/ortoprovidethe finalenergyaswell.

“directeffectsofICT”

FirstorderimpactsofICTasdefinedinSection3.1

“indirecteffectsofICT”or “enablingeffectsofICT”

Secondand/orthirdorderimpactsofICTasdefinedin section3.1

Material Flow Analysis e.g.(OECD 2008) Energy consuming products (incl.ICT hardware)

ICTeffects (OECD 2009)


page

13


Defining the baseline or “Business as Usual (BAU)” scenario InordertoassessICTrelatedenergyefficiency,isnecessarytocomparetwosystems,whichwecalledAandB inSection1.3.2.Veryoften,AandBarejustdifferenthypotheticaldevelopmentsofonesystem,i.e.AandB arescenarios.Aisthescenario“withICTapplicationforenergyefficiency”andBcanbecalledthe“baseline scenario”or“Businessasusual(BAU)scenario”. Veryoften,relativeefficienciesarepostulatedwithoutdefiningtheBAUscenario.Forexample,NathandHaas pointedoutthattheSmart2020study(GeSI2008b)isnotveryexplicitabouttheBAUscenario.Accordingto Smart2020, ICT could enable approximately 7.8 Gt CO2equivalents of global emissions savings in 2020. This wouldamountto15%ofemissionsin2020basedonaBAUestimation,economicallyitwouldtranslateinto approximately€600billionofcostsavings.However,thestudydoesnotdefinetheBAUscenario(exceptfora fewassumptionssuchasGDPgrowth).Itwouldbeessentialtoknowwhatothercomponentsofefficiency(not relatedtoICT)areincludedintheBAUscenarioandhowtheyarequantified;inthisregardthestudyrefersto theFourthAssessmentReportofIPCC,whichhoweverexplicitlynegatestodefineaBAUscenario(Nathand Haas2008).

2.3

Interest of the ICT Sector in the Field

As an indicator of the interest of the ICT sector in ICTrelated energy issues, the considerable number of initiativesandstudiesfromtheprivatesectorisdocumentedinthissection.Amorecomprehensiveoverviewis providedbytheOECD’sWorkingPartyontheInformationEconomy(WPIE)intheirlatestreport(OECD2009).

2.3.1

Private-Sector Initiatives

The 80 plus Program 80PLUSisaplatform“thatuniteselectricutilities,thecomputerindustryandconsumersinanefforttobring energyefficienttechnologysolutionstothemarketplace”.7 Themaingoalistointegratemoreenergyefficient powersuppliesintodesktopcomputersandservers.80PLUScertifiespowersupplyproductsforhighefficiency performanceinserverapplications. The Electronic Industry Code of Conduct (EICC) EICCisacodeofbestpracticeadoptedbynearly30majorelectronicsbrandsandtheirsuppliers. 8 Thegoalisto improveconditionsintheelectronicssupplychain.CooperatingwithGeSI,EICCiscreatingandimplementinga set of tools and methods with the aim of ensuring the standards in the Code are upheld throughout the electronicssupplychain. The “Green Grid” The“GreenGrid”isaglobalconsortiumofover100members–ITcompaniesandprofessionals–dedicatedto advancing energy efficiency in data centres and “business computing ecosystems”. 9 A “Framework for Data CenterEnergyProductivity”isintendedtomakepossiblecomparisonsofdatacentres.Itisdescribedinoneof threewhitepaperspublishedwithsupportoftheUSDepartmentofEnergy(GreenGridConsortium2008).

7

Seehttp://www.80plus.org/.

8

Seehttp://www.eicc.info/downloads/EICC_FAQs.pdf.

9

Seehttp://www.thegreengrid.org.


page

14


IBM’s Big Green Project According to IBM, this will be a $ 1 billion project with the aim of increasing the energy efficiency of IBM products: “New IBM products and services, announced as part of Project Big Green, include a five step approach to energy efficiency in the data centre that, if followed, will sharply reduce data centre energy consumptionandtransformclients'technologyinfrastructureinto“green”datacentres,withenergysavingsof approximately42percentforanaveragedatacentre.”Seebackgrounddocument(Ebbers,Galeaetal.2008). The Climate Savers Computing Initiative Climate Savers Computing is a nonprofit group of ecoconscious consumers, businesses and conservation organizationsstartedbyGoogleandIntelin2007,whichcollaborateswiththeU.S.EnvironmentalProtection Agency'sEnergyStarprogram. 10 TheinitiativewasstartedinthespiritofWWF’sClimateSaversprogramwhich mobilized over a dozen companies in the early 2000’s to cut carbon dioxide emissions, with the aim of demonstrating that reducing emissions can be good business. The goal was to promote development, deploymentandadoptionofsmarttechnologiesthatcouldbothimprovetheefficiencyofacomputer’spower deliveryandreducetheenergyconsumedwhenthecomputerisinaninactivestate. AsparticipantsintheClimateSaversComputingInitiative,computerandcomponentmanufacturerscommitted to producing products that met specified powerefficiency targets, and corporate participants committed to purchasing powerefficient computing products. The mission was to reduce global CO2 emissions from the operationofcomputersby54milliontonsperyearby2010.

2.3.2

Industry-Supported Studies

Saving the climate @ the Speed of Light This study was produced in a joint project of WWF and ETNO, the European Telecomunications Network Operator Association (Pamlin and Szomolányi 2006). Started in 2004, the project provided a road map for reducingCO2emissionsintheEUandbeyond.Thisroadmapaimedtoclosethegapbetweenacademicstudies ontheenvironmentalimpactsofICTandpolicymaking.Thefullroadmapisavailableforfreedownload. 11 WWFhasalsobeenactiveinthedevelopingworldonICTandsustainability.WWFIndiaandWiproLimited,a company providing global IT and R&D services, have had an initiative to explore the use of IT to drive sustainable development. WWFIndia and Wipro signed in 2008 a partnership agreement for sustainable development. 12 Smart 2020 This is a pivotal industrysupported study done by the McKinsey consultancy and organized by The Climate GroupandtheGlobaleSustainabilityInitiative(GeSI). 13 Thestudy’smainconclusionisthattheconsequentuse ofICTtowardsenergyefficiencythroughouttheentireeconomycoulddeliverbytheyear2020CO2savingsfive timeslargerthanICT’sownimpact(GeSI2008b).Smart2020identifiedthegreatestsavingsasfollows:2.03Gt CO2e(CO2equivalents)savedbysmartgrids,1.68GtCO2ebysmartbuildings,1.52GtCO2ebysmartlogistics and0.97GtCO2ebysmartmotorsystems.

10

Seehttp://www.climatesaverscomputing.org.

11

Seehttp://www.panda.org/news_facts/publications/ict/index.cfm.

12

Seehttp://www.wwfindia.org/news_facts/index.cfm?uNewsID=2661.

13

Seehttp://www.smart2020.org/.


page

15


Asforthe“smartlogistics”,thestudypredictsthatnewsoftwarecouldpromoteanintermodalshifttowards more efficient means of transport (mainly trains), ecodriving, route optimisation and inventory reduction. Furthermorethereportnotesthat“althoughthisfigureisrelativelymodestcomparedtoreductionsofferedby otherICTdrivensolutionsinthisreport,theopportunitiestomakethelogisticsindustrymoreefficienthave importanteconomicconsiderations,sinceitoperatessuchahighvaluemarket”(GeSI2008b).

2.4

Political Relevance of the Field

This section provides a survey of political initiatives that have created the political awareness for “ICT and energy”issuesattheEU,OECD,andUNlevel.ThefocusisontheEU,howeverwithoutinvestigatingthelevelof memberstates.ForasurveyonnationalinitiativesinOECDmembers(coveringmostofEUmemberstatesas well)wereferthereadertothelatestreportbytheOECD’sWorkingPartyontheInformationEconomy(OECD 2009).

2.4.1

EU Initiatives: Emergence of the Issue “ICT for Energy Efficiency”

NowadaysintheECthemainpoliticaldriveristhenew202020commitmentfortheyear2020describedin greater detail below under the section heading “Toplevel commitment to the future”. Although this developmentbeganintheECinthemidNineties,itreceivedfurtherimpetusin2006throughthegrantingof theNobelPrizetoAlGoreandtheIntergovernmentalPanelonClimateChange(IPCC),andthepublicationof thereportbytheformerWorldBankChiefEconomistSirNicholasStern(Stern2006). AsregardsICTforenergyefficiency,anewtrendistobefoundintheimprovedcollaborationbetweentheDGs for Information Society and Media (INFSO) and for Energy and Transport (TREN). In the words of the Commissionitself,“bringingtogethersectorsasdiverseanddistinctasICTsandenergyisratherchallenging,as theapproaches,andeventimelinesforinvestment,arequitecontrasting(shorttermforICTsversusverylong termforenergy)”(EuropeanCommission2008c). DGTRENaddstothemixitsdualfocusesonmobility(asonemightexpectunder“transport”)andIntelligent GridsandDemandSideManagementDSM(asonemightexpectunder“energy”).Sometimes,however,inthe ECmobilityistreatedasseparatefromEnergyEfficiency,perhapsforexpediencyintheorganizationofcalls. ThissectionlooksnextathistoricalbackgroundmaterialbehindthiscollaborationbetweenDGs. Formal Arrangements One should bear in mind the different types of activities pursued by the European Commission in its work programme. The concept of “initiatives” as it is used here includes Commission terms such as “thematic strategies”and“actionplans”(see(Pallemaerts,Wilkinsonetal.2006)forexplanations),whichareassociated witha“whitepaper”communicationprocess.Theseintermediatestepshelprelatethe“researchprojects”to policymaking. The normal sequence of these Commission activities thus proceeds from the research programme (see Figure 5) by means of an Impact Assessment process (Ruddy and Hilty 2007) and “white paper”communicationprocessinmanycasestobecomeaformoflaw.


page

16


Figure5.RoleofEuropeanCommissioninrelatingresearchtopolicymaking.

TheconceptualizationofresearchprioritiesisinthehandsofvariousCommission“services”,i.e.Directorates General(DGs);butmuchawardingandfundingofresearchisorganizedbyadedicatedDG–theDGResearch. In this way, the socalled framework programmes (FP) developed: “Various Community research activities, mainly in the energy sector, were combined in 1984 into a fiveyear framework programme (FP)”(Banchoff 2002).BanchoffdeploresthelackofCommissionpowertoharmonizetheresearchfundingbyMemberStates: “theequationofresearchpolicywiththeframeworkprogramme–acentrallyadministereddistributivepolicy –impededeffortstodefineitmorebroadlyinregulatorytermsasthecoordinationandintegrationofnational [research]policies”(Banchoff2002). Historical Background on EU Research on Energy Consumption and Energy Efficiency ThehistoricalstartingpointforenergyefficiencyinEuropewasthefoundingoftheEuropeanCoalandSteel Community (ECSC) in 1952. The ECSC predated the establishment of the European Economic Community propersixyearslaterin1958throughtheTreatyofRome.“ResearchpolicyisasoldastheEUitself”claims (Banchoff2002),p.7,lookingbackto1958,recountingitsoriginalentanglementwithnuclearpower,asecond importantaspectofenergypolicyintheearlydaysrightaftercoalandsteel. From a common interest in coal, steel and nuclear power, European integration proceeded with a view to securingpeaceforfuturegenerations,butemployedlargelyeconomicincentivestomotivatememberstatesto surrender sovereignty in carefully conceded increments. Rebuilding the Continent in the 1950’s after World WarIIentailedseekingeconomiesofscaleandefficiency.OneexampleofthiscampaigncanbecitedfromEU Member State Germany where there had been a government body to promote efficiency including energy efficiency, Rationalisierungskuratorium der Deutschen Wirtschaft e.V. since 1921, which after the war re establisheditsofficesinFrankfurt/Main. Energyefficiencyinitscoremaybesaidtofocusonthetechnicalaspectsofprocessmanagement.Inaddition, though,thereareorganizationalaspectssuchasinfrastructurerenewalandurbanplanningincludingbuilding insulation that traditionally characterize the field. In the 1970’s the oil crisis sparked interest in energy efficiency,andEUMemberStatesgrantedsubsidiesandtaxbreakstopromoteit.Unlikethem,theEUhadno power to levy taxes, but identified other areas where it could expand its competencies by comparing and analysingMemberStatessuccessesthroughresearch,andissuedirectives. Inthe1980and1990’s,traditionalsignalprocessingwasgraduallysupplantedbythedigitaltechnology.This newtechnologywavesweptthroughtheeconomybasedonanincreaseintheusageofdigitalsignals,causing


page

17


ittobedubbedtheDigitalRevolution.Asthistrendspread,thepotentialbecamemoreevidentandbeggedto beutilizedalongwiththeolder,largerbodyofenergyefficiencymeasuresdescribedabove. Commission Measures on ICT for Energy Efficiency (ICT for EE) In2006,anactionplanforenergyefficiencywasissued(ingeneralterms,EC2006),whichincludedamobility chapter. Given this relatively early start on that subtopic, the Commission did an Impact Assessment on Intelligent Transport Systems (ITS) in 2008 (European Commission 2008b). and one on ICT for EE (European Commission 2009b). As documented, for the Intelligent Transport System (ITS) Action Plan, “an interservice group composed of representatives of the DirectoratesGeneral concerned (SG, ECFIN, ENTR, EMPL, ENV, INFSO, RTD, TAXUD and JRC) was created to accompany the impact assessment. The group met four times betweenJanuaryandMay2008andprovidedinputtotheimpactassessment.Inadditiontothisinterservice group,anITSSteeringGroupwassetupinApril2007withDirectorsfromfivedifferentDirectoratesGeneral: INFSO, RTD, ENTR, ENV and TREN. This Group provided guidance on the preparation of the ITS Action Plan” 14 (European Commission 2008b) . DG TREN has a Greening transport package containing its Intelligent Car Initiativefrom2006anditsITSprogrammefrom2008. In May 2008, a first official communication specifically dedicated to ICT for EE was issued stating that “It is initially proposed to focus on the power grid, energysmart homes and buildings and smart lighting (due to their relative importance and potential for improvement). Other sectors with considerable energysaving potentialarethemanufacturingindustryandtransport(estimated,by2020,ataround25%and26%oftheir totalprimaryenergyconsumption).[...]TheneedtoimprovethepowergridiswelldocumentedintheAction PlanforEnergyEfficiency”(EuropeanCommission2008b),p.6. InNovember2008,the“SecondStrategicEnergyReview–SecuringourEnergyFuture”(EuropeanCommission 2008h) was published by DG TREN as a “wideranging energy package”. That package contained another package, “a new 2008 Energy Efficiency Package” (page 11) with an emphasis relevant here on reinforcing energyefficiencylegislationonbuildingsandenergyusingproducts.Impactassessmentshavebeendoneboth ontheEnergyPerformanceofBuildings(EuropeanCommission2008g)andtheEnergyEfficiencyofProducts (European Commission 2008a). These products were originally covered by the Energy Labelling Directive for Household Appliances (ELD) now to be recast as one of the elements of the Action Plan on Sustainable ConsumptionandProductionundertheleadofDGENVandonSustainableIndustrialPolicyundertheleadof DGENTRrecentlyreconciledas(SCP/SIP)(p.3). In January 2009, the second official Communication on ICT for EE was issued according to the Roadmap, (EuropeanCommission2009a)Section12,pp.41.ff. InitiativesmoredirectlyrelatedtoICTincludethe“EUStandbyInitiative”(EuropeanActionstoImprovethe EnergyEfficiencyofElectricalEquipmentwhileeitherOFForinStandbymode).Sincethenadraft“Codeof Conduct on Energy Efficiency of Data Centres has been issued (version 0.8 of 8.04.2008) prepared by the DirectorateGeneral JRC, Joint Research Centre, Institute for Environment and Sustainability, Renewable EnergiesUnit(JRC2008). 15 The creation of “technology platforms”, such as the SmartGrids European Technology Platform (ETF), comprises a step towards organizing in the newer form “initiative”, such as the “European electricity grid initiative”. Such initiatives form parts of the Strategic Energy Technology Plan (SET Plan) (for FAQs see (IHS 2008)).TheSETPlanisanattempttoimprovecollaborationbetweentheEUandmemberstatesdatingfrom November2007.

14

Seehttp://ec.europa.eu/transport/strategies/2008_greening_transport_en.htm.

15

Seehttp://www.smartgrids.eu/.


page

18


2.4.2

OECD

TheOECDheldafirstworkshopon“ICTsandEnvironmentalChallenges”inCopenhageninMay2008.Basedon thisWorkshopandtheOECDMinisterialMeetingontheFutureoftheInternetEconomyinSeoulinJune2008, a highlevel conference on “ICTs, the environment and climate change” was held by OECD and the Danish MinistryofScience,TechnologyandInnovationinHelsingoron2728May2009. 16 Outcomesofthisworkshop willcontributetoworktowardstheOECDCouncilatMinisterialLevelinJune2009andwillberelevantinthe context of theUnited Nations Climate Change Conference,718 December 2009 in Copenhagen, Denmark (COP15). Thereport“TowardsGreenICTStrategies–AssessingPoliciesandProgrammesonICTandtheEnvironment”, prepared as an input to the Helsingor Workshop, provides an analysis of 92 government programmes and businessinitiativesacross22OECDcountriesplustheEuropeanCommission(OECD2009).

2.4.3

UN Initiatives

TheInternationalTelecommunicationUnion(ITU),theoldestUNagencyandorganizeroftheWorldSummiton theInformationSociety(WSIS,20032005),hassetupaFocusGroup(FG)onICTsandClimateChange. 17 ITUhasasconstituentmembersmostoftheworld’sleadingtelecommunicationcompanies.Theyareinvolved inagreeingonfuturedefactostandardsformeasuringtheeffectsoftheuseofICTontheenvironmentand secondorderimpactsaswell.ThecurrentstatusoftheFG’sworkisthatinApril2009resultswillbepresented to the ITU’s influential Telecommunication Standardization Advisory Group (TASG). One further proposal pendingapprovalistomakeITUtheworld’spremierclearinghouseforICTandclimatechange. AfterthefirstyearofthisFocusGroup’swork,theITUhaspassedaresolutionwiththeaimsof x

“CreationofaframeworkforenergyefficiencyintheICTfield,takingaccountofWTSAResolution73” (ITUT2008),and

x

Creationofacentral“repositoryandknowledgebaseontherelationshipsbetweenICTsandclimate change”; with the intention of receiving a “report on progress on the application of this resolution annually to the ITU Council and to the 2012 world telecommunication standardization assembly” (ITUT2008),p.3.

16

Seehttp://www.oecd.org/sti/ict/greenict.

17

Seehttp://www.itu.int/ITUT/climatechange/.


page

19


3 State of the Ar t

3.1

Conceptual Framework

The conceptual framework used to structure the issue is based on a product lifecycle perspective, which is applied to two types of life cycles: (i) the life cycle of a given ICT product or service (such as a PC, a mobile phone,anoperatingsystem,atelecommunicationservice)and(ii)otherlifecyclesthataremodifiedthrough the availability of the information or communication service (IC service) provided by the first life cycle. We introducetheterm“linkedlifecycles”torefertothisconceptfurtheron. Figure6showsthegenericlifecyclemodelthatwillbeusedasabasicbuildingblockforthelinkedlifecycle framework.However,anyotherlifecyclemodelcouldalsobeused,provideditviewsthelifecycleasasystem thathasthepurposeofprovidingaserviceatthecostofmassandenergythatflowthroughit.Ontheirway trough the system, mass and energy are transformed. Roughly speaking, the production phase transforms inputresourcestoaproductandtheusephasetransformstheproducttowastewhileprovidingthedesired service. Both the physical inputs and outputs of the system, i.e. all flows crossing the system boundary, are evaluatedwithregardtotheirenvironmentalimpacts. Energyconsumedisrelatedtothefunctionalunitchosen,i.e.dividedbythenumberoffunctionalunitsthat are generated during the use phase. This implies that the service life – the length of the use phase – of a durableproductisanessentialparameterinanLCA.Example:IfthesystemunderstudyisthePClifecycleand thefunctionalunitisdefinedas“1yearofPCuse”,theenergyconsumptionofthenonusephases(production andendoflife)iscuttohalfifthePCisusedfor6insteadof3years.

Figure6.Genericmodeloftheecologicalproductlifecycle.

TheeffectofICTasanenablingtechnology,namelyasanenablerofchangeinproductionandconsumption withthepotentialtomitigateenvironmentalimpacts,canbeconceptualizedwiththeapproachoflinkedlife cycles,i.e.byshowinghowtheICTlifecyclecanaffectotherlifecycles.Figure7illustratestheconcept.The environmental impacts of the ICT life cycle (shown at the bottom) are also called “firstorder impacts” or “primaryimpacts”ofICT.Themodificationoftheimpactsoccurringatthelevelofthesecondlifecycle(shown atthetop)arecalled“secondordereffects”or“secondaryeffects”ofICT,becausetheyareindirectlycaused throughthemodificationofthesecondlifecyclebytheavailabilityoftheICservice.Theterminologytraces backtoEMPA(2005)andthesourcescitedthere. ThepotentialmodificationsofanICservicetoaproductlifecyclecanbedescribedgeneralizingatypologyof relationships between telecommunications and transportation, which is based on the idea to differentiate among optimization, substitution and induction effects: An IC service can help to better organize traffic (optimization),replacetraffic(substitution)orgenerateadditionaldemandfortraffic(induction).Thistypology isnowgeneralizedtoallICTappliedinanyfieldofapplicationandinterpretedformalifecycleperspective.


page

20


Optimization can refer to any phase of the life cycle in our framework, including the design phase, which usually is not regarded part of the life cycle in LCA methodology but included here as an important link betweenthetwolifecycles(seeFigure7).TheredarrowsrefertoeffectsoftheICservicewhichmodifythe efficiencyofpartsoftheotherlifecycle,i.e.theyrepresentoptimizationpotentials. Substitution and induction refer to the demand for the service which is affected by the availability of the IC service. If the original service is replaced by the IC service, the product becomes obsolete (at least in its functiontoprovidetheservice),whichisasubstitutioneffect.However,thedemandcouldalsoincreaseasa consequenceoftheICservice(suchasthedemandforpaperincreasesduetotheavailabilityofinkjetprinters), whichisaninductioneffect.

Figure7.Illustrationofthe“linkedlifecycles”concept.TheinformationorcommunicationserviceprovidedbytheICTlife cycle at the bottom can modify the life cycle of another product (providing any service) in two ways: by modifying the design, production, use or endoflife phase of that product (red arrows) or by influencing the demand for the service it provides(yellowarrow).Source:Empa

Table2showshowthistypologycanbeinterpretedinmoredetailinthelinkedlifecycleframework.Thistable canalsoserveasachecklisttoscreenagivenICserviceforpotentialsecondorderimpacts. AninterestingspecialcaseoccursifbothlifecyclesareICTlifecycles.Inthiscase,ICThassecondordereffects in the ICT sector. Example: The IC service is PC system software, the other life cycle PC hardware. If the software helps the hardware to come closer to the ideal of loadproportional power demand, it has an optimization effect on the use phase. If a new software version demands for more hardware capacity, it increasesthedemandforPChardwarebyshorteningtheusephase.ThelattereffecthasbeencalledSoftware InducedHardwareObsolescenceorSIHO(Hilty2008).


page

21


Table2.ExplanationoftheLinkedLifeCycleApproach:Howaninformation/communicationservicecaninfluencethelife cycleofanotherproduct(Source:Empa)

Contactpoint

EffectofICservice

Designphase (optimization)

Samedesignwithlesseffort Positive,smallpotential(no multiplication)

ConventionalCAD

Newdesignenablingmore efficientproduction

CIM

Production phase (optimization)

Secondorder environmentalimpact

Positive,highpotential

Examples

3Dprinting

Complexityreduction, designingmultiuseparts

Newdesignenablingmore efficientuse

Energyefficient architecture

Newdesignenablinglonger use

Designformaintainability

Newdesignenablingmore efficientrecycling

Designforrecyclability

Moreefficientproduction process

Positive,highpotential (unlessalreadyused)

Processoptimization Integratedprocesschain Optimizedlogictics

Usephase (optimization)

Moreefficientuse


Energymanagementofuser appliances Intelligentheating,cooling andventilation Smartgrids

Longeruse


Remotemaintenanceand selfmaintenance Informationsystemson replacementparts

Endoflifephase Moreefficientendoflife treatment


Smartsortingtechniques forrecycling

Demandforthe service

ReplacedbyICservice (substitution)


Virtualmeetingsreplace travel

Reduceddemand (increasedawareness)


Reducedenergy consumptionduetosmart metering

Increaseddemand (induction)

Negative,highpotential

Increasedpapercon sumptionduetoPCprinters Hardwareobsolescencedue tonew,demanding softwareversions


page

22


3.2

EU-Funded Projects and Studies Contributing to the Field

OvertheearlydecadesoftheEU’sexistence,theUnion’sinfluenceoverthefieldofenergygraduallyincluded moreICTs,asthetechnologiesevolvedandconvergedinnewcontexts.Thiswasorganizedmainlyunderthe DirectorateGeneral for Transport and Energy (DG TREN). In the mid1990s, a new DirectorateGeneral was createdfortheemergingInformationSociety(DGINFSO).Henceforth,itsharedwithDGTRENthecompetence forenergyandICT. TheprojectslistedimmediatelybelowinapproximatechronologicalorderwerecarriedoutunderDGINFSO, whereasDGTREN’sinfluencecanbeseenmoredistinctlyinitsleadroleinthesuperprogrammeIntelligent Energy Europe (IEE) and in projects described at the end of this section under headings such as Buildings, meteringandthegridandIntelligentTransportSystems.Therelativeimportanceofthesectorsisdealtwith under the “Studies” heading, leading into the next section with its discussion of how to measure the direct energyconsumptionofICT.

3.2.1

EU-Funded Projects

Early Discourse-Oriented Projects GlobalSocietyDialogue:InSeptember2001,DGINFSOorganisedaworkshopentitled“TheChallengeofthe DigitalDivide”.TheworkshopwasheldinViennabytheGlobalSocietyDialogue(GSD)projectsupportedbythe EuropeanCommissionandcarriedoutbytheResearchInstituteforAppliedKnowledgeProcessing(FAW)based inUlm,Germany.Theworkshopwasdocumentedinabrochure(SchauerandRadermacher2001). ASIS: Under the 5th Framework Programme, a series of projects involving Information Society Technologies (IST), entitled “Living and Working in the Information Society,” followed up on previous projects organized under Advanced Communications Technologies and Services (ACTS) and Esprit. In 1998, ACTS participants joinedanewAllianceforaSustainableInformationSociety(ASIS)assembledbyKlausTochtermannatFAWin 18 Germany. TheGlobalSocietyDialogueprojectwascarriedoutin2001byDGINFSOunderRobertPesteland FAW’s Thomas Schauer. This was followedup with the Terra 2000 project on the optimisation of ISTs’ contribution to sustainability, 19 also with DG INFSO, Robert Pestel, the RAND Corporation and FAW Thomas Schauer,andBarryHughesofInternationalFutures(IFs). DigitalEurope:FollowinggoodexperienceintheU.K.,theForumfortheFutureconductedaprojectknownas DigitalEurope,whichproducedasseriesofstudiesonteleworkanddematerialization,suchas“DigitalEurope –VirtualDematerialisation”,September2003,aswellasabookeditedbyJamesWilsdon(Wilsdon2001). EPICICT: Under the 6th Framework Programme a project called EPICICT was conducted to measure ICT products’environmentalimpacts. IEE Projects TheEfficientServersproject 20 waspartofthesuperprogrammeIntelligentEnergyEurope(IEE)describedin section5below.Theprojectaimedtodemonstratethelargepotentialsavingsthatcouldbeachievedthrough efficienttechnologyintheareaofserversandtodrivemarketdevelopmenttowardsenergyefficientservers. TheprojectconsortiumincludedtheAustrianEnergyAgency(asprojectcoordinator),IBM,Sun,Universityof Karlsruhe, the French energy agency ADEME, and Robert Harrison Associates LTD. Compilation of IDG data

18

Seehttp://cordis.europa.eu/infowin/acts/ienm/newsclips/arch1998/981001es.htm.

19

Seehttp://www.Terra2000.org.

20

Seehttp://www.efficientservers.eu.


page

23


fromboththeUSAandtheEUinvariousscenariosshowedthat2070%savingscouldbeachieveddepending ontheapplication.TheEfficientServersprojectlastedfortwoyearsculminatinginthespringof2009. TheSAVEprogrammewasafurtherprogrammeinsideIntelligentEnergyEurope.SAVEincludedmanyprojects suchasthosechosenfortheirICTcontentandlistedbelow.SAVEdidnotincludepilotactionsforlargesmart buildings, which are covered under CIPPSP 2008 Work Programme Theme 2: “ICT for Energy Efficiency and Sustainabilityinurbanareas”(EuropeanCommission2008d). OdysseewasanotherwellknownprojectwithinIEE,whichproducedtwodatabases:oneonenergyefficiency data&indicators,andanotheronpolicymeasures.Thedatabasesarestillcontinuallyupdatedandaccessis freetoregisteredusers.OdysseewasincludedinalistofinnovativeprojectsbyECEEEEuropeanCouncilforan EnergyEfficientEconomy. 21 ARTEMIS Projects The ARTEMIS Joint Undertaking (JU) was set up as an early major initiative to bring privatesector research actors together with the European Commission and a large number of contributing Member States. An association was formed in 2007 for R&D actors (such as DaimlerChrysler, Nokia, Philips Electronics, STMicroelectronics,andThales)inAdvancedResearch&TechnologyforEmbeddedIntelligenceandSystems 22 tosupporttheARTEMISJUprogramme.Itissuesitsowncalls. x

One majorprojects done in the context ofthe ArtemisiaAssociation was CESAR; it was intended to reducedevelopmenttime,andhadatotalcostofEuros58.5million.

x

AnothermajorprojectisSOFIAonimprovingtheinteroperabilityamongmultivendordevices,which startedin2009forthreeyearswithabudgetofEuros36.5million.

x

SCALOPES also started in 2009, for two years with Euros 36 million to be used for an “industrially sustainablepathfortheevolutionoflowpowermulticorecomputingplatforms.”

x

eDIANAisanothersuchprojectrunningfromFeb.1/2009toJan.31/2012calledEmbeddedSystems forEnergyEfficientBuildings(eDIANA)addressestheneedofachievingenergyefficiencyinbuildings throughinnovativesolutionsbasedonembeddedsystems.ItisfundedwithEuros17.3millionfor3 yearsfrom1.1.09.

Electricity Supply and Demand Projects CLEVERFARM was an FP5 project for the advanced management and surveillance of wind farms. 23 It began makingwindfarmsmoreintelligentbyprovidingremotemeasuringofthestatusofthewindfarm,andsending warningstothemaintenancecrewifsomethinggoeswrong.Itwillalsopredictitsownoutputandschedule maintenance. A similar project called ANEMOS dealt with Development of a Next Generation Wind Resource Forecasting SystemfortheLargeScaleIntegrationofOnshoreandOffshoreWindFarms. 24 ThePOWERSAVERprojectconductedin2006aimedtoimprovehouseholdapplianceswithintelligence. The Intelligent Metering project looked at energy and water savings which could be obtained in local and regionalpublicsectorbuildings. 25

21

See http://www.odysseeindicators.org/, http://ec.europa.eu/intelligentenergy, and http://www.eceee.org/european_directives/, respectively.

22

Seehttp://artemisiaassociation.org/.

23

Seehttp://www.cleverfarm.com/.

24

Seehttp://anemos.cma.fr/modules.php?name=News&file=article&sid=1.


page

24


TheSEPOWERFOILprojectwasdevotedtothedevelopmentofrolltorollmanufacturingtechnologyforthe productionofcosteffective,highefficiency,flexiblephotovoltaicmodules. Intelligent Transport Systems (ITS) Projects ITS projects seek to apply ICTs to transport. The main innovation is said to be the integration of existing technologies to create new services. ITSs can be applied in every transport mode (road, rail, air, water) and servicescanbeusedbybothpassenger(Intelligentmobility)andfreighttransport. TheFREIGHTWISEprojectdevelopedtheoreticalmodelsandIntermodalManagementServices,whicharenow beingimplementedaspartsoftheFREIGHTWISEFramework(FWF). INTElligent inteGration of RAILway systems (INTEGRAIL), 26 was an Integrated Project that ran for two years endinginDecember2008.Ithadabudgetof20.66millionEuros,halfofwhichcamefromtheCommission.It was intended to test an allpervasive, wideband communication infrastructure providing digital data links between trains and ground installations, with the hope that it could to be used as the future reference standardinrailways. “Intelligentroads”(INTRO),asynergisticclusteringactionledbyFEHRL(ForumofEuropeanNationalHighway ResearchLaboratories), 27 strovetocombinenewandexistingsensortechnologiesinpavementsandbridgesin order to prevent accidents, enhance traffic flows and significantly extend the lifetimes of existing infrastructure.Itranfrom2005until2008andhadabudgetof3.5millioneurosofwhich2millioneuroswere fromtheCommission. Modularurbanguidedrailsystems(MODURBAN),28 wasacoordinationactiontodesign,developandtestan open common core system architecture and its key interfaces to pave the way for the next generations of urbanguidedpublictransportsystems.Itranuntil2008andhadabudgetof19.42millioneuros,halfofwhich camefromtheCommission.

3.2.2

EU-Funded Studies

The Future Impact of ICTs on Environmental Sustainability TheInstituteforProspectiveTechnologicalStudies(IPTS)attheCommission’sJointResearchCentreinSevilla, Spain, commissioned an early study – in fact, “the first quantitative projection” of the impact of ICTs on environmental sustainability (Erdmann, Hilty et al. 2004; Hilty, Arnfalk et al. 2006). Using a methodology combining qualitative scenariobuilding and quantitative modelling, the general conclusion reached was that ICTscouldeitherimprovethesituation,reinforcingpositiveeffectsontheenvironment,ortheycouldworsen thesituation.Thisconclusionsuggestedthatenvironmentalpolicieshavetobedesignedinsuchaswayasto ensurethatICTapplicationsmakeabeneficialcontributiontoenvironmentaloutcomes,and,atthesametime, suppress rebound effects. The sectors “housing” and “passenger and freight transport” were identified as crucialwithregardtoICTapplications. Impacts of information and communication technologies on energy efficiency Looking at possible future developments in the area of ICTs and energy efficiency as well, this study (Bio IntelligenceService2008)constructed“businessasusual”(BAU),andEcoscenarios.IntheEcoscenario,ICTs 25

Seehttp://www.managenergy.net/products/R1951.htm.

26

Seehttp://www.integrail.info/structure.htm.

27

Seehttp://intro.fehrl.org/.

28

Seehttp://www.modurban.org.


page

25


wereshowntoprovidethelargestenergyefficiencygainsintheareaofbuildingHeating,VentilationandAir Conditioning (HVAC). The total possible gains (for the Eco scenario) would outweigh seven times ICT’s own CO2efootprint.

3.3 3.3.1

Measuring ICT-Related Energy Consumption Approaches Based on Final Energy Consumption

AlargenumberofstudiesfocusonthefinalenergyconsumptionofICTintheusephase(IBM2006;EPA2007; Koomey 2007; BioIntelligenceService 2008; DEFRA 2008; Fichter, Clausen et al. 2008; Fraunhofer IZMISI 2008). The studies we reviewed calculate the energy consumption by ICT enduser devices (such as PCs, telephones)andinsomecasesalsobyICTinfrastructure(suchasdatacenters,servers,routers).Thisisdone foragivengeographicareawithabottomupapproachbasedonthestockofICTdevicesandinfrastructure andontheirassumedaverageenergyconsumption. ThestudiesevaluatedhavesimilarbutnotthesamescopewithregardtotheICTconsidered.Thus,insome cases data centres and telecom networks are included and in some others not. Further, in the studies we evaluated,theauthorsconsiderdifferentperipherals(seeTable3andTable5).Ingeneral,theICTevaluatedis ratherpoorlydocumented,particularlyregardingtherepresentativedevicesselectedforcalculations.Contrary tothat,theprocedurefollowedtoestimatethestockofICTunitsismostlywellexplained.Usually,statistics fromindustryonshipmentsandstockturnoverareapplied.However,alsoassumptionsonthelifespanofthe devices and infrastructure have to be made, which introduces uncertainties in calculations. The procedure followedtodeterminetheaverageenergyconsumptionofthedifferentICTitemsconsideredisdescribedwith different levels of detail. Here also assumptions on usage patterns are made, which might contribute to uncertainty. With regard to the last point, for example, important differences in energy consumption values can be expectedforICTorevenforacategoryofICT,suchasdatacentres,dependingonwhethertheirrelatedcooling andlightingserviceswereconsideredornot(Cremer,Eichhammeretal.2003;BioIntelligenceService2008; BITKOM2008).Anotherissuethatmakesitdifficulttocompareresultsformdifferentstudiesisthescarcityof standardizedproceduresfordeterminingthetypicaluseofICTdevicessuchascomputerswhenassessingtheir 29 electricityconsumption(JönbrinkandZackrisson2007).InthesocalledEuPpreparatorystudies ithasbeen carefullydocumentedhowtheenergyconsumptionwascalculatedforawiderangeofICTdevices,usingtest standardswhentheywereavailable.TheseEuPpreparatorystudiesandtheirvaluesonenergyconsumption are quoted in a large number of studies and they will certainly contribute to a harmonization of calculation proceduresinEuropeanstudies.Anadditionalsupportcanbefoundinthespecificationsfortestproceduresto befollowedforcomputerswithintheENERGYSTAR®Program 30 andintheEuropeanStandardEN62018for methodsofmeasurementofelectricalpowerconsumptionbyICTdevicesindifferentusemodes. Thedifferencesinthescopedefinitionandmethodologyfollowedinthestudiesevaluatedmakesitdifficultto compareresults. Table3presentsvaluesoffinalenergyconsumptionorCO2emissionscausedbyICTworldwide.Interestingly, mostofthesestudiesonlypresentCO2emissionvaluesandnottheenergyconsumptionvaluesthatmustbe behind. This does not only obscure the origins of the results, it also adds uncertainty to them because additional(implicit)assumptionsaboutpowergenerationanddistributionareincluded.

29

TheEuP preparatory studies were conducted in preparation for the Directive 2005/32/EC on the ecodesign of Energyusing Products (EuP),suchaselectricalandelectronicdevicesorheatingequipment.ThisDirectiveprovidescoherentEUwiderulesforecodesignand ensurethatdisparitiesamongnationalregulationsdonotbecomeobstaclestointraEUtrade.

30

Seehttp://www.euenergystar.org/en/index.html.


page

26


Table3.EstimatesofICTrelatedenergyconsumptionorCO2emissionsworldwide Reference

Year

EnergyconsumptionorCO2 emissions

ICTtypesincluded

Lifecyclephases considered

Usephase

Allphases considered

2006 (Mingay2007b)with additionalinformation from(Gartner2007)

460MtCO21

600MtCO21

PCs(desktopandlaptop),serversand cooling,fixedandmobiletelecommuni cations,localareanetworks(LAN),office telecommunicationsandprinters.

ForPCsandcell phones:design, manufacturing, distribution,use;for others:use

(Malmodinand Jonsson2008)

2007

n.a.

650MtCO2

Fixedandmobiletelecommunications, otherICTcommercialuse,otherICT householdsuse

Notdocumented

(GeSI2008b)

2007

640MtCO2

830MtCO2

PCsandperipherals(workstations, laptops,desktops,monitors,printers),IT services(datacentresandtheir componentservers,storageand cooling),andtelecommunications networksanddevices(network infrastructurecomponents,mobile phones,chargers,broadbandrouters, IPTVboxes)

Manufacturing, distribution,use, endoflifetreatment

(Koomey2007)

1

2005

123billion kWh

123billion kWh

Servers(highend,midrangeandvolume Use servers)andcoolingandauxiliary equipmentassociatedtoserverpower

Mingayusedanemissionfactorof0.6kgCO2/kWhelectricity

Table 4 presents the results of the Study by Bio Intelligence Service, which calculated the final energy consumptionofICTdevicesintheusephaseonly(BioIntelligenceService2008). Table5presentsserverrelatedconsumptionvaluesintheUSA,alsobasedonfinalenergyconsumptioninthe usephase.EnergyconsumptionrelatedtoserversasreportedbyKoomeyiscomparabletotheconsumptionby colorTVsetsintheUS.,theEPAvaluefordatacentersismuchhigher,correspondingtotheelectricityusedby the entire US transportation manufacturing industry (which includes the manufacture of ships, trucks, automobilesandaircraft).


page

27


Table4:ICTrelatedenergyconsumption(usephaseonly)inEU25(BioIntelligenceService2008). Year

Energyconsumption

ICTtypesincluded

Amount

Assumptions

2007

34TWh/a

Theinstalledbaseofserverswas calculatedbasedon(Fichter2007; Koomey2007)

2005

29.1TWh/a

Serversanddatacentres:smalltolargecentresand infrastructureincludingairconditionssystems,ventilation systems,lightingsystems,switchesanduninterruptiblepower supplies(UPS).

2005

14.3TWh/a

Theauthorsuseddatafromfive leadingEuropeantelecomcompanies (British,German,French,Italianand SpanishTelecom)andassumedthat thesecompaniesrepresent70%of EU25.

CoretelecomnetworksandTV/radiobroadcasting:this categorycontainscopperoropticalfibretelecommunications lineswithrespectiveterminal,routerandswitches.Wealso coverRadio/TVbroadcastequipmentincludingradiorelays, directorialradioantennas,etc.

2005

13TWh/a

SincenodatawasavailableforEurope, theauthorshadtousedatafromtwo surveysforGermanyandextrapolated thevaluestoEU25

Cellularphonenetworks:thiscategorycontainsmobilephone nd rd telecommunicationsofthe2 (GMS/GPRS)and3 generation (UMTS/WCDMA)withbasetransceiverstations(nodeB),main switchcontrolsandothernetworkcomponentsincludingtheir infrastructure(cooling,etc)

2005

56.4TWh/a

(seeabove)

TotalICTinfrastructure,thesumoftheprevious3.

2005

42TWh/a

30%ofthereportedvaluerelatesto officedevicesand70%tohome devices.

Computers:Desktopcomputers,laptops,andCRT/LCD Monitors.

2005

54TWh/a

Television:CRT(cathoderaytube),LCD(liquidcrystaldiode), PDP(plasmadisplaypanel),RP(rearprojection),TVsandTV componentunits.

2005

7.8TWh/a

Imagingequipment:Inkjetandelectrophotographybased copiers,printersandmultifunctionaldevicesinmonochrome andcolour.

2005

0.5TWh/a

Mobiledevices:digitalcameras,camcorder,etc

2005

27.8TWh/a

ExtrapolatedfromdataforGermany.

Audiosystems:compactsystems,stereosystemsandclock radios.

2005

4.5TWh/a

ExtrapolatedfromdataforGermany.

VHS/DVDequipment

2005

9.1TWh/a

SetTopBoxes:personalvideorecorders

2005

4.3TWh/a

Telephones:DECT(cordless)telephonesandsmartphones(the latterincludesphoneswithmanyadditionalfunctions)

2005

1TWh/a

Faxmachines

2005

4.1TWh/a

Modems

2005

2.7TWh/a

Mobilephones

2005

158.1TWh/a

TotalICTendusedevices

2005

214.5TWh/a (8%)

Total electricity consumption in EU25 =2691TWh/a(Eurostat)

TotalICT(endusedevices+infrastructure)


page

28


Table5ConsumptionofenergyandCO2emissionoftheusephaseofICTinUSA Reference

Year

Energyconsumption Amount

(Koomey 2007)

2005

45billionkWh/a

(EPA2007)

2006

61billionkWh/a (1.5%oftotal USenergy consumption)

ICTtypesincluded

Assumptions Highendservers(>500000USDperunit),mid rangeservers(25000–500000USDperunit), volumeservers(