AGreenInternet

1、a green internetinformation logistics and energyrod tuckerarc special research centre forultra-broadband information networks (cubin)my friends:jayant baliga, kerry hinton, rob ayreenergy and the internetthe internetvideoenergyincore networkdistributionnetworkdatadatacentermetro/edgenetworkaccessnet

2、workoutwhy is energy important? opex greenhouse impact energy-limited capacity bottlenecks (“hot spots”) enabling energy efficiencies in other sectorsgtco22putting things into contextinformation and communication technologies(ict)global emissionsgtcoeebusiness asusual 2020global emissions2020 abatem

3、entsotherictemissions5% of total“smart 2020: enabling the low carbon economy in the information age,”gesi, 2008 summaryestimating energy consumption of the internetwhere is the energy consumed- core, metro, access network, data centres?information logistics- cloud computing- travel repla

4、cement- airmail vs. internetthe khazzoom-brookes postulatecaveat: “making predictions is difficult especially about the future.”network energy modeltier 1 networkcore networkmetro/edge networkaccess networkcore routeredfaopticalcross connectfiberedgeroutersbroadbandnetworkgatewaysethernetswitchdslam

5、oltcabinetsplittercufiberdslponhotspotsoltswitchcabinetdslamonufibercufttnptpserverserverstoragestoragevideo distribution networkdata centeroversubscriptionoversubscription 3 mb/s in 2009m =peak access rate sold to useraverage access ratem = 25 0.12 mb/s in 2009%ofelectricitysupplypower(w/user)5bali

6、ga et al., 2008power consumption of ip network25oversubscription = 2520152009 technologytodays internet( 3 mb/s)totalrouters1.010access (pon)sdh/wdm links0.5005010015020025020092020peak access rate (mb/s)traffic growth = 40% p.apower consumption in access networksaccess n/wwave7 ont-g1000iedge nodes

7、plitterfibergponcisco6513hitachi 1220splittercabinetnec vf200f6cufttnvdsl2nec am3160cisco 4503tc communications tc3300fiberptphitachi 1220axxceleraexcelmax btsaxxceleraexcelmaxciscoubr10012rfgatewayfibernodecucisco dcp3000wimaxhfcrf amppowerperuser(w)300mpower consumption in access networks20 users

8、per sector=1m= oversubscription20wimaxfttnm=1m=1ptp10m= 1m= 1032 customershfcponm=1m 1001101001000peak access rate (mb/s)pon ftth is “greenest”low-power states in user modemseuropean commissiondirectorate -general jrcjoint research centreinstitute for the environment and sustainabilityrenewable ener

9、gies unitcode of conducton energy consumption of broadband equipmentdraft version 3issue 15 17 july 2008“with implementation of this code of conduct, 5.5 millions tons of oilequivalent (toe) will be saved per year.”extract:adsl-cpevdsl2-cpegpon onuptp onuoff-state(w)0.3low-powerstate (w)3.5

10、4.55.03.0on-state(w)4.06.09.05.0energyperbit(j)0.1network energy consumption per bit1000100 j/b1001 j/b10total1.0routersaccess (pon)wdm (optical fibre trunk links)0.012.5252502500peak access rate (mb/s)energyperbit(nj)2009energy per bit in network devices10001001010.10.012009opticalpos ethernet core

11、ponserverampwdm switch router onutx/rx (10 mb/s)cisco crs-1 routerlargest routing system available todaylinecard chassis: 1.28 tb/s, 13.6 kwswitch fabric chassis: 8 kwpower consumptionbuffers, 5%switch fabric, 10%power and heatmanagement, 35%control plane, 11%i/o, 7%ip look-up andforwarding engine,3

12、2%energy in electronic integrated circuitscwirelinecardcmos asiccmosgatesenergy =egate +1 2 cwire v 2power = energy x bit rateswitchingenergyes,jfereswitching energy in cmos10-1425010-16130cmos65atus32ize(nm)10-1840% decrease /18 months(40% efficiency gain p.a.)18?10-20200020102020moores law: number

13、 of transistors x 2 each 18 monthssource: itrs 97-06 roadmapsx1993performance100008xwore0%o y 4mfficieserouter capacity growthrouter capacity x 2.5/18 m12416crs-1(1.3 tb/s,13.6 kw/rack)100010010(0.3 tb/s/rack)12016(80 gb/s/rack)cmom2/1s la p.a.encnergyrouter energy efficiencyimproving at 20% p.a.nei

14、lson, jstqe 20061199419961998200020022004200620082010yearbased on g. epps, cisco, 2006power(w/user)%ofelectricitysupply5effect of efficiency gains252015overall technology efficiencyimprovement rate= 0% p.atotal1.0routers10access (pon)sdh/wdm links0.5005010015020025002020peak access rate (mb/s)baliga

15、 et al., 2008power(w/user)%ofelectricitysupply5effect of efficiency gains252015overall technology efficiencyimprovement rate= 0% p.a5% p.a10% p.a1.010target20% p.a0.500501001502002502020peak access rate (mb/s)baliga et al., jlt, 2009accessnetworkglobalpacket corenetworktowards energy-efficient netwo

16、rksapplicationcontent serverefficient clients andhome networksaccessnetworkservers anddata centresglobalpacket corenetworkapplicationcontent server alternatives for sleep mode adaptive sharing of energy managementcomplex pc adaptive control ofcontentsolutions e.g.optical/electronic sleep mode andcom

17、ponents based highly efficient powerbypassfast “wake up”,adaptive modes ofon communicationrequirementsadaptive processing content deliverynetwork and clients virtualizationnetworks (e.g. iptv)after: gladish et al., ecoc 200812data centersdata center electricity consumption is 1% of the global total1

18、energy consumption of data centers worldwide doubled between 2000 and2006 2incremental us demand for data centre energy between 2008 and 2010 isthe equivalent of 10 new power plants21 mw data centerkoomey, 2008revolutionizing data centre efficiencykey analyses”mckinsey & company, april 2008.3020

19、energy consumption in a data centreconversionlosslightingictinfrastructurecooling5040“burden factor” of 1.8 to 2.5 for powerconsumption associated with cooling,1050%35%15%conversion/distribution and lighting0serverstoragenetworksources: eyp mission critical facilities, cisco it, network world,custom

20、er interviews, apcinformation logisticsmanagement of data, including, transport, storage, and processingsome typical issues/questions: wired vs. wireless broadband access can cloud computing save energy? renewable energy sources physical vs. internet delivery of data can internet travel replacement

21、save the planet? optimum design of video on demand systemsstoragecloud computingenterprise data centerservers,processorspon access networkcomputer,storagecore networkoltsplitterfiberonucorerouteroxcservers,processorsbroadbandnetworkgatewaysedgeroutersethernetswitchmetro/edge networktotal energyconsu

22、mption =energy/bit used in storage transport processingxstoragepublic data centercloudtotal number of bitsthree cloud computing scenariossoftware-as-a-service stored on users computer with updates downloaded regularlyservice bureau most tasks done on lower end user machine, outsource the“big” jobsco

23、mputing-as-a-service hosted and run on provider computer “farm” with data initiallyuploaded from user (thin client model)can cloud computing save energy ?service bureauhere is my rawcomputingresourceservicecentermetro/edgenetworkvideo footageand editcommandsaccessnetworkmid-rangecomputersservice bur

24、eau hardware & softwareservice bureaucomputingresourceservicecentermetro/edgenetworkhere is yourdvd imageaccessnetworkmid-rangecomputersservice bureau hardware & softwareaveragepowerconsumption,w16040energy consumption of service bureau modelcomputer used for 20 hrs/week, plus some video enc

25、oding of hour videos180140120older computertransport energydominates100mid-range computer8060low-end computer & outsourcing200020406080100120number of encodings per weekfollowing the sun and the windmaking best use of renewable energy locate computing and storage resources near sources of renewa

26、ble energy move data to follow the sun and winds requires many orders of magnitude increase in a data transport capacity transport/efficiency trade-offdata by airmail vs. data by the internetdata by airmail:3x10532-gbcargo jetmelbourneflashdrives107 gb103 kg co220 nj/bsydneydata by internet:(24 hour

27、s)107 gbthe internet1000 gb/s for 24 hours2 j/b105 kg co2energyperbit(j)energy consumption per bit1000100 j/b100101 j/btotal2 j/b1.0routersaccess (pon)0.10.01flash chips by airmailwdm links2.5252502500peak access rate (mb/s)using the internet for travel replacementvideo conferencingsource: cisco, 20

28、08travel replacement greenhouse impact (c02)air travelbusiness meeting500 kg/person returnmelbournevideo conferencing2 x 1 gb/s for 6 hours= 4 tb5 kg/personsydneydistancetravelled(km)rods telecommute calculator100,00010,000tele-workbusiness meetingin sydney1,000100bicycledaily worktrainplanecar101tr

29、avel0.1101001101001000mb/s-hrgb/s-hrbitrate-time productlargethe khazzoom-brookes postulate energy efficiency at micro levelreduction of energy use at this level but leads to an increase in energy use, at the macro level example:wide bodied passenger aircraftlower costs per passengerincrease in air travelincreased greenhouse emissions“energy efficiency