A Green Internet

1、a green internet information logistics and energy rod tucker arc special research centre for ultra-broadband information networks (cubin) my friends: jayant baliga, kerry hinton, rob ayre energy and the internet the internet video energy in core networkdistribution network data data center metro/edg

2、e network access network out why is energy important? opex greenhouse impact energy-limited capacity bottlenecks (“hot spots”) enabling energy efficiencies in other sectors gtco22 putting things into context information and communication technologies (ict) global emissions gtcoee business as usual 2

3、020 global emissions 2020 abatements otherict emissions 5% of total “smart 2020: enabling the low carbon economy in the information age,” gesi, 2008 summary estimating energy consumption of the internet where is the energy consumed - core, metro, access network, data centres? informatio

4、n logistics - cloud computing - travel replacement - airmail vs. internet the khazzoom-brookes postulate caveat: “making predictions is difficult especially about the future.” network energy model tier 1 network core network metro/edge network access network core router edfa optical cross connect fi

5、ber edge routers broadband network gateways ethernet switch dslam olt cabinet splitter cu fiber dsl pon hot spots olt switch cabinet dslam onu fiber cu fttn ptp server server storage storage video distribution network data center oversubscription oversubscription 3 mb/s in 2009 m = peak access rate

6、sold to user average access rate m = 25 0.12 mb/s in 2009 %ofelectricitysupply power(w/user) 5 baliga et al., 2008 power consumption of ip network 25 oversubscription = 25 20 15 2009 technology todays internet ( 3 mb/s) total routers 1.0 10 access (pon) sdh/wdm links 0.5 0 050100150200250 2009 2020

7、peak access rate (mb/s) traffic growth = 40% p.a power consumption in access networks access n/w wave7 ont-g1000i edge node splitterfiber gpon cisco 6513 hitachi 1220 splitter cabinet nec vf200f6 cu fttn vdsl2 nec am3160 cisco 4503 tc communications tc3300 fiber ptp hitachi 1220 axxcelera excelmax b

8、ts axxcelera excelmax cisco ubr10012 rf gateway fiber node cu cisco dcp3000 wimax hfc rf amp powerperuser(w) 30 0 m power consumption in access networks 20 users per sector =1 m= oversubscription 20 wimax fttn m=1 m=1 ptp 10 m= 1 m= 10 32 customers hfc pon m=1 m 10 0 1 10100 1000 peak access rate (m

9、b/s) pon ftth is “greenest” low-power states in user modems european commission directorate -general jrc joint research centre institute for the environment and sustainability renewable energies unit code of conduct on energy consumption of broadband equipment draft version 3 issue 15 17 july 2008 “

10、with implementation of this code of conduct, 5.5 millions tons of oil equivalent (toe) will be saved per year.” extract: adsl-cpe vdsl2-cpe gpon onu ptp onu off-state (w) 0.3 0.3 0.3 0.3 low-power state (w) 3.5 4.5 5.0 3.0 on-state (w) 4.0 6.0 9.0 5.0 energyperbit(j) 0.1 network energy consumption p

11、er bit 1000 100 j/b 100 1 j/b 10 total 1.0 routers access (pon) wdm (optical fibre trunk links) 0.01 2.5252502500 peak access rate (mb/s) energyperbit(nj) 2009 energy per bit in network devices 1000 100 10 1 0.1 0.01 2009 opticalpos ethernet coreponserver ampwdm switch router onu tx/rx (10 mb/s) cis

12、co crs-1 router largest routing system available today linecard chassis: 1.28 tb/s, 13.6 kw switch fabric chassis: 8 kw power consumption buffers, 5% switch fabric, 10% power and heat management, 35% control plane, 11% i/o, 7% ip look-up and forwarding engine, 32% energy in electronic integrated cir

13、cuits cwire linecard cmos asic cmos gates energy = egate + 1 2 cwire v 2 power = energy x bit rate switchingenergyes,j fe re switching energy in cmos 10-14 250 10 -16 130 cm os 65 atu s 32 ize (nm) 10-18 40% decrease /18 months (40% efficiency gain p.a.) 18 ? 10-20 200020102020 moores law: number of

14、 transistors x 2 each 18 months source: itrs 97-06 roadmaps x1993performance 10000 8 x w ore0% o y 4 m ffici e se router capacity growth router capacity x 2.5/18 m 12416 crs-1 (1.3 tb/s,13.6 kw/rack) 1000 100 10 (0.3 tb/s/rack) 12016 (80 gb/s/rack) cmo m 2/1 s la p.a. enc nergy router energy efficie

15、ncy improving at 20% p.a. neilson, jstqe 2006 1 19941996199820002002 2004200620082010 year based on g. epps, cisco, 2006 power(w/user) %ofelectricitysupply 5 effect of efficiency gains 25 20 15 overall technology efficiency improvement rate = 0% p.a total 1.0 routers 10 access (pon) sdh/wdm links 0.

16、5 0 050100150200250 0 2020 peak access rate (mb/s) baliga et al., 2008 power(w/user) %ofelectricitysupply 5 effect of efficiency gains 25 20 15 overall technology efficiency improvement rate = 0% p.a 5% p.a 10% p.a 1.0 10 target20% p.a 0.5 0 050100150200250 2020 peak access rate (mb/s) baliga et al.

17、, jlt, 2009 access network global packet core network towards energy-efficient networks application content server efficient clients and home networks access network servers and data centres global packet core network application content server alternatives for sleep mode adaptive sharing of energy

18、management complex pc adaptive control of contentsolutions e.g. optical/electronic sleep mode andcomponents based highly efficient powerbypass fast “wake up”, adaptive modes of on communication requirements adaptive processing content delivery network and clients virtualizationnetworks (e.g. iptv) a

19、fter: gladish et al., ecoc 2008 1 2 data centers data center electricity consumption is 1% of the global total1 energy consumption of data centers worldwide doubled between 2000 and 2006 2 incremental us demand for data centre energy between 2008 and 2010 is the equivalent of 10 new power plants2 1

20、mw data center koomey, 2008 revolutionizing data centre efficiencykey analyses” mckinsey & company, april 2008. 30 20 energy consumption in a data centre conversion loss lighting ict infrastructurecooling 50 40 “burden factor” of 1.8 to 2.5 for power consumption associated with cooling, 10 50%35% 15

21、% conversion/distribution and lighting 0 serverstoragenetwork sources: eyp mission critical facilities, cisco it, network world, customer interviews, apc information logistics management of data, including, transport, storage, and processing some typical issues/questions: wired vs. wireless broadban

22、d access can cloud computing save energy? renewable energy sources physical vs. internet delivery of data can internet travel replacement save the planet? optimum design of video on demand systems storage cloud computing enterprise data center servers, processors pon access network computer, storage

23、 core network olt splitter fiber onu core router oxc servers, processors broadband network gateways edge routers ethernet switch metro/edge network total energy consumption = energy/bit used in storage transport processing x storage public data center cloud total number of bits three cloud computing

24、 scenarios software-as-a-service stored on users computer with updates downloaded regularly service bureau most tasks done on lower end user machine, outsource the “big” jobs computing-as-a-service hosted and run on provider computer “farm” with data initially uploaded from user (thin client model)

25、can cloud computing save energy ? service bureau here is my raw computing resource service center metro/edge network video footage and edit commands access network mid-range computers service bureau hardware & software service bureau computing resource service center metro/edge network here is your

26、dvd image access network mid-range computers service bureau hardware & software averagepowerconsumption,w 160 40 energy consumption of service bureau model computer used for 20 hrs/week, plus some video encoding of hour videos 180 140 120 older computer transport energy dominates 100 mid-range compu

27、ter 80 60 low-end computer & outsourcing 20 0 020406080100120 number of encodings per week following the sun and the wind making best use of renewable energy locate computing and storage resources near sources of renewable energy move data to follow the sun and winds requires many orders of magnitud

28、e increase in a data transport capacity transport/efficiency trade-off data by airmail vs. data by the internet data by airmail: 3x105 32-gb cargo jet melbourne flash drives 107 gb 103 kg co220 nj/b sydney data by internet: (24 hours) 107 gb the internet 1000 gb/s for 24 hours 2 j/b 105 kg co2 energ

29、yperbit(j) energy consumption per bit 1000 100 j/b 100 10 1 j/b total 2 j/b 1.0 routers access (pon) 0.1 0.01 flash chips by airmail wdm links 2.5252502500 peak access rate (mb/s) using the internet for travel replacement video conferencing source: cisco, 2008 travel replacement greenhouse impact (c

30、02) air travel business meeting 500 kg/person return melbourne video conferencing 2 x 1 gb/s for 6 hours = 4 tb 5 kg/person sydney distancetravelled(km) rods telecommute calculator 100,000 10,000 tele-work business meeting in sydney 1,000 100 bicycle daily work train plane car 10 1 travel 0.1 101001

31、101001000 mb/s-hr gb/s-hr bitrate-time product large the 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 passenger increase in air travelincreased green