A Green Internet. Information Logistics and Energy. Rod Tucker. ARC Special Research Centre for Ultra-Broadband Information Networks (CUBIN)

Transcription

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

2 Energy and the Internet The Internet Energy Core Network Video Distribution Network In Data Center Metro/Edge Network Access Network Data Out Why is Energy Important? OPEX Greenhouse Impact Energy-limited capacity bottlenecks ( hot spots ) Enabling energy efficiencies in other sectors

3 Putting Things into Context Information and Communication Technologies (ICT) Global Emissions GtCO GtCo 22 e Business as Usual 2020 Global Emissions 2020 Abatements Other ICT Emissions 5% of total SMART 2020: Enabling the low carbon economy in the information age, GeSI,

4 Summary Estimating energy consumption of the Internet Where is the energy consumed - Core, metro, access network, data centres? Information Logistics - Cloud Computing - Travel Replacement - Airmail vs. Internet The Khazzoom-Brookes postulate Caveat: Making predictions is difficult especially about the future.

5 Network Energy Model Core Network Core Router Optical cross connect EDFA Server Fiber Storage Data Center Hot spots Server Edge Routers Tier 1 Network Metro/Edge Network Storage Broadband Network Gateways Ethernet Switch Video Distribution Network Access Network DSLAM Cu DSL Cabinet Fiber OLT PON Splitter Cabinet Cu OLT FTTN ONU DSLAM Switch Fiber PtP

6 Oversubscription Oversubscription Peak access rate sold to user M = Average access rate ~ 3 Mb/s in 2009 M = 25 ~ 0.12 Mb/s in 2009

7 Power Consumption of IP Network 25 Oversubscription = 25 Power (W/user) Technology Today s Internet (~ 3 Mb/s) Access (PON) Routers Total SDH/WDM Links % of Electricity Supply Peak Access Rate (Mb/s) 250 Traffic Baliga et al., Growth 2008 = 40% p.a

8 Power Consumption in Access Networks Access N/W Wave7 ONT-G1000i Edge Node Splitter Fiber GPON Cisco 6513 Hitachi 1220 Splitter Cabinet NEC AM3160 NEC VF200F6 Cu FTTN VDSL2 Cisco 4503 TC Communications TC3300 Fiber PtP Cisco ubr10012 Hitachi 1220 RF Gateway Axxcelera ExcelMax BTS Fiber Node Cu RF Amp Axxcelera ExcelMax Cisco DCP3000 WiMAX HFC

9 Power Consumption in Access Networks users per sector M= Oversubscription M= 10 Power Per User (W) M = 1 M= 1 HFC WiMAX FTTN M= 10 PON M = 1 M = 1 32 Customers PtP M ~ Peak Access Rate (Mb/s) PON FTTH is greenest

10 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 July 2008 With implementation of this Code of Conduct, 5.5 Millions tons of oil equivalent (TOE) will be saved per year. Extract: Off-State (W) Low-Power State (W) On-State (W) ADSL-CPE VDSL2-CPE GPON ONU PtP ONU

11 Network Energy Consumption per Bit 1000 Energy per bit (μj) ~100 μj/b ~1 μj/b Total Routers Access (PON) WDM (Optical Fibre Trunk Links) Peak Access Rate (Mb/s)

12 Energy per Bit in Network Devices Energy per bit (nj) Optical Amp PoS WDM Tx/Rx Ethernet Switch Core Router PON Server ONU (10 Mb/s)

13 Cisco CRS-1 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% Control Plane, 11% Power and Heat Management, 35% I/O, 7% IP Look-Up and Forwarding Engine, 32%

14 Energy in Electronic Integrated Circuits C wire Linecard CMOS ASIC CMOS Gates Energy 1 = E + gate C 2 wire V 2 Power = Energy x Bit Rate

15 Switching Energy in CMOS Switching energy E s, J CMOS Feature size (nm) % decrease /18 months (40% efficiency gain p.a.) 18? Moore s Law: Number of transistors x 2 each 18 months Source: ITRS Roadmaps

16 Router Capacity Growth Router capacity x 2.5/18 m CRS-1 (~1.3 Tb/s,13.6 kw/rack) X 1993 Performance (~0.3 Tb/s/rack) (~80 Gb/s/rack) Moore s law x 2/18 m CMOS energy efficiency 40% p.a. Router energy efficiency improving at 20% p.a. Neilson, JSTQE Year Based on G. Epps, CISCO, 2006

17 Effect of Efficiency Gains 25 Power (W/user) Overall Technology Efficiency Improvement Rate = 0% p.a Access (PON) Routers Total SDH/WDM Links % of Electricity Supply Peak Access Rate (Mb/s) Baliga et al., 2008

18 Effect of Efficiency Gains 25 Power (W/user) Overall Technology Efficiency Improvement Rate = 0% p.a Target 5% p.a 10% p.a 20% p.a % of Electricity Supply Peak Access Rate (Mb/s) Baliga et al., JLT, 2009

19 Towards Energy-Efficient Efficient Networks Application Content Server Efficient Clients and home networks Access Network Servers and Data Centres Global Global Packet Packet Core Core Network Application Content Server Alternatives for complex PC Sleep mode and fast wake up, adaptive modes of network and clients Sleep mode Adaptive control of components based on communication requirements Adaptive sharing of content Highly efficient power adaptive processing Virtualization Energy management solutions e.g. optical/electronic bypass Content delivery networks (e.g. IPTV) After: Gladish et al., ECOC 2008

20 Data Centers Data center electricity consumption is ~1% of the global total 1 Energy consumption of data centers worldwide doubled between 2000 and Incremental US demand for data centre energy between 2008 and 2010 is the equivalent of 10 new power plants 2 1 MW Data Center 1 Koomey, Revolutionizing Data Centre Efficiency Key Analyses McKinsey & Company, April 2008.

21 Energy Consumption in a Data Centre Conversion Loss Lighting ICT Infrastructure Cooling % 35% 15% Server Storage Network Burden factor of 1.8 to 2.5 for power consumption associated with cooling, conversion/distribution and lighting Sources: EYP Mission Critical Facilities, Cisco IT, Network World, Customer Interviews, APC

22 Information Logistics Management of data, including, transport, storage, and processing Some 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 save the planet? Optimum design of Video on Demand systems

23 Cloud Computing Servers, Processors Core Network Enterprise Data Center Storage PON Access Network OLT Splitter Fiber ONU Computer, Storage Core Router OXC Public Data Center Servers, Processors Storage Edge Routers Broadband Network Gateways Ethernet Switch Metro/Edge Network Cloud Total Energy Consumption = Energy/bit used in Storage Transport Processing X Total number of bits

24 Three Cloud Computing Scenarios Software-as-a-Service Stored on user s 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) Can cloud computing save energy?

25 Service Bureau Computing Resource Service Center Metro/Edge Network Here is my raw video footage and edit commands Access Network Mid-range Computers Service Bureau Hardware & Software

26 Service Bureau Computing Resource Service Center Metro/Edge Network Here is your DVD Image Access Network Mid-range Computers Service Bureau Hardware & Software

27 Energy Consumption of Service Bureau Model Computer used for 20 hrs/week, plus some video encoding of ½ hour videos Average Power Consumption, W Older Computer Mid-range Computer Low-end Computer & outsourcing Transport energy dominates Number of Encodings per Week

28 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 magnitude increase in a data transport capacity Transport/efficiency trade-off

29 Data by Airmail vs. Data by the Internet Data by Airmail: Melbourne 3x GB flash drives 10 7 GB Cargo Jet Sydney 10 3 kg CO 2 20 nj/b Data by Internet: (24 hours) 10 7 GB 1000 Gb/s for 24 hours The Internet 2 μj/b 10 5 kg CO 2

30 Energy Consumption per Bit Energy per bit (μj) ~100 μj/b ~1 μj/b Total 2 μj/b Routers Access (PON) WDM Links Flash chips by Airmail Peak Access Rate (Mb/s)

31 Using the Internet for Travel Replacement Video Conferencing Source: CISCO, 2008

32 Travel Replacement Greenhouse Impact (C0 2 ) Air Travel Business Meeting Melbourne ~500 kg/person return Sydney Video Conferencing 2 X 1 Gb/s for 6 hours = 4 TB ~5 kg/person

33 Rod s s Telecommute Calculator 100,000 10,000 Tele-work Business Meeting in Sydney Distance travelled (km) 1, Bicycle Daily work Train Plane Car Travel Mb/s-hr Gb/s-hr Bitrate-Time Product

34 The Khazzoom-Brookes Postulate Energy efficiency at micro level reduction of energy use at this level But leads to an increase in energy use, at the macro level Example: Wide bodied passenger aircraft lower costs per passenger increase in air travel increased greenhouse emissions large Energy efficiency is absolutely the wrong approach to network design. The objective should be to make the network carbon neutral - Bill St. Arnaud, CANARIE D. Khazzoom Energy Journal,10,1987, L. Brookes Energy Policy, 20,1992 Inhaber, Why Energy Conservation Fails, Quorum, 2002 H. Herring,

35 Some Interest Groups and Organizations

36 Summary The Way Forward Energy consumption of the Internet is small, but growing Internet energy consumption dominated by Access network today Core network in the future A multi-faceted approach needed to build a green Internet: Improved efficiency in electronic and photonic devices Low-energy switching techniques Improved architectures New protocols Information Logistics A new approach to network design and management Melbourne Information Logistics Centre (MILC)? Centre for Information LOgistics (CILO)?