HANDBOOK ON GREEN INFORMATION AND COMMUNICATION SYSTEMS Chapter 18: Towards Energy-Oriented Telecommunication Networks 1 Sergio Ricciardi, 2 Francesco Palmieri, 3 Ugo Fiore, 1 Davide Careglio, 1 Germán Santos-Boada, 1 Josep Solé-Pareta 1 Technical University of Catalonia, Spain 2 Second University of Naples 3 University of Naples Federico II © Technical University of Catalonia - Second University of Naples - University of Naples Federico II 1
Table of contents q Introduction q Background & Motivations q Network energy consumption q Energy-aware architectures q Energy models q Multilevel approach q Energy Efficiency & Energy Awareness q Energy Oriented Architectures © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Network energy consumption q In Italy, Telecom Italia, and in France, France Telecom, are the second largest consumers of electricity after the National Railway systems: 2 TWh per year. q In the UK, British Telecom is the largest single power consumer. [7][8] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
The Growing Dynamics q CPU computing power doubles every 18 months (Gordon Moore, 1964) q The available network bandwidth doubles every 6 months (George Gilder, 1992) Gilder’s Law Moore’s Law © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
![Network energy consumption q Moore’s and Gilder’s laws [2][3] have not had the expected](https://present5.com/presentation/ece2fe1436a6addb2cf5ff6774ec2449/image-5.jpg "Network energy consumption q Moore’s and Gilder’s laws [2][3] have not had the expected")
Network energy consumption q Moore’s and Gilder’s laws [2][3] have not had the expected counterpart in power consumption reduction (Jevons paradox [4]) q As a consequence, the total power required per node is growing faster and faster. Network infrastructures consume: h 22 GW 1, 16% worldwide produced electrical energy h Growth rate: 12% per year [24] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
![Network energy consumption Evolution of the bandwidth capacity and energy-per-bit consumption Source: [24] Period](https://present5.com/presentation/ece2fe1436a6addb2cf5ff6774ec2449/image-6.jpg "Network energy consumption Evolution of the bandwidth capacity and energy-per-bit consumption Source: [24] Period")
Network energy consumption Evolution of the bandwidth capacity and energy-per-bit consumption Source: [24] Period windows: 10 years Bandwidth increment: x 1000 Energy per bit: ÷ 100 Energy consumption increment: x 10 more than 10 years ago Technological advancements foreseen by Moore’s law have not been fully compensated by the same growth in energy-efficiency [47] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Energy consumption: access vs backbone Backbone energy Consumption 2009: < 10% l Backbone energy Consumption 2017: ~ 40% Energy consumption currently dominated by the access network because of the high number of end -point devices [31] l ADSL link: 2. 8 W, while using as the access infrastructure l GPON (gigabit-capable passive optical network): only 0. 5 W (80% improvement) l With rising traffic volume, the major consumption is expected to shift from access to core networks [12] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Environmental impacts l Human’s activities have severe impacts on the environment h h l Human ecological footprint h h l measures the humanity’s demand on the biosphere 1, 5 planet Earths in 2007 [48] Carbon footprint h l Energy-consumption, resource exploitation GHG emissions, climate changes, global warming & dimming, pollution Measures the total set of GHG emissions Three dimensions h h h Energy consumption (Wh) GHG emissions (kg CO 2) Energy Cost (€) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Alternate energy sources l l Major contributors to the GH effect: h water vapor 36– 72% h carbon dioxide CO 2 9– 26% h methane 4– 9% h ozone 3– 7% h nitrous oxide ~1% h chlorofluorocarbons ~1% Global Warming Potential (GWP) CO 2 equivalent (CO 2 e) q Green & dirty energy sources § Green: solar, wind, tide, geothermic § Dirty: fossil fueled (coal, oil, gas) and nuclear plants q Green sources are renewable q Use Green source as possible! © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Renewable energies q Advantages: § § Virtually endless (5. 000 millions years) Zero carbon (green): no GHG emissions nor pollution during the use phase Free energy: zero costs in the use phase (except maintenance operations) Renewable Energy Sources are beneficial over their entire Life-Cycle [34] q Drawbacks § § § Low yield (solar panels efficiency ~25% as maximum) Not always available (e. g. , day/night cycle) nor always foreseeable Not always applicable as the only energy sources (e. g. mobility) q Allow paradigm shift from centralized to distributed energy system (Smart Grid) § § Centralized system: 1 big power plant giving energy to the whole region Distributed system: n small power plants providing energy for private use and putting surplus energy into the power grid q Follow-the-whatever paradigm q Energy form a cost to a revenue source © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Towards energy-efficiency q Energy-Efficiency § refers to a technology designed to reduce the equipment energy consumption without affecting the performance, according to the do more for less paradigm. It takes into account the environmental impact of the used resources and constraints the computations to be executed taking into account the ecological and potentially the economic impact of the used resources. Such solutions are usually referred to as eco-friendly solutions. © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Is energy efficiency sufficient? q The simplest approach is improving the efficiency of the individual devices involved q Energy efficiency alone is not sufficient to achieve effective results (the Jevons paradox occur) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
The energy-oriented paradigm q Energy-Awareness § refers to an intelligent technology that adapts its behavior or performance based on the current working load and on the quantity and quality of energy that the equipment is expending (energy -feedback information). It implies knowledge of the (dirty or green) sources of energy that supply the equipment thus differentiating how it is currently being powered. Energy-aware solutions are usually referred to as ecoaware solutions. q Energy-Oriented Infrastructures § Energy-Efficiency + Energy-Awareness in a holistic, sustainable and systemic approach with smart grid power distribution network and entire life cycle assessment (LCA) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
An energy oriented approach q Energy as an additional constraint to operate into network and data centers infrastructures § § § Current worldwide energy shortages Rising costs of energy (as fossil sources become scarcer) Green House effect, Global Warming and Pollution Need for alternative and renewable energy sources Growing interests of governments and society into eco-concerns q Lack of a comprehensive energy-oriented paradigm § Energy-efficient + energy-aware solutions in a systemic approach § Renewable energy sources § Energy models q The basic Principle § Do more for less!!! © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Energy-aware architectures q Current router architectures are not energy-aware average offered load of 75% UDP traffic with different packet sizes Source: [25] UDP traffic with medium packet size with different features enabled The difference between idle and heavily loaded router vary only of 3% (about 25 W on 750 W)… Energy consumption is a function of capacity, not throughput © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Energy-aware architectures …but… Power consumption for different installed line cards configurations of the GSR Cisco Router Power consumption for different installed line cards configurations of the 7507 Cisco Router …power consumption of base system is 50% of full line cards configuration Focus on energy-aware architectures that can adapt their behavior, and so, their energy consumption, to the current traffic loads (advocated both by standardization bodies and governmental programs and assumed in many literature sources [25]) Source: [25] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Optical Networks are cheaper q Optical transport (WDM) consumes relatively little energy q Access network dominates at low rates q Eliminating the O/E/O converters is not the only solution q Network routers dominate at higher rates: § § reduce hop count improve router efficiency (technology) manage routers better (sleep states) develop better network architectures using fewer routers § manage distribution and replication of contents © Technical University of Catalonia - Second University of Naples - University of Naples Federico II Number of Hops in the Internet [31]
Optical vs Electronic devices Total power consumption vs the aggregated bandwidth for electronic and optical router technologies Source: [24] Optical Cross-Connect node (OXC) with micro-electro-mechanical system (MEMS) switching logic consumes about 1. 2 W per single 10 Gb/s capable interface, whereas a traditional IP router requires about 237 W per port. © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
3 R Regeneration, Optical amplifiers & dispersion compensation q 3 R regeneration should be avoided as much as possible in planning, designing and managing new paths throughout an optical infrastructure (60 W per channel) use optical amplifiers, try to avoid 3 R as much as possible q EDFA are more performing (higher gain, lower insertion loss, noise and cross-talk effects) than SOA but have also higher energy consumption (respectively 25 W and 3 W) [24] q Use dispersion compensation fibers (DSF ITU-T G. 653, NZ-DSF ITU-T G. 655/656) instead of “simple” single mode fiber (SMF, ITU-T G. 652) will reduce the dispersion of the optical signal traversing the fiber and reduce the number of required optical amplifiers [47] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Energy-aware architectures q Router power consumption: § Fixed part due for the device to stay on § Variable part somehow proportional to the traffic load Power (P) k. Watt Total power consumption Variable power consumption Fixed power consumption due to the base system Idle Router aggregated bandwidth © Technical University of Catalonia - Second University of Naples - University of Naples Federico II Load (L) Tbps
Sleep mode q Putting entire network nodes down when they are not used q Generic problems § § § Load balancing Time consuming Start-up & configuration problem + peak in power usage Lifetime (MTBF) Economic CAPEX & OPEX Per-interface sleep mode / Adaptive rate / Low Power Idle[39] / STOP-START Energy proportional computing / Downclocking © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
ALR & LPI hybrid concept l Idea: temporarily switching off or downclocking unloaded interfaces and line cards (per interface sleep mode) to save energy l Adaptive Link Rate (ALR) and Low Power Idle (LPI) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
ALR & LPI q ALR: Native and working link rate § dynamically modify the link rate according to the real traffic needs q LPI: transmission on single interface is stopped when there is no data to send and quickly resumed when new packets arrive § in contrast with the continuous IDLE signal used in legacy systems § few microseconds Ts = ~2 s , Tw = ~ 4 s (@10 Gbps) § Power consumption in LPI: ~10% of active mode § Power consumption in transition ~50%-100% of active mode § Depends on the packet arrival distribution -> buffering, packet coalescing and coordinated Ethernet [6][12][13] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
A Multi-layer approach q Network Administration Procedures § Are aware of the power state of the nodes and the currently set power policy q Network Engeneering and Management § Load balancing, scheduling & task distribution § Shut down idle nodes q Networking Protocols § § Energy-efficiency, energy-awareness, energy-oriented paradigm Decrease overall usage of network Optimized placement/replication of data to minimize equipment usage Minimize time to transfer data & reduce data to be transmitted (forecasting techniques) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
![Energy-oriented network infrastructure Source: [47] © Technical University of Catalonia - Second University of](https://present5.com/presentation/ece2fe1436a6addb2cf5ff6774ec2449/image-25.jpg "Energy-oriented network infrastructure Source: [47] © Technical University of Catalonia - Second University of")
Energy-oriented network infrastructure Source: [47] © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Green network control plane © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Conclusions q Energy as a new constraint – considering the current energy consumption growth rate – even stronger than the bandwidth capacity q Energy-aware architectures are necessary to minimize the energy consumption q Assess the power consumption of ICT and network infrastructures l In general: l Energy-efficient architectures l Energy-aware algorithms & protocols l Energy-oriented infrastructures (sustainable systemic approach + LCA) © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
References 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. Gartner press release, http: //www. gartner. com/it/page. jsp? id=503867, 2007. An inefficient Truth by the Global Action Plan, http: //www. globalactionplan. org. uk/upload/resource/Full-report. pdf. SMART 2020: Enabling the low carbon economy in the information age, The climate group, 2008. EU Spring Summit, Brussels, Mar. 2007. Global Action Plan Report, An inefficient truth, http: //www. globalactionplan. org. uk/, 2007 C. Lange, “Energy-related Aspects in Backbone Networks”, in Proc. ECOC 2009, Vienna, Austria, Sep. 2009. S. Pileri, “Energy and Communication: engine of the human progress”, INTELEC 2007 keynote, Rome, Italy, Sep. 2007. L. Souchon Foll, “TIC et Énergétique: Techniques d'estimation de consommation sur la hauteur, la structure et l'évolution de l'impact des TIC en France”, Ph. D. dissertation, Orange Labs/Institut National des Télécommunications, 2009. BT Press, “BT announces major wind power plans”, Oct. 2007, http: //www. btplc. com/News/Articles/Showarticle. cfm? Article. ID=dd 615 e 9 c-71 ad-4 daa-951 a-55651 baae 5 bb. Telefónica supplement, The environment and climate change, 2008 special report on corporate responsibility, Apr. 2009. H. D. Saunders, “The Khazzoom-Brookes postulate and neoclassical growth”, The Energy Journal, Oct. 1992. C. Lange, D. Kosiankowski, C. Gerlach, F. Westphal, A. Gladisch, “Energy Consumption of Telecommunication Networks”, in Proc. ECOC 2009, Vienna, Austria, Sep. 2009. Nature Photonics technology conference 2007, Tokyo, Japan, Oct. 2007. M. Gupta, S. Singh, “Greening of the Internet”, in Proc. ACM SIGCOMM 2003, Karlsruhe, Germany, Aug. 2003. R. S. Tucker et al. , “Evolution of WDM Optical IP Networks: A Cost and Energy Perspective”, IEEE/OSA Journal of Lightwave Technologies, vol. 27, no. 3, pp. 243 -252, Feb. 2009. B. St Arnaud, “ICT and Global Warming: Opportunities for Innovation and Economic Growth”, http: //docs. google. com/Doc? id=dgbgjrct2767 dxpbdvcf. A. Muhammad, Paolo Monti, Isabella Cerutti, Lena Wosinska, Piero Castoldi, Anna Tzanakaki, “Energy-Efficient WDM Network Planning with Protection Resources in Sleep Mode”, accepted for Globecom 2010, ONS 01. © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
References 18. L. Chiaraviglio, M. Mellia, F. Neri, “Energy-aware Backbone Networks: a Case Study”, Green. Comm – First International Workshop on Green Communications, Dresden, Germany, Jun. 2009. 19. W. Van Heddeghem, M. De Groote, W. Vereecken, D. Colle, M. Pickavet, P. Demeester, “Energy-Efficiency in Telecommunications Networks: Link-by-Link versus End-to-End Grooming”, in Proc. of ONDM 2010, Feb. 1 -3 2010, Kyoto, Japan. 20. R. S. Tucker, “Modelling Energy Consumption in IP Networks”, retrieved from: http: //www. cisco. com/web/about/ac 50/ac 207/crc_new/events/assets/cgrs_energy_consumption_ip. pdf. 21. W. Vereecken, W. Van Heddeghem, D. Colle, M. Pickavet, P. Demeester, “Overall ICT footprint and green communication technologies”, in Proc. of ISCCSP 2010, Limassol, Cyprus, Mar. 2010. 22. Juniper, http: //www. juniper. net. 23. M. Z. Feng, K. Hilton, R. Ayre, R. Tucker, “Reducing NGN Energy Consumption with IP/SDH/WDM”, in Proc. 1 st International Conference on Energy-Efficient Computing and Networking, Passau, Germany, pp: 187 -190, 2010, ISBN: 978 -1 -4503 -0042 -1. 24. BONE project, 2009, “WP 21 Topical Project Green Optical Networks: Report on year 1 and updated plan for activities”, No. E, FP 7 -ICT-2007 -1 216863 BONE project, Dec. 2009. 25. J. Chabarek, J. Sommers, P. Barford, C. Estan, D. Tsiang, and S. Wright, “Power awareness in network design and routing”, in Proc. IEEE INFOCOM, 2008. 26. S. Aleksic, “Analysis of Power Consumption in Future High-Capacity Network Nodes”, Journal of Optical Communications and Networking, vol. 1, no. 3, pp. 245 -258, Aug. 2009. 27. I. Cerutti, M. Tacca, A. Fumagalli, “The multi-hop multi-rate wavelength division multiplexing ring”, IEEE/OSA Journal of Lightwave Technologies, vol. 18, 2000. 28. Energy Star, “Small network equipment”, http: //www. energystar. gov/index. cfm? c=new_specs. small_network_equip. 29. Gordon E. Moore, “Cramming more components onto integrated circuits”, Electronics, Volume 38, Number 8, April 19, 1965. 30. George F. Gilder, “Telecosm: How Infinite Bandwidth Will Revolutionize Our World”, The Free Press, NY, 2000. © Technical University of Catalonia - Second University of Naples - University of Naples Federico II
Thanks for your attention! © Technical University of Catalonia - Second University of Naples - University of Naples Federico II 30
Скачать презентацию HANDBOOK ON GREEN INFORMATION AND COMMUNICATION SYSTEMS Chapter
ece2fe1436a6addb2cf5ff6774ec2449.ppt