1 / 63

Energetic Efficiency (EE) over Wired Networks

Energetic Efficiency (EE) over Wired Networks. Alberto E. García Telematic Engineering Group, University of Cantabria, Spain agarcia@tlmat.unican.es. Disclaimer. The contents of this presentation include figures and results from several Internet sources .

kirby
Download Presentation

Energetic Efficiency (EE) over Wired Networks

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. Energetic Efficiency (EE) over Wired Networks Alberto E. García Telematic Engineering Group, University of Cantabria, Spain agarcia@tlmat.unican.es

  2. Disclaimer • Thecontents of thispresentationinclude figures and resultsfromseveral Internet sources. • Author has triedtoincludeallthecorrespondingreferences, included at theend of thispresentation • Withall, themajorpart of thepresentedtechnologies and solutions are actuallyoverresearching and thelecturerrecommendsto complete theinformationforeachoptiondirectlyoverthesourcetoobtain concrete details. • Ifyou are notreferencedyet, I’msorry and I’llincludeyourname in nextversions of thisdocument.

  3. Agenda • Powerconsumptionovertheaccess • Powerconsumptionoverthebackbone • WhyEnergyEfficiencyovernetworks • EE overthepractice: someproposals • Mainkeywords: • EE techniques • Circuit switching • Slow down • Turn off hardware • EE applications/protocols

  4. Internet: theorigin Traditional hierarchical Internet topology Sprint, MCI, AGIS, Uunet, PSINet Settlement Free NationalBackboneOperators • Most traffic flowing between the core tier-1 provider NAP NAP Regional Access Providers Payfor BW Payforaccess BW ISP3 ISP2 ISP1 Local Access Providers Customer IP Networks Customers and businesscustomers

  5. Internet: actually Emerging flat Internet topology • Most traffic flowing directly between content providers and customer IP networks • Part flows to the global Internet through Internet Exchange Points (IXP) Global Transit / NationalBackbones “HyperGiants” Large Content, Consumer, Hosting CDN Settlement Free Global Internet Core IXP IXP IXP Payfor BW Regional / Tier2 Providers ISP1 ISP2 Payforaccess BW Customer IP Networks Customers and businesscustomers

  6. Internet: thepracticalview Source: Tucker, Baliga et all

  7. And …cloudcomputing??? Publiccloud Privatecloud Source: Tucker, Baliga et all

  8. EE: Whatistheobjective? Communication from the Commission - Energy efficiency: delivering the 20% target /* COM/2008/0772 final */ • The optimization of energy consumption is a crucial topic • European Commission promoted the ambitious “20-20-20” • 20% cut in emissions of greenhouse gases by 2020 (compared with 1990 levels); a • 20% increase in the share of renewable sources in the energy mix • 20% cut in energy consumption. • Transport networks are into the hurricane eye

  9. Whichtransportnetworks? Source Bolla et all

  10. Isitpossible??? Source: TELIT

  11. NGN isanexplanationtoo Source STRONGEST Project

  12. It’sthe time of the (ICT) networks

  13. Thereality of theaccess • Coexistence of severalaccessworlds • Limitsdepending of technology

  14. Access: What are thelimits? Source: “Powerconsumption in telecommunicationnetworks: overview and reductionstrategies”, IEEE Communications Magazine, June 2011 SourceECOnet Project (lowEnergyCOnsumptionNETworks) https://www.econet-project.eu/

  15. Access: Howmany and When? • Power as a function of access rate with an oversubscription rate of 20. • Expected power consumption of latest • Base access rate in 2010 = 5 Mb/s.

  16. Access: and thewinneris…

  17. Comparingwiththerest • Opticaltechnologies are mosteficient • Backboneequipment more efficientthanaccessequipment

  18. And thebackbone? • Backboneequipmentsevolvedtoo • But, whatistheproblem? • comsumptionproportionaltothebandwidth Source: Tucker, Baliga et all

  19. Inside a router Source: Tucker et al

  20. EE in routers: whoiswho?

  21. Whydoesthenetworkneed EE ? • Currently: • Access isdominatingoverallenergyconsumption • In future…nearfuture!!! • Corewillbecomeincreasinglyimportant • Feasibilitylimitsmay be reached (growingtrafficvolumes & more nodeconsolidation) • Innovationneeded: • To keep this under control • Tostaywell below these limits • Even avoid further growth of energy

  22. EE: somesuggestions

  23. Practical EE: a formal approach Scalable, Tunable and Resilient Optical Networks Guaranteeing Extremely-high Speed Transport Source: STRONGEST project http://www.ict-strongest.eu/

  24. The “Strongest” Keywords • A. Improved inherent energy-efficiencies as offered by electronics technologies • (1) More efficient CMOS technologies • (2) High temperature operation of ICs • B. More sophisticated management and exploitation of network resources • (3) Source coding & caching • (4) Multi-layer traffic engineering (MLTE), • (5) Powering down, sleep/idle modes and burst-mode operation. • C. The inherent energy efficiencies as offered by optics technology solutions • (6) Optical bypass • (7) Coherent detection • (8) Polarisation multiplexing. • D. More environmentally sustainable approaches to network design such as • (9) Micro-power generation • (10) Increased reliability and robustness of network equipment.

  25. EE: somesolutions • Techniques for improving energy efficiency in communication networks: • EE technics • Circuit switching less work/buffering • Slow down • Turn off hardware

  26. EE Technics SourceAleksic

  27. Routing vs. switching Source: ECOnet

  28. CircuitSwitching • Internet traffic: large share of predictableservices (e.g. video) • Circuitswitching = EE • Pipeline forwarding • No buffers & no headerprocessing • ++ faster • Application: video file/stream per opticalcircuit

  29. Switching: theevolution Source: Tucker et al

  30. SDH/SONET + WDM • All traffic is passed up through the lower layers and processed by the IP router • Traffic is processed by the SDH/SONET switch, bypassing the IP router • Traffic is switched by the optical cross-connect, bypassing both the SDH/SONET and IP layers. Lower layers are progressively more energyefficient.

  31. Point topoint WDM netowork

  32. Optical IP network

  33. Wabeband IP network

  34. Source: Tucker et al

  35. OpticalBurstswitching • Requires increased number of lightpaths for a given blocking probability • Switch technology requires fast reconfiguration time: • More costly per port than “slow”OXC’s(MEMS etc.)

  36. OBS: application • Pro’s • Requiresfewer OXC ports • Con’s • OXC ports must be wideband • Requires waveband (i.e. multi-channel) wavelength conversion • Dispersion issues Source: Y. Huang et al., OFC 2004, Pathibanet al., OFC 2006

  37. Opticalpacketswitching Source: Tucker et al • No viable optical buffering technology in sight • Optical switch fabrics may become competitive with CMOS • Not clear whether optical packet switching will solve the energy bottleneck problem

  38. Slowdowntechniques • Desktop links have low utilization • Snapshot of a typical 100 Mb Ethernet link • Shows time versus utilization (trace from Portland State Univ.) • Some Server links have low utilization • Snapshot of a File Server with 1 Gb Ethernet link • Shows time versus utilization (trace from LBNL)

  39. Slowdown & buffering • Buffering a Selectivesleep: • Bufferingthepacketstolargerbursts • Twovariations • Managed: monitorsschedules links being up • Autonomous: waitsdependingon buffer size • Disadvantage: • More burstytraffic • LowerQoS

  40. Adaptive link rate • Based on works of Dr. Ken Christensen from University of South Florida and Bruce Nordmanfrom LBNL • “Ethernet Adaptive Link Rate: System Design and Performance Evaluation”, Gunaratne, C., Christensen, K.; Proceedings 2006 31st IEEE Conference on Local Computer Networks, Nov. 2006 Page(s):28 - 35 SourceBianzino et al

  41. Adaptive Link Rate • No power-awareoptimizations • Only idle logic • Only performance scaling • Performance scaling and idle logic

  42. Adaptive Link Rate: expectedresults

  43. Lowpower idle • Smart switching of interface to low power mode • LPI replaces the continuous IDLE signal • Energy consumption around 10% of active mode • Synchronisationis kept with signalling during short periodic refresh intervals (Tr) • Trade-off energyforlatency • Transitions between active/sleep modes take time • Ts– Time to go to low power (sleep) mode • Tw – Time to wake up link

  44. LPI • CombinedwithAdaptive Link Rateisthebasis of 802.3az • September 2010 – IEEE Std 802.3az approved

  45. Turn off hardware • MultilayerTrafficEngineering • Control IP/MPLS flowrouting & logicaltopology • Adaptthesetochangingtrafficdemand Fast MLTE (off-peak = switch-off) Slow MLTE (Adaptnetworktotraffic) Static full mesh Source: “PowerreductionTechniques in MLTE, Bart Puype et al, IBCN

  46. Turn off basedontrafficdemand Source: http://asert.arbornetworks.com/2009/08/what-europeans-do-at-night

  47. EE Powerover Ethernet Source: “Powerover Ethernet (PoE): AnEnergy-EfficientAlternative”, RomanKleinerman, Marvell and Daniel Feldman, MicrosemiCorporation, in Marvell Whitepaper, 2011

  48. Energy-awarerouting • Routers are put to sleep when the network load is low, while preserving connectivity. This technique may increase the load on some links SourceBianzino et al

  49. MLTE results • Twooptimizationmodes: • Slow: days/week • Fast: hours • Powerscales • Betterwithtrafficvolume as MLTE actsfaster Source: “Powerreductiontechniques in multilayertrafficengineering”, Bart Puype et al, IBCN

More Related