Energy Consumption in IP Networks

Transcription

1 Energy Consumption in IP Networks Rodney S. Tucker, Jayant Baliga, Robert Ayre, Kerry Hinton, Wayne V. Sorin ARC Special Research Centre for Ultra-Broadband Information Networks (CUBIN) University of Melbourne

2 Energy Consumption of the Network Hot spot Why should we be interested in energy? OPEX Greenhouse Impact Managing Hot Spots - Getting the energy in - Getting the heat out Energy-limited capacity bottlenecks Enabling energy efficiencies in other sectors Power In

3 Where are We Heading? More users More data-intensive applications, e.g. video More often and for longer periods Increasing demand operators provide faster access and increased core capacity New applications enabled by faster access Energy Consumption Grows

4 Summary Modeling energy consumption of the Internet - Core, metro, and access networks Where does the energy go? Will (can) optical switching technologies help to reduce energy consumption? A Word of Warning

5 What is the Carbon Footprint of Telecoms? Global Telecoms Footprint (devices & infrastructure) Fixed Narrowband Mobile Network Mobile handsets X increase 1450+% growth Footprint (MtCO 2 p.a.) Broadband Fixed Broadband Modems Adapted from SMART 2020: Enabling the low carbon economy in the information age, GeSI,

6 Where does the Energy Go? CRS-1 ~ 10 kw / rack Packet over Sonet Edge ~ 4 kw Fibre Amps Core Core Edge Core WDM Core Edge Core Core Core Metro Passive Optical Network OLT - 100W OLT - 100W Curb Curb Curb Curb Access ONU ~ 5-10W Baliga et al., Mb/s to the user

7 Number of Hops in the Internet Data Pr [ H = k] Multi ISP Network: hops Number of Hops, k Source: P. Van Mieghem, Performance Analysis of Computer Systems and Networks, Cambridge (2006)

8 Power Consumption of IP Network router hops 2007 Technology 20 Contention ratio = Power (W/user) Today s Internet (~ 2.5 Mb/s) Total Routers Access (PON) SDH/WDM Links 0.5 % of Electricity Supply Peak Access Rate (Mb/s) Baliga et al., 2007

9 Ultra-Broadband Access router hops 2007 Technology Power (W/user) 50 Contention ratio = 25 Total Access (PON) Routers SDH/WDM 2.5 % of Electricity Supply Peak Access Rate (Mb/s) Baliga et al., 2007

10 Energy Consumption in Access Networks Access N/W Cabinet NEC CM7710T Edge Node Splitter PON Cisco NEC CM7700S Cabinet NEC VF200F6 FTTN with VDSL2 Cisco 4503 Zyxel VES-1616F-34 NEC GM100 PtP Axxcelera ExcelMax CPE Axxcelera ExcelMax BTS WiMAX

11 Power Consumption in Access Networks 40 Oversubscription = Power Per User (W) 20 WiMAX PON FTTN PtP Peak Access Rate (Mb/s) Wireless access consumes more energy than optical access PON FTTH is greener than FTTN

12 Network Energy Consumption per Bit Energy per bit (J) 10-3 ~100 μj/b 20 hops ~1 μj/b Total Routers Access (PON) 10-8 WDM Links Peak Access Rate (Mb/s)

13 Observations Optical transport (WDM) consumes relatively little energy < 5% of energy > 25% of CAPEX Eliminating the O/E/O converters provides no significant benefit Access network dominates at low rates Standby/Sleep mode is key to reducing energy consumption Network routers dominate at higher rates Need to reduce hop count improve router efficiency (technology) manage routers better (sleep states) develop better network architectures using fewer routers manage distribution and replication of content (IPTV)

14 Power Consumption in Routers 1,000,000 Power consumption (W) 100,000 10,000 1, P ~ 10 P = C 2/3 where P is in Watts where C is in Mb/s 100 nj/bit? 10 nj/bit 1 1 Mb/s 1 Gb/s 1 Tb/s 1 Pb/s Router Throughput Source: METI, 2006, Nordman, 2007

15 Energy Bottleneck High-end router: Cisco CRS-1 Linecard Chassis Capacity: 0.64 Tb/s Power: 13.6 kw Switch Fabric Chassis: Power: 8 kw Fully equipped: Multi-rack router Capacity: 41 Tb/s Power ~ 1 MW X2 every 18 months Per Rack Source: Neilsen, 2006; Deutche Telekom, 2007

16 Electronic Routers Forwarding Engine Fibers J Switch Fabric Demutiplexers O/E Converters Switch Fabrics Buffers Multiplexers Reduced bit rate (i.e. parallel processing) Speed (throughput) is not a limitation Electronics Optics

17 Energy in Electronic and Optical Routers Line Card Amp I/O Amp O/E O/E Buffer Forwarding Engine Buffer Buffer Forwarding Engine Data Plane Switch Fabric Switch Control Routing Tables Routing Engine Control Plane Power supply inefficiency Fans and blowers Electronic (2008) Energy/bit 0.7 nj 3.2 nj 0.5 nj 1.0 nj 1.1 nj 3.5 nj Total 10 nj Electronic (2018) 10 pj 65 pj 10 pj 20 pj 25 pj 80 pj 210 pj Optical (2018) 20 pj 65 pj 20 pj? 20 pj 25 pj 80 pj 220 pj Optical Packet Switching is not a compelling alternative G. Epps, Cisco, 2007, ITRS, 2005, R. Tucker, JLT, 2006

18 Contention Resolution in the Wavelength Domain λ 1 λ n Forwarding Engine Forwarding Engine Forwarding Engine Switch Fabric Optical Burst Switch Optical Label Switch Fatal Flaw: Require large n for low blocking probability (n ~3-10 x) Power (W/user) Total (Wavelength-domain contention resolution ) Total (Conventional router) Wong, JLT 2006 Pathiban et al., JLT 2009 Routers 1.2 X Routers WDM 5 X Peak Access Rate (Mb/s) Access WDM

19 The Challenge 10 % - 20 % p.a. continuous improvement in efficiency Total Power Per User (W) Efficiency Improvement Rate = 0% p.a Target 5% p.a 10% p.a 20% p.a Peak Access Rate (Mb/s) % of Electricity Consumption Baliga, et al, 2008, unpublished

20 Summary Energy consumption currently dominated by the access network The energy bottleneck in routers is looming - More significant than the so-called electronic speed bottleneck Key strategies for efficient network design - Control energy in the access network (e.g. sleep mode in modems) - Reduce the hop count (i.e. agile optical bypass) - Energy-efficient network architectures - Continuous improvement in (electronic) router efficiency - Caching and content distribution networks - O/E/O conversions not necessarily a problem