for reducing energy usage -Dr. Anders S.G. Andrae, Huawei, Nov. 1, - PowerPoint PPT Presentation

The role of cloud computing, telepresence and telecommuting for reducing energy usage -Dr. Anders S.G. Andrae, Huawei, Nov. 1, 2013

Outline  Experience of Huawei  Introduction to energy consumption of ICT, Entertainment&Media and cloud  Cloud computing  micro implications  macro perspectives  Implications for energy saving – micro and macro  Telepresence  Telecommuting  Summary  Next steps

Experience of Huawei Life cycle assessments performed  Radio Base Stations  All sorts of mobile phones  Tablets  Metals  FTTx Networks  Radio Access Networks  Cloud Computing Networks  Quick LCA method developed Sector analysis performed  ICT+Entertainment&Media(E&M) Sector with a defined scope

Energy (≈electricity) usage by ICT and cloud  Cloud computing is readily available internet computing for many services (software, storage, computing , etc)  Energy saving is not main purpose of cloud which is mobility, availability, cost, security, scalability  Digital technologies will all move into the Cloud more or less via wireless transmission  The scope of ”Cloud” needs to be defined for each analysis of energy usage  Public cloud and Private cloud  Energy usage of cloud is correlated to private/public + low/high frequency of use of service  ”Cloud Computing” used ≈700 TWh in 2012, i.e. ≈40% of ICT and E&M electricity

Traffic types: ICT and cloud Access traffic trends 1800 1600 1400 1200 1000 ExaByte Mobile Data Share of Data Center traffic trends 800 Fixed Data Fixed + Wifi 20000 600 18000 400 16000 200 14000 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 12000 Year ExaByte Access (“Data -center-to- user”) 10000 Within and between Data centers Increasing Traffic Trends: 8000 Global Data Center IP Traffic 6000 4000  Mobile data share of access 2000  Fixed+WiFi share of access 0 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020  ”Within and between data Year centers” share of global data center  wireless cloud increasing

Emerging wireless cloud Shares of Access traffic Mobile Data Fixed Data Fixed + Wifi 36% 39% 42% 44% 47% 49% 52% 54% 55% 55% 55% Shares of Global Data Center IP traffic 25% 63% 60% 56% 29% 52% 49% 33% Access (“Data -center-to- user”) Within and between Data centers 45% 40% 37% 20% 16% 12% 9% 8% 6% 5% 3% 2% 1% 2% 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 77% 78% 82% 83% 83% 83% 83% 84% 84% 85% 85% 23% 22% 18% 17% 17% 17% 17% 16% 16% 15% 15% 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020

Trends for global IP traffic and energy Study 1: Total IP traffic 29x, 15% Trend is very clear. Study 2: Total IP traffic 20x, 25% annual improvement annual improvement 31 Data created is 26 26 growing steeply, 21 21 16 16 however, the related Traffic units Traffic units 11 11 Energy units Energy units 6 6 energy usage is 1 1 under control until 2010 2012 2014 2016 2018 2020 2010 2012 2014 2016 2018 2020 Year Year 2020. Suppliers,Operators Global (data center) IP traffic & electricity trends Study 3: Total IP traffic 11x, 10% 35000 annual improvement and Research 30000 13 Community work 25000 11 Traffic (EB) ExaByte & TWhrs 9 20000 together. 7 Electricity for ICT 15000 Traffic units (TWhrs) 5 10000 Energy units Total global electricity 3 (TWhrs) 5000 1 2010 2012 2014 2016 2018 2020 0 Year 2010 2015 2020 Year

Energy consumption by mobile: Theory >32% annual Mobile traffic (voice+data) and energy consumption: improvement (AI) relation to energy efficiency of energy/traffic 100 needed to reduce 51 energy as mobile traffic grows 51x 18 Traffic and energy units 10 EE =energy Traffic units 5 efficiency Energy units, 40% AI of EE Energy units, 20% AI of EE Energy units, 32% AI of EE • Mobile: 4G and 1.08 1 1 Energy units, 10% AI of EE 2010 2012 2014 2016 2018 2020 SDN radio access solutions will overall 0.31 become as efficient as WiFi solutions 0.1 Year

Energy consumption by mixed and mobile: Practical >12% and 19%, uJ/bit facts 2009-2012 and predictions 2013-2020 for wireless and mixeds network annual 10000 improvement of 2728 energy/traffic 1000 expected from 759 measurements of uJ/bit operator 1, mostly fixed mixed networks. 100 uJ/bit operator 2, mix of fixed and mobile 70 62 uJ/bit global mobile network scenario 15 Semi-empirical 21 10 case study shows that large improvements 1 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 ≈35% AI could be possible High utilization of mobile networks is key to their energy efficiency. The Cloud is accessed more and more via mobile access.

Energy consumption by ICT+E&M: Theory PMC 2013 2012: 2017: ≈1800TWh ≈2500 TWh LCA, 18% LCA, 18% Devices Devices, Devices 32% Data Networks Devices, Networks Data Centers, 47% Data Centers Centers, 15% Data 21% LCA Centers Networks, Networks, 20% 29% Observations on this projected 2017 data: • Direct consumption by devices is less than 1/3 of electricity; compared to 1/2 in 2012. • Data centers + networks combined will represent 1/2 of electricity usage • LCA (Manufacturing Upstream) remains approximately at the same level of contribution http://vmserver14.nuigalway.ie/xmlui/bitstream/handle/10379/3563/CA_MainArticle14_ all-v02.pdf?sequence=4

Micro: Cloud computing Case study: Physical Desktop (PD) vs Virtual Desktop (VD) Generally there are two types of office users: PD and VD users PD use the Desktops as Servers VD users instead use Servers in the Data Center. VD use Thin Clients to connect. This case study represents a private cloud using wired transmission. Wireless transmission might render different conclusions.

Micro: Cloud computing – PD scope End-user devices: Desktops+Screens+Mouses+ Private Keyboards Network Intranet Equipment: with Switches, Server/Applications/Storage/Firew Access all Gateway

Micro: Cloud computing – VD scope End-user devices: Thin Clients+Screens+Keybo Servers Switches Storages Firewalls ards+Mouses Virtual Data Center desktop APP OS Private VM Network Equipment: Switches, Access Gateway Data Center: Cabinets Batteries UPS Intranet Cooling Equipment

Micro: Cloud computing cont. End-User Equipment PD VD Mass Power [W] Life [years] Annual type [#] [#] [kg/#] time electricit y [kWh] Keyboards 488 488 1.25 - 3 Mouses 488 488 0.12 - 3 Thin Clients 0 488 0.605 <15.2 5 17,421 Screens 488 488 5.1 Lenovo tool 3 22,936 Desktops 488 0 11.3 -- ” -- 3 139,568

Micro: Cloud computing Data Center Equipment PD VD Mass Power [W] Life [years] Annual type [#] [#] [kg/#] time electricit y [kWh] Servers 0 2 10 (est.) Blade Servers Blade Servers 5 -Blade Servers 10 20 (est.) 90-130 (CPU 5 16000 model) Storages 0 2 90 (est.) 650 5 11120 Switches 0 4 10 (est.) 91 5 3108 Firewalls 0 2 10 ( Eudemon 75 5 1306 1000E Series Firewall) Annually 10 Batteries 80 (50Ah12V) 16.75 8 batteries UPS 1 105-430 20 Annually 0.05 UPS Annually 0.2 Cabinets 2 100 10 Cabinets Air Conditioners (20kW-40kW) 1 332-388 10 Annually 0.1 AC

Micro: Cloud computing Private Network PD VD Mass Power [W] Life [years] Annual Equipment type [#] [#] [kg/#] time electricit y [kWh] Switches 488/4 488/40=12.2 10 (est.) 91 5 9479.4 0=12. 2 Gateways 1 1 10 (est.) 70 (est.) 5 574 Cables Cut- Cut-off off

Micro: Cloud computing  VD is 36% lower than PD for CO2e and 43% in electricity  Typically micro cloud computing energy analyses show >50% reductions in CO2 emissions  Overheads energy could be noticable

Micro: Cloud computing VD is advantageous to PD mainly due to -less impact of end- user devices

Micro: Cloud computing VD is advantageous to PD mainly due to differences associated with the Desktop and Thin Clients life cycles.

Micro: Cloud computing With PUE 1.7 the cooling electricity became the highest individual contributor. (1.7-1) * electricity consumption in data center = 22,100 kWh/year

Micro: Cloud computing Green Power (≈0.1 kg CO2e/kWh)  VD CO2e per user reduced from 270 to 211 kg CO2e/user/year. The Physical Desktop less chance shifting to Green Power as it is more confined to grids?

Micro: Cloud computing Virtual Desktop Huawei World, 488 users Switch Use Firewall Network Switch Production Keyboard Mouse Production 1% Use 1% Production 1% 1% 3% Network Switch Use Thin Client Production 5% 2% Thin Client Use Data Center Production 8% 1% Storage Production Screen Use 1% Data Center Use 11% 11% Storage Use 5% Server Use 8% Server Screen Production Production 40% 2% Screen (Monitor) Production becomes the highest contributor.