Reduced and Alternative Energy for Cloud and Telephony Applications - PowerPoint PPT Presentation

Reduced and Alternative Energy for Cloud and Telephony Applications James Hughes, Fellow Cloud Computing www.huawei.com HUAWEI TECHNOLOGIES CO., LTD.

Agenda Energy usage in a Telco   Trends and direction Energy usage in data centers   Total Energy  Server Energy  Consolidation  Saving energy by selecting data center location  Latency issues Promising research areas   FAWN  Ceph What about data center recycling  HUAWEI TECHNOLOGIES CO., LTD.

Green Radio Scenarios Two Market Profiles:  Developed World  Developed Infrastructure  Saturated Markets  Quality of Service Key Issue  Drive is to Reduce Costs  Emerging Markets  Less Established Infrastructure  Rapidly Expanding Markets  Large Geographical Areas  Often no mains power supply  power consumption a major issue Peter Grant, Green Radio – The Case for More E ffj cient Cellular Base Stations, May 2009, University of Edinburgh HUAWEI TECHNOLOGIES CO., LTD.

CO 2 emissions per subscriber Energy Cost per Subscriber per year 3 electricity use is in radio access 2.6kg 9kg dominant energy requirement at Operation CO 2  Operation vs Embodied CO 2  Data Center a small part 8.1kg Embodied CO 2  Room for improvement 4.3kg energy CO 2 Base station Base station Mobile Tomas Edler, Green Base Stations – How to Minimize CO2 Emission in Operator Networks, Ericsson, Bath Base Station Conference 2008 HUAWEI TECHNOLOGIES CO., LTD.

Base Station Power use Central Equipment Combining/Demultiplexing Power Amplifier Transceiver Power conversion Cooling Fans Transceiver Idling Power Supply Tomas Edler, Green Base Stations – How to Minimize CO2 Emission in Operator Networks, Ericsson, Bath Base Station Conference 2008 HUAWEI TECHNOLOGIES CO., LTD.

Where does the data center power go?  Energy losses in the U.S. T&D system are ~7.2% Chiller 33%  1w server savings is a 3.3w Electrical Waste overall savings Humidifier 3% Power Heat IN OUT CRAC 9%  Ambient cooling  UPS IT Equipment 30% Indoor Data Center PDU 5% Heat UPS 18% Switchgear/generator 1% Lighting 1% http://climatetechnology.gov/library/2003/tech-options/tech-options-1-3-2.pdf http://doe.thegreengrid.org/files/temp/E12A2B5D-B0E1-CA1A-97C1553AF4A01249/Green_Grid_Guidelines_WP.pdf HUAWEI TECHNOLOGIES CO., LTD.

Economic arguments Labor  Servers and Power are 70% of Networking data center TCO Facilities Servers Other IT Power Data Center 3 Year TCO http://scap.nist.gov/events/2009/itsac/presentations/day2/Day2_Cloud_Blakley.pdf HUAWEI TECHNOLOGIES CO., LTD.

Breakdown of Server Power Consumption Server Power Consumption (Source: Intel Labs, 2008) HUAWEI TECHNOLOGIES CO., LTD.

The Argument for Server Consolidation Power usage vs Processor Utilization  Average server is 10% 400 Server Power Consumption utilized 300  Consolidation increases utilization 200 100 0 0 20 40 60 80 100 Server Utilization CPU Utilization and Power Consumption (Source: Blackburn 2008) HUAWEI TECHNOLOGIES CO., LTD.

: Current Profile – WRITEs n ACT= 36 RAM Power ACT WR WR PRE ACT WR WR Data In Data In Data In  Energy based mostly on access  25% static power utilization  RAM Dedup  Reduce static power utilization  Embodied Energy WRITEs 17: Power Consumption per Device 500 450 400 Device Power (mW) 350 300 Total RD/WR/Term Pow er Total Activate Pow er 250 Total Background Pow er 200 150 100 50 http://download.micron.com/pdf/technotes/ddr3/TN41_01DDR3%20Power.pdf 0 HUAWEI TECHNOLOGIES CO., LTD.

Data center location matters  On how many days is cooling Harbin Urumqi necessary in each city?  Assume air cooling Lanzhou  Cost of electricity?  Latency? Yichang  Network Capacity? Shenzhen HUAWEI TECHNOLOGIES CO., LTD.

Ambient Cooling, Harbin, China Average Temperature 100 75 50 25 0 Low High Air cooling Limit -25 Jan Mar May Jul Sep Nov http://www.travelchinaguide.com/climate/harbin.htm HUAWEI TECHNOLOGIES CO., LTD.

Eliminating Transmission loss  Colocate data centers with electricity generation  Dams have water to cool data centers http://en.wikipedia.org/wiki/Three_Gorges_Dam HUAWEI TECHNOLOGIES CO., LTD.

Hydroelectric power  Not Constant http://en.wikipedia.org/wiki/Three_Gorges_Dam HUAWEI TECHNOLOGIES CO., LTD.

Improvements in Networking  2,250 miles  24ms optical latency  130ms measured (San Francisco to Chicago)  5x improvement with AON HUAWEI TECHNOLOGIES CO., LTD.

Fast Array of Wimpy Nodes 2500 2000  Scales linearly qps (value = 100B) 1500  Using small processors 1000  High performance 500  Energy e ffj ciency qps 0  75x of a “Server” 0 2 4 6 8 10 # of fawn nodes  Can Fawn nodes scale to 1000s? System QPS Watts Queries/Joule Alix 704 6 117  Can the cost of Flash be Soekris (1) 334 3.75 89 Soekris (8) 2431 30 81 competitive to disk? Desktop+SSD 2728 80 34  When? Gumstix 50 2 25 Macbook Pro 53 29 1.8 Desktop 160 87 1.8 Server 600 400 ∗ 1.2 http://www.cs.cmu.edu/~dga/papers/fawn-pdl-tr-08-108.pdf HUAWEI TECHNOLOGIES CO., LTD.

Ceph: A Scalable, High-Performance Distributed File System …  Problem? File (ino,ono) oid … Objects  12 x 1TB disk server from Dell hash(oid) & mask pgid … PGs  ~$8000 … … … CRUSH(pgid) (osd1, osd2)  12 x 1TB disk sells for ~$1000  750W OSDs (grouped by failure domain)  Can this be reduced?  Ceph uses a $100 ARM server per 2 x 1TB disks  Does it scale?  Can the centralized metadata server be eliminated?  Reliability at scale?  180W? http://www.ssrc.ucsc.edu/Papers/weil-osdi06.pdf HUAWEI TECHNOLOGIES CO., LTD.

Recycling  Average server / storage system has a 3 year life  Companies like Amazon replace racks at the 3 year mark  How do we reduce the embodied CO 2 ?  Can we recover value from the waste?  Does “Fail in place” make a di fg erence?  “DESIGN FOR DISASSEMBLY TO RECOVER EMBODIED ENERGY”  Buildings http://eprints.qut.edu.au/2846/1/Crowther-PLEA1999.PDF HUAWEI TECHNOLOGIES CO., LTD.

Innovation over time  Moore’s law is not cause but e fg ect  Doubling every 2 years is 41%/yr  Feature size improvement of 19%/yr  Continuous process improvement  Green IT will be similar  Many improvements can be made  No single fundamental roadblock  Energy use (by itself) has no value  Increases in e ffj ciency lowers cost and increases competitiveness  In a commodity market, can lower cost indicate more green?  If government force manufacturers to pay recycling fee? HUAWEI TECHNOLOGIES CO., LTD.

Conclusion  Energy usage in a Wireless Telco focuses on base stations  Energy usage in data centers has much room for improvement  Location  Cooling, low transmission loss, UPS  Latency  Server Energy  Consolidation  Necessary but not su ffj cient  Can more/smaller processors be an answer?  Now that we know how to horizontally scale?  The future for energy savings is bright HUAWEI TECHNOLOGIES CO., LTD.

Thank You www.huawei.com