Project GreenLight

679 views

Published on

Dr. Gregory Hidley
California Institute for Telecommunications and Information Technology, UCSD

Published in: Technology, Business
0 Comments
0 Likes
Statistics
Notes
  • Be the first to comment

  • Be the first to like this

No Downloads
Views
Total views
679
On SlideShare
0
From Embeds
0
Number of Embeds
31
Actions
Shares
0
Downloads
9
Comments
0
Likes
0
Embeds 0
No embeds

No notes for slide
  • Packet network in a box --- E1200. Passive DWDM enables huge bandwidth.
  • Project GreenLight

    1. 1. Project GreenLight<br />Presentation for the<br />…<br />…<br />Dr. Gregory Hidley<br />California Institute for Telecommunications and Information Technology, UCSD<br />
    2. 2. ICT is a Key Sector in the Fight Against Climate Change<br />Applications of ICT could enable emissions reductions of 7.8 Gt CO2e in 2020, or 15% of business as usual emissions. But it must keep its own growing footprint in check and overcome a number of hurdles if it expects to deliver on this potential.<br />www.smart2020.org<br />
    3. 3. ICT Industry is Already Actingto Reduce Carbon Footprint<br />
    4. 4. Application of ICT Can Lead to a 5-Fold GreaterDecrease in GHGs Than its Own Carbon Footprint<br />“While the sector plans to significantly step up the energy efficiency of its products and services, ICT’s largest influence will be by enabling energy efficiencies in other sectors, an opportunity that could deliver carbon savings five times larger than the total emissions from the entire ICT sector in 2020.”<br />--Smart 2020 Report<br />Major Opportunities for the United States*<br />Smart Electrical Grids<br />Smart Transportation Systems<br />Smart Buildings<br />Virtual Meetings<br /> * Smart 2020 United States Report Addendum<br /> www.smart2020.org<br />
    5. 5. GreenLight Motivation: The CyberInfrastructure (CI) Problem<br />Compute energy/rack : 2 kW (2000) to 30kW in 2011<br />Cooling and power issues now a major factor in CI design<br />IT industry is “greening” huge data centers … but today every $1 spent on local IT equipment will cost $2 more in power and overhead<br />Academic CI is often space constained at departmental scale<br />Energy use of growing departmental facilities is creating campus crises of space, power, and cooling<br />Unfortunately, little is known about how to make shared virtual clusters energy efficient, since there has been no campus financial motivation to do so<br />Challenge: how to make data available on energy efficient deployments of rack scale hardware and components?<br />
    6. 6. The NSF-Funded GreenLight ProjectGiving Users Greener Compute and Storage Options<br />PI is Dr. Thomas A. DeFanti<br />$2.6M over 3 Years to construct GreenLight Instrument<br />Start with Sun Modular Data Center<br />Sun Has Shown up to 40% Reduction in Energy<br />Measures Temperature at 5 Levels in 8 Racks<br />Measures power Utilization in Each of the 8 Racks<br />Chilled Water Cooling input and output temperatures<br />Add additional power monitoring at every receptacle<br />Add web and VR interfaces to access measurement data<br />Populate with a variety of computing clusters and architectures<br />Traditional compute and storage servers<br />GP GPU arrays and specialized FPGA based coprocessor systems<br />DC powered servers<br />SSD equipped systems<br />Turn over to investigators in various disciplines <br />Measure, Monitor and Collect Energy Usage data<br />With the goal of maximizing work/watt<br />
    7. 7. The GreenLight Project: Instrumenting the Energy Cost of Computational Science<br />Focus on 5 Communities with At-Scale Computing Needs:<br />Metagenomics<br />Ocean Observing<br />Microscopy <br />Bioinformatics<br />Digital Media<br />Measure, Monitor, & Web Publish Real-Time Environmental Sensor Output<br />Via Service-oriented Architectures<br />Allow Researchers Anywhere To Study Computing Energy Cost<br />Enable Scientists To Explore Tactics For Maximizing Work/Watt<br />Develop Middleware that Automates Optimal Choice of Compute/RAM Power Strategies for Desired Greenness<br />Partnering With Minority-Serving Institutions Cyberinfrastructure Empowerment Coalition <br />Source: Tom DeFanti, Calit2; GreenLight PI<br />
    8. 8. UCSD Energy Instrumented Cyberinfrastructure<br />Active Data Replication<br />10 Gigabit L2/L3 Switch<br />Eco-Friendly <br />Storage and Compute<br />N x 10 Gbit<br />N x 10 Gbit<br />Wide-Area 10G<br /><ul><li>Cenic/HPR
    9. 9. NLR Cavewave
    10. 10. Cinegrid
    11. 11. Cisco
    12. 12. …</li></ul>“Network in a box “<br /><ul><li> > 200 Connections
    13. 13. DWDM or Gray Optics</li></ul>On-Demand Physical Connections<br />Your Lab Here<br />Microarray<br />Source:Phil Papadopoulos, SDSC/Calit2<br />
    14. 14. GreenLight Goals: More Work/Watt<br />Build a full-scale virtualized device, the GreenLight Instrument<br />Measure then minimize energy consumption<br />Develop middleware to automate optimal choice of compute/RAM power strategies<br />Discover better power efficiency configurations and architectures <br />Teach future engineers who must scale from an education in Computer Science to a deeper understanding in engineering physics<br />Measure, monitor, and make publicly available, via service-oriented architectures, real-time sensor outputs<br />Focus on 5 communities: metagenomics, ocean observing, microscopy, bioinformatics, and digital media<br />Allow researchers anywhere to study the energy cost of at-scale scientific computing<br />
    15. 15. GreenLight Research activitiesLeading to Greener CI Deployments<br />Computer Architecture – FPGA, GP GPU systems <br />Rajesh Gupta/CSE<br />Software Architecture – Virtualization, memory management, networking and modeling<br />Amin Vahdat, Ingolf Kruger/CSE<br />CineGrid Exchange – mixed media storage, streaming, and management<br />Tom DeFanti/Calit2<br />Visualization – Using 2D and 3D modeling on display walls and CAVEs <br />Falko Kuster/Structural Engineering, Jurgen Schulze/Calit2<br />Power and Thermal Management <br />Tajana Rosing/CSE<br />DC Power Distribution<br />Greg Hidley/Calit2<br />http://greenlight.calit2.net<br />
    16. 16. Calit2/UCSD [http://greenlight.calit2.net]<br />Monitoring, Modeling and Management<br />GLIMPSE<br />Decision Support System<br />http://glimpse.calit2.net<br />
    17. 17. Situational Awareness<br />Calit2/UCSD [http://greenlight.calit2.net]<br />12<br />Dashboard interface<br />“Tap”<br />for details<br />Power utilization<br />Enterprise reach<br />Multiple perspectives<br />
    18. 18. Datacenter vitals<br />2010.08.20<br />Calit2/UCSD [http://greenlight.calit2.net]<br />13<br />Input/Output<br />sampling<br />Live/average Fan speeds<br />Live/Average<br />data<br />Live Temperature<br />Heat Exchangers<br />Environmentals<br />
    19. 19. Domain specific views<br />2010.08.20<br />Calit2/UCSD [http://greenlight.calit2.net]<br />14<br />Control elements<br />Real-time heatmap<br />Realistic models<br />
    20. 20. Airflow dynamics<br />2010.08.20<br />Calit2/UCSD [http://greenlight.calit2.net]<br />15<br />Live <br />fan speeds<br />Airflow dynamics<br />
    21. 21. Heat distribution<br />2010.08.20<br />Calit2/UCSD [http://greenlight.calit2.net]<br />16<br />Combined heat + fans<br />Realistic correlation<br />
    22. 22. Heat Trends<br />Calit2/UCSD [http://greenlight.calit2.net]<br />17<br />Heat exchangers<br />Hotspot identification<br />Trends over past 24h<br />
    23. 23. Past changes in airflow<br />Calit2/UCSD [http://greenlight.calit2.net]<br />18<br />Fan slices<br />rpm<br />Heat distribution changes<br />Potential for failures<br />Trends over past 24h<br />
    24. 24. Power spikes<br />Calit2/UCSD [http://greenlight.calit2.net]<br />19<br />IT assets<br />Computation zone<br />Unused asset<br />Average load<br />Peak computation<br />1 minute<br />resolution<br />
    25. 25. Zoom-in Analysis<br />Calit2/UCSD [http://greenlight.calit2.net]<br />20<br />History over several days.<br />Zoom on desired time range.<br />Hint on each sample point.<br />Automatic average area.<br />Multiple sensors per asset with up to 1 min sampling resolution.<br />
    26. 26. The GreenLight Project Focuses on Minimizing Energy for Key User Communities<br />Microbial Metagenomics<br />Ocean Observing<br />Microscopy<br />Bioinformatics<br />Digital Media—CineGrid Project<br />Calit2 will Host TBs of Media Assets in GreenLightCineGrid Exchange to Measure and Propose Reductions in the “Carbon Footprint” Generated by:<br />File Transfers and <br />Computational Tasks <br />Required for Digital Cinema and Other High Quality Digital Media Applications<br />
    27. 27. GreenLight Project: Putting Machines To Sleep Transparently<br />22<br />Rajesh Gupta, UCSD CSE; Calit2<br />Laptop<br />Network interface<br />Peripheral<br />Low power domain<br />Secondary processor<br />Network interface<br />Management<br />software<br />Main processor,<br />RAM, etc<br />SomniloquyEnables Servers to Enter and Exit Sleep While Maintaining Their Network and Application Level Presence <br />
    28. 28. Improve Mass Spectrometry’s Green Efficiency By Matching Algorithms to Specialized Processors <br />Inspect Implements the Very Computationally Intense MS-Alignment Algorithm for Discovery of Unanticipated Rare or Uncharacterized Post-Translational Modifications<br />Solution: Hardware Acceleration with a FPGA-Based Co-Processor<br />Identification and Characterization of Key Kernel for MS-Alignment Algorithm<br />Hardware Implementation of Kernel on Novel FPGA-based Co-Processor (Convey Architecture)<br />Results: <br />300x Speedup & Increased Computational Efficiency<br />Large Savings in Energy Per Application Task<br />
    29. 29. Virtualization at Cluster Level for Consolidation and Energy Efficiency<br />Source: Amin Vadhat, CSE, UCSD<br />Fault Isolation and Software Heterogeneity, Need to Provision for Peak Leads to:<br />Severe Under-Utilization<br />Inflexible Configuration<br />High Energy Utilization<br />Usher / DieCast enable:<br />Consolidation onto Smaller Footprint of Physical Machines<br />Factor of 10+ Reduction in Machine Resources and Energy Consumption<br />Original Service<br />Usher<br />Virtualized Service<br />
    30. 30. DC Power: UCSD is Installing Zero Carbon EmissionSolar and Fuel Cell DC Electricity Generators<br />UCSD 2.8 Megawatt Fuel Cell Power Plant Uses Methane<br />Available Late 2011<br />San Diego’s Point Loma Wastewater Treatment Plant Produces Waste Methane<br />2 Megawatts of Solar Power Cells at <br />UCSD, 1 MW to be<br />Installed off campus<br />
    31. 31. Zero Carbon GreenLight Experiment:DC-Powered Modular Data Center <br />Concept—Avoid DC to AC to DC Conversion Losses<br />Computers Use DC Power Internally<br />Solar and Fuel Cells Produce DC<br />Both Plug into the AC Power Grid<br />Can We Use DC Directly (With or Without the AC Grid)?<br />DC Generation Can Be Intermittent <br />Depends on Source <br />Solar, Wind, Fuel Cell, Hydro<br />Can Use Sensors to Shut Down or Sleep Computers<br />Can Use Virtualization to Halt/Shift Jobs<br />Experiment Now Underway<br />Collaboration with Sun, EPRI, DPTI and LBNL<br />NSF GreenLight Year 2 and Year 3 Funds <br />Sun Box <200kWatt<br />Source: Tom DeFanti, Calit2; GreenLight PI<br />
    32. 32. DC Power Distribution<br />Today's data center AC based distribution model:<br />480-V AC power from the grid is converted to <br />DC to charge a battery-based UPS <br />DC stream from the UPS is converted to AC and <br />transformed to 208-VAC for distribution<br />The above AC is then rectified back to 380-VDC <br />in each server's power supply<br />DC distribution offers a comparatively simpler scheme:<br />a single rectifier turns 480-V AC into 380-V DC that both charges the UPS and supplies the servers. <br />Energy Power Research Institute (EPRI) and Duke Energy Corp. measured a 15 percent reduction in power consumption in a test of 380-V DC distribution at the utility's Charlotte, N.C., data center. <br />Net energy savings could be twice that, they claim, once the cooler-running equipment's reduced air conditioning burden is factored in.<br />
    33. 33. UCSD Two Rack DC Experiment<br />
    34. 34. Questions?<br />

    ×