F.E.A.R. Returns: 5 Steps Toward Data Center Efficiency Today
Objectives Benefits for ITS professionals Brief review of the FEAR data center Define data center efficiency Link efficiency to the FEAR data center Increase data center efficiency
Benefits Increase exposure to Owners ITS needs are addressed during budget development Educated Owner makes value instead of budget decisions Increase input in the design process Develop peer relationships with traditional consultants Ensure standards and requirements are implemented Increase revenue for the ITS professional
FEAR Model - Review Footprint determine the rack size based on planned equipment Energy evaluates the power allowance for process and support equipment Arrangement optimizes layout for space efficiency Redundancy considers space for support equipment to meet requirements
FEAR Model - Review Cell Module Block
FEAR Model - Review Detailed Block Sample Data Center
FEAR Model - Summary A simple process for concept planning Considers Footprint, Energy, Arrangement, and Redundancy Creates a modular plan that can be scaled to a project
Efficiency - Background What is today s Data Center? Comms and Network Equipment Storage Area Networks (SAN) Enterprise Process Servers
Efficiency - Background What affects data center efficiency? Facilities Power Management Infrastructure Cooling Compute Storage Network 2009 The Green Grid
Efficiency - Factors Where is energy consumed? Category Energy IT equipment and processes Processor duty cycle Application deployment density Internal power utilization Processors 170 W Pwr Supply Loss 158 W Storage 47 W Networking 47 W Peripherals 88 W Fans 78 W Power Draw 588 W
Efficiency - Factors Where is energy consumed? Facility equipment Electrical losses HVAC equipment effectiveness Redundancy considerations Category Chillers Fans Elec Losses Lights Energy 300 W 125 W 101 W 13 W Operations and controls Power Draw 539 W
Efficiency - Factors Where is energy consumed? Pwr Supply Losses 14% Cooling 38% Processors 15% Peripherals 8% Lights 1% Elec Dist 9% Fans 7% Storage 4% Networking 4%
Efficiency - Metrics How is data center efficiency measured? Data Center Productivity (DCP) Site Infrastructure Energy Efficiency Ratio (SI-EER) Deployed Hardware Utilization Ratio (DH-UR) Compute Units Per Second (CUPS)
Efficiency - Metrics How is data center efficiency measured? Power Utilization Effectiveness (PUE) [The Green Grid Consortium] Corporate Average Data Center Efficiency (CADE) [McKinsey Report] Data Center Performance Per Energy (DPPE) [Japan s Green IT Promotion Council]
Efficiency - Metrics How is data center efficiency measured? Consensus opinions: Efficiency does not capture full potential Economy measures performance throughout operating range US EPA & US DoE
Efficiency - Metrics How is data center efficiency measured? For now PUE is the standard PUE Total Data Center Power IT Equipment Power
Efficiency - Metrics How is data center efficiency measured? For now PUE is the standard 1,127127 kw 1.92 588 kw
Efficiency Meets FEAR Model conceived to emphasize efficiency Adapts to variable power densities Applies to Raised Floor or Slab Emphasizes Hot / Cold Aisle Configuration Flexible for Whole Room or In-Row Solutions
Solution #1 - Footprint Manage IT Utilization Consolidate 40% in Development/Test vs. Production Eliminate 30% of servers are "dead" Virtualize Average 10% processor utilization
Solution #1 - Footprint Manage IT Utilization Advantages Maximizes space Eliminates low use losses Drives economical operations Disadvantages Potential security / audit issues Virtualization can be complex and expensive
Solution #2 - Energy Expand the Envelope
Solution #2 - Energy Expand the Envelope Advantages No cost increase in equipment efficiency Subsequent increase in capacity / redundancy Disadvantages Warranty issues with legacy equipment Difficult to overcome paradigm Could affect back-up / recovery time
Solution #3 - Arrangement Plan and manage room load Option A - Spread the load Design room based on average density Allocate functions to manage average density Option B - Create density zones Designate specific rows for high / med / low density
Solution #3 - Arrangement Plan and Manage Room Load Advantages Easy to plan Enables right sizing of support equipment Disadvantages Difficult to implement in existing room Can create sprawl, countering efficiency efforts
Solution #4 - Redundancy Leverage site conditions Free cooling Water side economizer Air side Economizer Other renewable solutions Evaporative cooling Geothermal e
Solution #4 - Redundancy Leverage site conditions
Solution #4 - Redundancy Leverage site conditions Advantages Increase redundancy with reduced energy Enables right sizing of support equipment Major step toward sustainable operations Disadvantages Trends toward limited geography Potential air quality issues
Solution #5 - FEAR Measure & Control You can t improve what you can t measure Facilities Power Management Infrastructure t Cooling Compute Storage Network
Solution #5 - FEAR Measure & Control Advantages Automation enhances response to efficiency measures Enables direct feedback for efficiency metrics Provides predictive data for growth management Multiple levels available Disadvantages Can be expensive and complicated to implement
Solutions - Summary Manage Utilization - Footprint Expand the Envelope - Energy Manage Layout - Arrangement Leverage Site Conditions Redundancy Measure & Control - FEAR
Tools & Resources Resource Organizations BICSI Data Center Design Manual The Green Grid www.thegreengrid.com library and tools ASHRAE TC 9.9 9 DataCom Series
Tools & Resources Useful Tools DoE EERE ITP DC Pro profiling and system assessment tools The Green Grid Performance calculators and estimators Application white papers
F.E.A.R. Returns: 5 Steps Toward Data Center Efficiency Today Thank You!