SlideShare una empresa de Scribd logo
1 de 30
Descargar para leer sin conexión
©Microsoft®
2009
Performance/
Watt/Dollar
Dileep Bhandarkar, Ph. D.
Distinguished Engineer
Global Foundation Services
Microsoft Corporation
Energy Efficiency
in Datacenters 2010
©Microsoft®
2009
CLOUD SERVICES
59 markets and
36 languages
76 markets and
48 languages
Our Consumer Cloud Services
©Microsoft®
2009
Our Commercial Cloud Services
©Microsoft®
2009
Service Key Metric
More than 1 billion authentications per
day
~ 400 million active accounts
Processes 3 – 5 billion e-mail messages a
day
More than 4 billion queries per month
10B messages each day/ 750M active
accounts
Over 450 million unique visitors monthly
Key Operational Metrics
©Microsoft®
2009
Datacenter Construction Costs
Land: 0%-2%
Core & Shell Costs: 5%-9%
Architectural: 4%-7%
Mechanical / Electrical: 70%-85%
Land
Core / Shell
Mech /Elec
Arch
Where the costs are:
>80% scale with power
<10% scale with space
Reduce!
©Microsoft®
2009
Why Power is Important?
Energy Consumption: US power rate 10.27
cents per Kilowatt hour) in 2008 according
to DOE/eia (http://www.eia.doe.gov/cneaf/electricity/epm/table5_3.html)
In a typical data center for every watt in
server power there can be another 0.5 to 1
watt consumed for power distribution
losses and cooling.
Power (Switch Gear,
UPS, Battery backup,
etc)
Cooling (Chillers,
CRACs, etc)
Building load
Demand from grid
IT load
Demand from servers,
storage, telco equipment,
etc
PUE = Total Facility Power/IT Equipment Power
©Microsoft®
2009
But There is More!
Datacenters can cost between $10M and
$20M per megawatt
Capital depreciation costs are important
Server acquisition costs may not dominate
TCO Basic 1U Server - 5 year TCO
Total Equipment cost
per server
Energy Usage cost per
server
Datacenter Capital cost
per server
Datacenter operating
cost per server
©Microsoft®
2009
Beyond Just Reducing PUE
PUE is an excellent metric for evaluating
effectiveness of datacenter infrastructure
Server performance per watt can vary
Rightsizing of server configuration can
improve overall energy efficiency
Low Power processors often offer best
performance per watt
Microsoft optimizes the total solution for
best performance per watt per TCO$
8
©Microsoft®
2009
Holistic Optimization
Datacenter and Servers together
PUE is an excellent metric for evaluating
effectiveness of datacenter infrastructure
Server performance per watt can vary
Rightsizing of server configuration can
improve overall energy efficiency
Low Power processors often offer best
performance per watt
Microsoft optimizes the total solution for
best performance per TCO$
9
Drive change in the industry
through strong partnerships
12
2 Socket server based on a half-
width server design
•6.3” x 16.7”
•2 CPU Sockets
•4 DIMM Slots per CPU Socket
•4 SATA HDD Connections
• Dual 1 GbE NIC
•One 16x PCIe Card Slot
16x PCIe Slot
4x SATA HDD
Connectors
Optional 16x PCIe
Card, not present in
basic configuration
2x CPU
Memory
DIMMs
13
2 to 4 TOR
(2 to 4 RU)
96 Servers
(Up to 48 RU)
2 Bulk Power
(2x 3RU)
Upper
Power
Domain
Lower
Power
Domain
1.0
1.07
1.18
1.0
1.10
1.6
1.0
1.24
1.56
1.0
0.87
0.76
1.0
0.79
0.46
2.33 GHz (50W) 2.67 GHz (80W) 3.167 GHz (120W)
Performance System Price System Power Perf/W Perf/W/$
Assumption:
Server Price = $2000 + CPUs, Server Power = 150W + CPUs
The sweet spot is often at low power processors,
especially when system price and power are considered.
Your mileage may vary!
ChicagoQuincy
Dublin
Amsterdam
Hong
Kong
Singapore
Japan
"Datacenters havebecomeasvitaltothefunctioning
ofsocietyaspowerstations."--TheEconomist
San
Antonio
Multiple global CDN locations
in the Americas, Europe and
APAC.
Microsoft has more than 10 and
less than 100 DCs worldwide.
Quincy, Washington: approx 500K sq ft, 27MW, uses entirely
hydro-electric power
San Antonio, Texas: approx 477K sq ft, 27MW, uses recycled
water for cooling
Chicago, Illinois: 707,000 square feet with critical power of
60 MW, uses water side economization, containers
Dublin, Ireland: approx 570K sq ft, up to 27MW, uses
outside air for cooling.
Microsoft’s Data Center Evolution
Containers
Scalability
Sustainability
Data Center
Colocation
Generation 1
Generation 4
Modular Data Center
Server
Capacity
Rack
Density
and
Deployment
Quincy and
San Antonio
Generation 2
Chicago and
Dublin
Generation 3
Deployment Scale Unit
IT PAC
Time to Market
Lower TCO
Scalable Data Center
©Microsoft®
2009
Chicago Datacenter
Two-story datacenter
First floor: container bay
medium reliability
Second floor: high-
reliability
traditional co-location
rooms
Can deploy at scale
Optimized energy and
cooling efficiency
Allows OEMs to build
customized solutions
based on Microsoft
specifications
©Microsoft®
2009
Container Selection Process
“CBlox” specification defined all of the
interface requirements to the Data Center:
Physical Dimensions – length, width, height, weight
Chilled Water temperature & flow rate
Power – voltage and wattage
Selection based primarily on:
$/Server
# of Servers/container
©Microsoft®
200921
©Microsoft®
2009
Chicago Container Installation
22
video
http://www.microsoft.com/showcase/en/us/details/2778257d-b0b3-4bbc-8ef0-63fc5a75acf2
©Microsoft®
2009
Modularize and pre-manufacture
the entire datacenter
Lower the Total Cost of
Ownership
Increase scalability and Right
Time to Market
Standardize components to
improve operations and reliability
Paradigm Shift to Ultra Modularity
©Microsoft®
2009
Microsoft’s Future Datacenters
• No mechanical cooling
• Ultra-efficient water
utilization
• Utilize more renewable
building materials
• Reduce time to market by >50%
• Low initial capital investment
• Scale with business demand
• 30 % - 50 % more cost effective
for the same class of service
• Outstanding PUE
©Microsoft®
2009
Gen 4 Datacenter Strategy
Standardized Gen, UPS, IT compute PACs and support
equipment
Multiple Datacenter Classes, Colo and Modular PAC’s
(Pre-Assembled Components)
Support equipment designed to be installed on
indoor skids or outdoor PAC’s
Primary Cooling via Airside Economizers
Secondary Cooling via Air Cooled or Water Cooled
Chillers, or Fluid Cooler
Warmest server intake temperatures compared to all
other Generations
Colo - No Raised Floor with Hot/Cold Aisle
containment
©Microsoft®
2009
ITPAC Design
26
video
http://www.microsoft.com/showcase/en/us/details/84f44749-1343-4467-8012-9c70ef77981c
20 foot, 170 KW, 9 rack Proof of Concept ITPAC
CE and ETL certified as a Air Handler
Operating outside for over 1 year
ITPAC POC Units
©Microsoft®
2009
Conclusions
Holistic optimization of server and
datacenter design
Platform energy efficiency remains a major
focus area (Perf/Watt/Dollar)
Stranding Power is expensive
Focus on Infrastructure Cost reduction via
high temperature and free air cooling
Workload analysis is critical to rightsizing
Total Cost of Ownership is important
Measure and Control Server Parameters
Conventional Wisdom is often wrong!
28
©Microsoft®
2009
Questions?
29
More information at:
www.globalfoundationservices.com
© 2008 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries.
The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market
conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. Microsoft
is a trademark of the Microsoft companies. The Microsoft Financing marks are used by CIT Financial Ltd. under license.
MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.

Más contenido relacionado

La actualidad más candente

Emc isilon technical deep dive workshop
Emc isilon technical deep dive workshopEmc isilon technical deep dive workshop
Emc isilon technical deep dive workshopsolarisyougood
 
Emc ecs 2 technical deep dive workshop
Emc ecs 2 technical deep dive workshopEmc ecs 2 technical deep dive workshop
Emc ecs 2 technical deep dive workshopsolarisyougood
 
Enterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologyEnterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologysolarisyougood
 
Accelerate Innovation in Your Business with HP
Accelerate Innovation in Your Business with HPAccelerate Innovation in Your Business with HP
Accelerate Innovation in Your Business with HPSpiceworks Ziff Davis
 
IBM Power 7
IBM Power 7IBM Power 7
IBM Power 7None
 
32960 lar visit 022713v2
32960 lar visit 022713v232960 lar visit 022713v2
32960 lar visit 022713v2gmazuel
 
EMC: VNX Unified Storage series
EMC: VNX Unified Storage seriesEMC: VNX Unified Storage series
EMC: VNX Unified Storage seriesASBIS SK
 
EMC Vipr srm-technical Deep dive
EMC Vipr srm-technical Deep diveEMC Vipr srm-technical Deep dive
EMC Vipr srm-technical Deep divesolarisyougood
 
Announcement Overview 4Q14 (ext)
Announcement Overview 4Q14 (ext)Announcement Overview 4Q14 (ext)
Announcement Overview 4Q14 (ext)David Spurway
 
Ibm power sales bootcamp
Ibm power sales bootcampIbm power sales bootcamp
Ibm power sales bootcampsolarisyougood
 
EMC Vnx master-presentation
EMC Vnx master-presentationEMC Vnx master-presentation
EMC Vnx master-presentationsolarisyougood
 
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...Filipe Miranda
 
Emc vi pr controller tecnical customer presentation
Emc vi pr controller tecnical customer presentationEmc vi pr controller tecnical customer presentation
Emc vi pr controller tecnical customer presentationxKinAnx
 
Student guide power systems for aix - virtualization i implementing virtual...
Student guide   power systems for aix - virtualization i implementing virtual...Student guide   power systems for aix - virtualization i implementing virtual...
Student guide power systems for aix - virtualization i implementing virtual...solarisyougood
 
POWER9 AC922 Newell System - HPC & AI
POWER9 AC922 Newell System - HPC & AI POWER9 AC922 Newell System - HPC & AI
POWER9 AC922 Newell System - HPC & AI Anand Haridass
 

La actualidad más candente (20)

Emc isilon technical deep dive workshop
Emc isilon technical deep dive workshopEmc isilon technical deep dive workshop
Emc isilon technical deep dive workshop
 
Emc ecs 2 technical deep dive workshop
Emc ecs 2 technical deep dive workshopEmc ecs 2 technical deep dive workshop
Emc ecs 2 technical deep dive workshop
 
Emc vipr srm workshop
Emc vipr srm workshopEmc vipr srm workshop
Emc vipr srm workshop
 
Enterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technologyEnterprise power systems transition to power7 technology
Enterprise power systems transition to power7 technology
 
Emc vplex deep dive
Emc vplex deep diveEmc vplex deep dive
Emc vplex deep dive
 
Accelerate Innovation in Your Business with HP
Accelerate Innovation in Your Business with HPAccelerate Innovation in Your Business with HP
Accelerate Innovation in Your Business with HP
 
IBM Power 7
IBM Power 7IBM Power 7
IBM Power 7
 
32960 lar visit 022713v2
32960 lar visit 022713v232960 lar visit 022713v2
32960 lar visit 022713v2
 
EMC: VNX Unified Storage series
EMC: VNX Unified Storage seriesEMC: VNX Unified Storage series
EMC: VNX Unified Storage series
 
EMC Vipr srm-technical Deep dive
EMC Vipr srm-technical Deep diveEMC Vipr srm-technical Deep dive
EMC Vipr srm-technical Deep dive
 
Announcement Overview 4Q14 (ext)
Announcement Overview 4Q14 (ext)Announcement Overview 4Q14 (ext)
Announcement Overview 4Q14 (ext)
 
Ibm power sales bootcamp
Ibm power sales bootcampIbm power sales bootcamp
Ibm power sales bootcamp
 
EMC Vnx master-presentation
EMC Vnx master-presentationEMC Vnx master-presentation
EMC Vnx master-presentation
 
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...
New Generation of IBM Power Systems Delivering value with Red Hat Enterprise ...
 
IBM PureSystems
IBM PureSystemsIBM PureSystems
IBM PureSystems
 
Emc vi pr controller tecnical customer presentation
Emc vi pr controller tecnical customer presentationEmc vi pr controller tecnical customer presentation
Emc vi pr controller tecnical customer presentation
 
Student guide power systems for aix - virtualization i implementing virtual...
Student guide   power systems for aix - virtualization i implementing virtual...Student guide   power systems for aix - virtualization i implementing virtual...
Student guide power systems for aix - virtualization i implementing virtual...
 
POWER9 AC922 Newell System - HPC & AI
POWER9 AC922 Newell System - HPC & AI POWER9 AC922 Newell System - HPC & AI
POWER9 AC922 Newell System - HPC & AI
 
Building the Case for System z Linux
Building the Case for System z LinuxBuilding the Case for System z Linux
Building the Case for System z Linux
 
IBM PureFlex System
IBM PureFlex SystemIBM PureFlex System
IBM PureFlex System
 

Similar a Datacenter Dynamics Chicago 30 sept 2010

Moscow conference keynote in 2012
Moscow conference keynote in 2012Moscow conference keynote in 2012
Moscow conference keynote in 2012Dileep Bhandarkar
 
Oracle Database Consolidation with FlexPod on Cisco UCS
Oracle Database Consolidation with FlexPod on Cisco UCSOracle Database Consolidation with FlexPod on Cisco UCS
Oracle Database Consolidation with FlexPod on Cisco UCSNetApp
 
IBM Power Systems at FIS InFocus 2019
IBM Power Systems at FIS InFocus 2019IBM Power Systems at FIS InFocus 2019
IBM Power Systems at FIS InFocus 2019Paula Koziol
 
NEC’s Smart Enterprise Solutions - Did You Know That…
NEC’s Smart Enterprise Solutions - Did You Know That…NEC’s Smart Enterprise Solutions - Did You Know That…
NEC’s Smart Enterprise Solutions - Did You Know That…InteractiveNEC
 
Virtualización en la Red del Data Center - Extreme Networks
Virtualización en la Red del Data Center - Extreme NetworksVirtualización en la Red del Data Center - Extreme Networks
Virtualización en la Red del Data Center - Extreme NetworksAEC Networks
 
Implementing AI: High Performance Architectures: Solving Core Recommendation ...
Implementing AI: High Performance Architectures: Solving Core Recommendation ...Implementing AI: High Performance Architectures: Solving Core Recommendation ...
Implementing AI: High Performance Architectures: Solving Core Recommendation ...KTN
 
Virtualization for efficiency: by Kathrin Winkler, The green grid
Virtualization for efficiency: by Kathrin Winkler, The green gridVirtualization for efficiency: by Kathrin Winkler, The green grid
Virtualization for efficiency: by Kathrin Winkler, The green gridDCC Mission Critical
 
Desktop Private Cloud
Desktop Private CloudDesktop Private Cloud
Desktop Private CloudPaul Morse
 
MIG 5th Data Centre Summit 2016 PTS Presentation v1
MIG 5th Data Centre Summit 2016 PTS Presentation v1MIG 5th Data Centre Summit 2016 PTS Presentation v1
MIG 5th Data Centre Summit 2016 PTS Presentation v1blewington
 
Superior Cloud Economics with Power Systems
Superior Cloud Economics with Power Systems Superior Cloud Economics with Power Systems
Superior Cloud Economics with Power Systems IBM Power Systems
 
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPC
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPCExceeding the Limits of Air Cooling to Unlock Greater Potential in HPC
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPCinside-BigData.com
 
Architecting for Hyper-Scale Datacenter Efficiency
Architecting for Hyper-Scale Datacenter EfficiencyArchitecting for Hyper-Scale Datacenter Efficiency
Architecting for Hyper-Scale Datacenter EfficiencyIntel IT Center
 
Interchangeable skids provide plug and-play capabilities, improve uptime
Interchangeable skids provide plug and-play capabilities, improve uptimeInterchangeable skids provide plug and-play capabilities, improve uptime
Interchangeable skids provide plug and-play capabilities, improve uptimeIntelligentManufacturingInstitute
 
Accelerating Cloud Services - Intel
Accelerating Cloud Services - IntelAccelerating Cloud Services - Intel
Accelerating Cloud Services - IntelAmazon Web Services
 
Presentation to dm as november 2007 with dynamic provisioning information
Presentation to dm as   november 2007 with dynamic provisioning informationPresentation to dm as   november 2007 with dynamic provisioning information
Presentation to dm as november 2007 with dynamic provisioning informationxKinAnx
 
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...Salman Shaikh ヅ
 
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...VMworld
 
CtrlS: Cloud Solutions for Retail & eCommerce
CtrlS: Cloud Solutions for Retail & eCommerceCtrlS: Cloud Solutions for Retail & eCommerce
CtrlS: Cloud Solutions for Retail & eCommerceeTailing India
 
Cisco MDS Main Session EMC World 2015
Cisco MDS Main Session EMC World 2015Cisco MDS Main Session EMC World 2015
Cisco MDS Main Session EMC World 2015ldangelo0772
 

Similar a Datacenter Dynamics Chicago 30 sept 2010 (20)

Moscow conference keynote
Moscow conference keynoteMoscow conference keynote
Moscow conference keynote
 
Moscow conference keynote in 2012
Moscow conference keynote in 2012Moscow conference keynote in 2012
Moscow conference keynote in 2012
 
Oracle Database Consolidation with FlexPod on Cisco UCS
Oracle Database Consolidation with FlexPod on Cisco UCSOracle Database Consolidation with FlexPod on Cisco UCS
Oracle Database Consolidation with FlexPod on Cisco UCS
 
IBM Power Systems at FIS InFocus 2019
IBM Power Systems at FIS InFocus 2019IBM Power Systems at FIS InFocus 2019
IBM Power Systems at FIS InFocus 2019
 
NEC’s Smart Enterprise Solutions - Did You Know That…
NEC’s Smart Enterprise Solutions - Did You Know That…NEC’s Smart Enterprise Solutions - Did You Know That…
NEC’s Smart Enterprise Solutions - Did You Know That…
 
Virtualización en la Red del Data Center - Extreme Networks
Virtualización en la Red del Data Center - Extreme NetworksVirtualización en la Red del Data Center - Extreme Networks
Virtualización en la Red del Data Center - Extreme Networks
 
Implementing AI: High Performance Architectures: Solving Core Recommendation ...
Implementing AI: High Performance Architectures: Solving Core Recommendation ...Implementing AI: High Performance Architectures: Solving Core Recommendation ...
Implementing AI: High Performance Architectures: Solving Core Recommendation ...
 
Virtualization for efficiency: by Kathrin Winkler, The green grid
Virtualization for efficiency: by Kathrin Winkler, The green gridVirtualization for efficiency: by Kathrin Winkler, The green grid
Virtualization for efficiency: by Kathrin Winkler, The green grid
 
Desktop Private Cloud
Desktop Private CloudDesktop Private Cloud
Desktop Private Cloud
 
MIG 5th Data Centre Summit 2016 PTS Presentation v1
MIG 5th Data Centre Summit 2016 PTS Presentation v1MIG 5th Data Centre Summit 2016 PTS Presentation v1
MIG 5th Data Centre Summit 2016 PTS Presentation v1
 
Superior Cloud Economics with Power Systems
Superior Cloud Economics with Power Systems Superior Cloud Economics with Power Systems
Superior Cloud Economics with Power Systems
 
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPC
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPCExceeding the Limits of Air Cooling to Unlock Greater Potential in HPC
Exceeding the Limits of Air Cooling to Unlock Greater Potential in HPC
 
Architecting for Hyper-Scale Datacenter Efficiency
Architecting for Hyper-Scale Datacenter EfficiencyArchitecting for Hyper-Scale Datacenter Efficiency
Architecting for Hyper-Scale Datacenter Efficiency
 
Interchangeable skids provide plug and-play capabilities, improve uptime
Interchangeable skids provide plug and-play capabilities, improve uptimeInterchangeable skids provide plug and-play capabilities, improve uptime
Interchangeable skids provide plug and-play capabilities, improve uptime
 
Accelerating Cloud Services - Intel
Accelerating Cloud Services - IntelAccelerating Cloud Services - Intel
Accelerating Cloud Services - Intel
 
Presentation to dm as november 2007 with dynamic provisioning information
Presentation to dm as   november 2007 with dynamic provisioning informationPresentation to dm as   november 2007 with dynamic provisioning information
Presentation to dm as november 2007 with dynamic provisioning information
 
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...
Ciscounifiedcomputingsystemucschangingtheeconomicsdatacenter 130514165541-php...
 
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...
VMworld 2013: US Air National Guard - DoD Private Cloud Initiative –How Virtu...
 
CtrlS: Cloud Solutions for Retail & eCommerce
CtrlS: Cloud Solutions for Retail & eCommerceCtrlS: Cloud Solutions for Retail & eCommerce
CtrlS: Cloud Solutions for Retail & eCommerce
 
Cisco MDS Main Session EMC World 2015
Cisco MDS Main Session EMC World 2015Cisco MDS Main Session EMC World 2015
Cisco MDS Main Session EMC World 2015
 

Más de Dileep Bhandarkar

Open Compute Summit Keynote 17 June 2011
Open Compute Summit Keynote 17 June 2011Open Compute Summit Keynote 17 June 2011
Open Compute Summit Keynote 17 June 2011Dileep Bhandarkar
 
Energy Efficiency Considerations in Large Datacenters
Energy Efficiency Considerations in Large DatacentersEnergy Efficiency Considerations in Large Datacenters
Energy Efficiency Considerations in Large DatacentersDileep Bhandarkar
 
Data center-server-cooling-power-management-paper
Data center-server-cooling-power-management-paperData center-server-cooling-power-management-paper
Data center-server-cooling-power-management-paperDileep Bhandarkar
 
New Delhi Cloud Summit 05 26-11
New Delhi Cloud Summit 05 26-11New Delhi Cloud Summit 05 26-11
New Delhi Cloud Summit 05 26-11Dileep Bhandarkar
 
Performance Characterization of the Pentium Pro Processor
Performance Characterization of the Pentium Pro ProcessorPerformance Characterization of the Pentium Pro Processor
Performance Characterization of the Pentium Pro ProcessorDileep Bhandarkar
 
Innovation lecture for hong kong
Innovation lecture for hong kongInnovation lecture for hong kong
Innovation lecture for hong kongDileep Bhandarkar
 
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...Dileep Bhandarkar
 
Qualcomm centriq 2400 hot chips final submission corrected
Qualcomm centriq 2400 hot chips final submission correctedQualcomm centriq 2400 hot chips final submission corrected
Qualcomm centriq 2400 hot chips final submission correctedDileep Bhandarkar
 
Innovation lecture for shanghai final
Innovation lecture for shanghai finalInnovation lecture for shanghai final
Innovation lecture for shanghai finalDileep Bhandarkar
 
Linaro connect 2018 keynote final updated
Linaro connect 2018 keynote final updatedLinaro connect 2018 keynote final updated
Linaro connect 2018 keynote final updatedDileep Bhandarkar
 
Dileep Random Access Talk at salishan 2016
Dileep Random Access Talk at salishan 2016Dileep Random Access Talk at salishan 2016
Dileep Random Access Talk at salishan 2016Dileep Bhandarkar
 
Server design summit keynote handout
Server design summit keynote handoutServer design summit keynote handout
Server design summit keynote handoutDileep Bhandarkar
 
My Feb 2003 HPCA9 Keynote Slides - Billion Transistor Processor Chips
My Feb 2003  HPCA9 Keynote Slides - Billion Transistor Processor ChipsMy Feb 2003  HPCA9 Keynote Slides - Billion Transistor Processor Chips
My Feb 2003 HPCA9 Keynote Slides - Billion Transistor Processor ChipsDileep Bhandarkar
 

Más de Dileep Bhandarkar (20)

Open Compute Summit Keynote 17 June 2011
Open Compute Summit Keynote 17 June 2011Open Compute Summit Keynote 17 June 2011
Open Compute Summit Keynote 17 June 2011
 
Energy Efficiency Considerations in Large Datacenters
Energy Efficiency Considerations in Large DatacentersEnergy Efficiency Considerations in Large Datacenters
Energy Efficiency Considerations in Large Datacenters
 
Samsung cio-forum-2012
Samsung cio-forum-2012 Samsung cio-forum-2012
Samsung cio-forum-2012
 
Data center-server-cooling-power-management-paper
Data center-server-cooling-power-management-paperData center-server-cooling-power-management-paper
Data center-server-cooling-power-management-paper
 
New Delhi Cloud Summit 05 26-11
New Delhi Cloud Summit 05 26-11New Delhi Cloud Summit 05 26-11
New Delhi Cloud Summit 05 26-11
 
Performance Characterization of the Pentium Pro Processor
Performance Characterization of the Pentium Pro ProcessorPerformance Characterization of the Pentium Pro Processor
Performance Characterization of the Pentium Pro Processor
 
Innovation lecture for hong kong
Innovation lecture for hong kongInnovation lecture for hong kong
Innovation lecture for hong kong
 
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...
Performance from Architecture: Comparing a RISC and a CISC with Similar Hardw...
 
Qualcomm centriq 2400 hot chips final submission corrected
Qualcomm centriq 2400 hot chips final submission correctedQualcomm centriq 2400 hot chips final submission corrected
Qualcomm centriq 2400 hot chips final submission corrected
 
Innovation lecture for shanghai final
Innovation lecture for shanghai finalInnovation lecture for shanghai final
Innovation lecture for shanghai final
 
Semicon2018 dileepb
Semicon2018 dileepbSemicon2018 dileepb
Semicon2018 dileepb
 
Linaro connect 2018 keynote final updated
Linaro connect 2018 keynote final updatedLinaro connect 2018 keynote final updated
Linaro connect 2018 keynote final updated
 
Hipeac 2018 keynote Talk
Hipeac 2018 keynote TalkHipeac 2018 keynote Talk
Hipeac 2018 keynote Talk
 
Alpha memo july 1992
Alpha memo july 1992Alpha memo july 1992
Alpha memo july 1992
 
China AI Summit talk 2017
China AI Summit talk 2017China AI Summit talk 2017
China AI Summit talk 2017
 
Dileep Random Access Talk at salishan 2016
Dileep Random Access Talk at salishan 2016Dileep Random Access Talk at salishan 2016
Dileep Random Access Talk at salishan 2016
 
Risc vs cisc
Risc vs ciscRisc vs cisc
Risc vs cisc
 
Server design summit keynote handout
Server design summit keynote handoutServer design summit keynote handout
Server design summit keynote handout
 
DileepB EDPS talk 2015
DileepB  EDPS talk 2015DileepB  EDPS talk 2015
DileepB EDPS talk 2015
 
My Feb 2003 HPCA9 Keynote Slides - Billion Transistor Processor Chips
My Feb 2003  HPCA9 Keynote Slides - Billion Transistor Processor ChipsMy Feb 2003  HPCA9 Keynote Slides - Billion Transistor Processor Chips
My Feb 2003 HPCA9 Keynote Slides - Billion Transistor Processor Chips
 

Último

From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .Alan Dix
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxhariprasad279825
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsMiki Katsuragi
 
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo Day
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo DayH2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo Day
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo DaySri Ambati
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity PlanDatabarracks
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Mattias Andersson
 
Search Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfSearch Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfRankYa
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsRizwan Syed
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningLars Bell
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostZilliz
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brandgvaughan
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLScyllaDB
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfAlex Barbosa Coqueiro
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.Curtis Poe
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii SoldatenkoFwdays
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionDilum Bandara
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Commit University
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Scott Keck-Warren
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Enterprise Knowledge
 

Último (20)

From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .From Family Reminiscence to Scholarly Archive .
From Family Reminiscence to Scholarly Archive .
 
Artificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptxArtificial intelligence in cctv survelliance.pptx
Artificial intelligence in cctv survelliance.pptx
 
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data PrivacyTrustArc Webinar - How to Build Consumer Trust Through Data Privacy
TrustArc Webinar - How to Build Consumer Trust Through Data Privacy
 
Vertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering TipsVertex AI Gemini Prompt Engineering Tips
Vertex AI Gemini Prompt Engineering Tips
 
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo Day
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo DayH2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo Day
H2O.ai CEO/Founder: Sri Ambati Keynote at Wells Fargo Day
 
How to write a Business Continuity Plan
How to write a Business Continuity PlanHow to write a Business Continuity Plan
How to write a Business Continuity Plan
 
Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?Are Multi-Cloud and Serverless Good or Bad?
Are Multi-Cloud and Serverless Good or Bad?
 
Search Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdfSearch Engine Optimization SEO PDF for 2024.pdf
Search Engine Optimization SEO PDF for 2024.pdf
 
Scanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL CertsScanning the Internet for External Cloud Exposures via SSL Certs
Scanning the Internet for External Cloud Exposures via SSL Certs
 
DSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine TuningDSPy a system for AI to Write Prompts and Do Fine Tuning
DSPy a system for AI to Write Prompts and Do Fine Tuning
 
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage CostLeverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
Leverage Zilliz Serverless - Up to 50X Saving for Your Vector Storage Cost
 
WordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your BrandWordPress Websites for Engineers: Elevate Your Brand
WordPress Websites for Engineers: Elevate Your Brand
 
Developer Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQLDeveloper Data Modeling Mistakes: From Postgres to NoSQL
Developer Data Modeling Mistakes: From Postgres to NoSQL
 
Unraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdfUnraveling Multimodality with Large Language Models.pdf
Unraveling Multimodality with Large Language Models.pdf
 
How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.How AI, OpenAI, and ChatGPT impact business and software.
How AI, OpenAI, and ChatGPT impact business and software.
 
"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko"Debugging python applications inside k8s environment", Andrii Soldatenko
"Debugging python applications inside k8s environment", Andrii Soldatenko
 
Advanced Computer Architecture – An Introduction
Advanced Computer Architecture – An IntroductionAdvanced Computer Architecture – An Introduction
Advanced Computer Architecture – An Introduction
 
Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!Nell’iperspazio con Rocket: il Framework Web di Rust!
Nell’iperspazio con Rocket: il Framework Web di Rust!
 
Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024Advanced Test Driven-Development @ php[tek] 2024
Advanced Test Driven-Development @ php[tek] 2024
 
Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024Designing IA for AI - Information Architecture Conference 2024
Designing IA for AI - Information Architecture Conference 2024
 

Datacenter Dynamics Chicago 30 sept 2010

  • 1. ©Microsoft® 2009 Performance/ Watt/Dollar Dileep Bhandarkar, Ph. D. Distinguished Engineer Global Foundation Services Microsoft Corporation Energy Efficiency in Datacenters 2010
  • 2. ©Microsoft® 2009 CLOUD SERVICES 59 markets and 36 languages 76 markets and 48 languages Our Consumer Cloud Services
  • 4. ©Microsoft® 2009 Service Key Metric More than 1 billion authentications per day ~ 400 million active accounts Processes 3 – 5 billion e-mail messages a day More than 4 billion queries per month 10B messages each day/ 750M active accounts Over 450 million unique visitors monthly Key Operational Metrics
  • 5. ©Microsoft® 2009 Datacenter Construction Costs Land: 0%-2% Core & Shell Costs: 5%-9% Architectural: 4%-7% Mechanical / Electrical: 70%-85% Land Core / Shell Mech /Elec Arch Where the costs are: >80% scale with power <10% scale with space Reduce!
  • 6. ©Microsoft® 2009 Why Power is Important? Energy Consumption: US power rate 10.27 cents per Kilowatt hour) in 2008 according to DOE/eia (http://www.eia.doe.gov/cneaf/electricity/epm/table5_3.html) In a typical data center for every watt in server power there can be another 0.5 to 1 watt consumed for power distribution losses and cooling. Power (Switch Gear, UPS, Battery backup, etc) Cooling (Chillers, CRACs, etc) Building load Demand from grid IT load Demand from servers, storage, telco equipment, etc PUE = Total Facility Power/IT Equipment Power
  • 7. ©Microsoft® 2009 But There is More! Datacenters can cost between $10M and $20M per megawatt Capital depreciation costs are important Server acquisition costs may not dominate TCO Basic 1U Server - 5 year TCO Total Equipment cost per server Energy Usage cost per server Datacenter Capital cost per server Datacenter operating cost per server
  • 8. ©Microsoft® 2009 Beyond Just Reducing PUE PUE is an excellent metric for evaluating effectiveness of datacenter infrastructure Server performance per watt can vary Rightsizing of server configuration can improve overall energy efficiency Low Power processors often offer best performance per watt Microsoft optimizes the total solution for best performance per watt per TCO$ 8
  • 9. ©Microsoft® 2009 Holistic Optimization Datacenter and Servers together PUE is an excellent metric for evaluating effectiveness of datacenter infrastructure Server performance per watt can vary Rightsizing of server configuration can improve overall energy efficiency Low Power processors often offer best performance per watt Microsoft optimizes the total solution for best performance per TCO$ 9
  • 10. Drive change in the industry through strong partnerships
  • 11.
  • 12. 12 2 Socket server based on a half- width server design •6.3” x 16.7” •2 CPU Sockets •4 DIMM Slots per CPU Socket •4 SATA HDD Connections • Dual 1 GbE NIC •One 16x PCIe Card Slot 16x PCIe Slot 4x SATA HDD Connectors Optional 16x PCIe Card, not present in basic configuration 2x CPU Memory DIMMs
  • 13. 13 2 to 4 TOR (2 to 4 RU) 96 Servers (Up to 48 RU) 2 Bulk Power (2x 3RU) Upper Power Domain Lower Power Domain
  • 14.
  • 15.
  • 16. 1.0 1.07 1.18 1.0 1.10 1.6 1.0 1.24 1.56 1.0 0.87 0.76 1.0 0.79 0.46 2.33 GHz (50W) 2.67 GHz (80W) 3.167 GHz (120W) Performance System Price System Power Perf/W Perf/W/$ Assumption: Server Price = $2000 + CPUs, Server Power = 150W + CPUs The sweet spot is often at low power processors, especially when system price and power are considered. Your mileage may vary!
  • 17. ChicagoQuincy Dublin Amsterdam Hong Kong Singapore Japan "Datacenters havebecomeasvitaltothefunctioning ofsocietyaspowerstations."--TheEconomist San Antonio Multiple global CDN locations in the Americas, Europe and APAC. Microsoft has more than 10 and less than 100 DCs worldwide. Quincy, Washington: approx 500K sq ft, 27MW, uses entirely hydro-electric power San Antonio, Texas: approx 477K sq ft, 27MW, uses recycled water for cooling Chicago, Illinois: 707,000 square feet with critical power of 60 MW, uses water side economization, containers Dublin, Ireland: approx 570K sq ft, up to 27MW, uses outside air for cooling.
  • 18. Microsoft’s Data Center Evolution Containers Scalability Sustainability Data Center Colocation Generation 1 Generation 4 Modular Data Center Server Capacity Rack Density and Deployment Quincy and San Antonio Generation 2 Chicago and Dublin Generation 3 Deployment Scale Unit IT PAC Time to Market Lower TCO Scalable Data Center
  • 19. ©Microsoft® 2009 Chicago Datacenter Two-story datacenter First floor: container bay medium reliability Second floor: high- reliability traditional co-location rooms Can deploy at scale Optimized energy and cooling efficiency Allows OEMs to build customized solutions based on Microsoft specifications
  • 20. ©Microsoft® 2009 Container Selection Process “CBlox” specification defined all of the interface requirements to the Data Center: Physical Dimensions – length, width, height, weight Chilled Water temperature & flow rate Power – voltage and wattage Selection based primarily on: $/Server # of Servers/container
  • 23. ©Microsoft® 2009 Modularize and pre-manufacture the entire datacenter Lower the Total Cost of Ownership Increase scalability and Right Time to Market Standardize components to improve operations and reliability Paradigm Shift to Ultra Modularity
  • 24. ©Microsoft® 2009 Microsoft’s Future Datacenters • No mechanical cooling • Ultra-efficient water utilization • Utilize more renewable building materials • Reduce time to market by >50% • Low initial capital investment • Scale with business demand • 30 % - 50 % more cost effective for the same class of service • Outstanding PUE
  • 25. ©Microsoft® 2009 Gen 4 Datacenter Strategy Standardized Gen, UPS, IT compute PACs and support equipment Multiple Datacenter Classes, Colo and Modular PAC’s (Pre-Assembled Components) Support equipment designed to be installed on indoor skids or outdoor PAC’s Primary Cooling via Airside Economizers Secondary Cooling via Air Cooled or Water Cooled Chillers, or Fluid Cooler Warmest server intake temperatures compared to all other Generations Colo - No Raised Floor with Hot/Cold Aisle containment
  • 26. ©Microsoft® 2009 ITPAC Design 26 video http://www.microsoft.com/showcase/en/us/details/84f44749-1343-4467-8012-9c70ef77981c 20 foot, 170 KW, 9 rack Proof of Concept ITPAC CE and ETL certified as a Air Handler Operating outside for over 1 year
  • 28. ©Microsoft® 2009 Conclusions Holistic optimization of server and datacenter design Platform energy efficiency remains a major focus area (Perf/Watt/Dollar) Stranding Power is expensive Focus on Infrastructure Cost reduction via high temperature and free air cooling Workload analysis is critical to rightsizing Total Cost of Ownership is important Measure and Control Server Parameters Conventional Wisdom is often wrong! 28
  • 30. © 2008 Microsoft Corporation. All rights reserved. Microsoft, Windows, Windows Vista and other product names are or may be registered trademarks and/or trademarks in the U.S. and/or other countries. The information herein is for informational purposes only and represents the current view of Microsoft Corporation as of the date of this presentation. Because Microsoft must respond to changing market conditions, it should not be interpreted to be a commitment on the part of Microsoft, and Microsoft cannot guarantee the accuracy of any information provided after the date of this presentation. Microsoft is a trademark of the Microsoft companies. The Microsoft Financing marks are used by CIT Financial Ltd. under license. MICROSOFT MAKES NO WARRANTIES, EXPRESS, IMPLIED OR STATUTORY, AS TO THE INFORMATION IN THIS PRESENTATION.