More Related Content Similar to Dc design (20) Dc design1. Data Center Design for
the 21st Century
Best Practices on Operational Efficiency from the IBM Data Center Study
The IT Services Site Webinar Sponsored by
3. Today’s Featured Speakers
Virginia Agee
IDC Research Director
Data Center Trends & Strategies
David Aubuchon
IBM Data Center Services Executive
4. Virginia Agee, IDC Research Director - Data Center Trends & Strategies
David Aubuchon, IBM Data Center Services Executive
May 22, 2012
Data Center Design for the 21st Century:
Best Practices on Operational Efficiency from the IBM Data Center Study
© 2012 IBM Corporation
5. The study addressed server, storage, network, facility and data
center operations to learn what is being implemented today
It represents 308 clients in 7 countries and 25 industries
Geography Sector Organization Size
Other Distribution >10,000 0 – 999
Asia Pacific Europe 18% 12%
29% employees employees
30% 13% 28%
Comms
10% Industrial
26%
Latin Public Financial
America North America 21% Services 1,000 – 9,999 employees
10% 31% 13% 59%
58% plan significant data center investments in next 2 years
68% are the first to adopt or rapid technology adopters
© 2012 IBM Corporation
6. Poll Question #1
What is the main industry for your company?
A. Distribution
B. Industrial
C. Financial Services
D. Public
E. Communications
F. Other
7. One in five, or 21%, of clients have highly efficient data centers—
and are allocating 50% more of their IT budget to new projects
How much of your IT budget Improving data center operational
is spent on: efficiency allows spending to go to:
Maintain existing
infrastructure • Improving public safety or
New projects providing more citizen services
Basic 35% • Transforming core banking and
data payment systems
centers
65%
• Implementing smart grids or
improving generation performance
New projects • Allocating funds to other business
Strategic priorities
data 53%
centers
47%
© 2012 IBM Corporation
8. Characteristics of highly efficient data centers
Optimize the IT and
facilities assets to maximize
capacity and availability
Design for flexibility to
support changing business
needs
Use automation tools to
improve service levels and
availability
Have a plan that aligns with
the business goals and
keep it current
© 2012 IBM Corporation
9. Design implications from highly efficient data centers
Right size capacity and
availability to get the best
return on investment
Design for flexibility to
support changing business
needs
Optimize long-term data
center operational
expenses, not just up-front
capital costs
© 2012 IBM Corporation
10. Highly efficient data centers right size by matching business needs to
current and future capability
Right Size Data Center Facilities
77%
41%
33%
25% 22%
14%
0% 2%
Forecast space Forecast Implement high and low Have fully redundant
(10-20 years) power demand density zones systems
Strategic Basic
10 © 2012 IBM Corporation
11. Right Size
Rightsizing can yield significant benefits. An IBM data center in
Lexington gained 8x IT capacity and avoided $50M capital cost
8x growth in
IT capacity
2x growth in
number of
10x servers
increase Compute Visibility,
Establish in asset Resources
baseline Automation,
utilization Control
energy use
20x more Consolidate 15-20% power savings
servers than and Virtualize Best from actual usage
needed Practices
Diagnose Right size infrastructure to
Improved virtualized IT workloads
Operations
10% improvement from
less over cooling
Use analytic tools and software monitoring to right-size the facility
infrastructure after server virtualization projects
© 2012 IBM Corporation
12. Right Size
Analytical tools can be used to right-size capacity providing a solution
tailored to meet your business growth
Banking and finance Business unit of
Distribution company in
customer in volatile hospitality organization
stable agricultural market
market in low-growth market
Business
plan 675 percent power increase
230 percent space increase
150 percent power increase
No change in power
Capacity No change in space
60 percent space decrease
forecast
1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10
Year Year Year
Design IBM created a modular IBM recommended IBM provided metrics for
expansion plan for the data retaining their existing a staged release of physical
direction center by linking capital and data center with power space, aiding lease
operational expenses to and cooling upgrades, negotiations and saving
business growth. reducing capital expense US$400,000 per year.
by US$13 million.
* Based on IBM client engagements; individual results can vary. © 2012 IBM Corporation
13. Right Size
Analysis shows the impact of your technology lifecycle providing
improved plans to maximize capacity and flexibility for change
Today 2020
50% of server density is < 5 kw/Rack or less 60% of server density is > 10-20 kw/rack
45%
40%
60%
35%
50%
30%
40% 25%
30% 20% h
h
20% 15%
10%
10%
5%
0%
0%
0 5 10 15 20 25 30 35
0 5 10 15 20 25 30 35
Kilowatts/rack
Kilowatts/rack
Predicted density over a period of time is a pre-requisite to better design
Analytics can create a density requirement model to plan for changes over time
Support unknown capacity: design for 3-5x power density over a 10-year time
Designing high / low density zones saved over $1 Million per year
Based on IBM estimates from client engagements
© 2012 IBM Corporation
14. Right Size
Availability analysis can be used to identify the cost and impact
trade-offs for your organization to frame a complex decision
Cost Value Optimal solution
Generate data center Forecast for the business Determine the most
designs alternatives benefit of each alternative efficient design
Generator redundancy1
US$980K
3.618 3.864 3.865 "2N"
US$765K
US$498K "N+1"
"N"
US$1M US$400M
"N" "N+1" "2N" "N" "N+1" "2N"
Additional annual expenses2 Reduction in annual outages2 Cost per outage to justify
additional expenses2
1Other applications include cooling system redundancy, electrical system distribution, UPS technology, etc …
2Based on IBM Proprietary Analytics using expected lifespan of equipment, data center vulnerabilities and operational quality
© 2012 IBM Corporation
15. Highly efficient data centers recognize that change occurs naturally and
have designed more flexibility into their approach
Design for Flexibility
86%
68%
53%
48% 47%
43%
27%
9%
First and fast technology Percentage of servers Build in smaller Upgrade Mechanical /
adoption virtualized increments Electrical equipment
without disruption
Strategic Basic
15 © 2012 IBM Corporation
16. Poll Question #2
What percent of virtualization have you achieved for your servers
A. > 60%
B. 40% - 60%
C. 20% - 40%
D. 10% - 20%
E. < 10%
17. Flexible Design
Capital and operational cost analysis shows data centers with on
demand capability is a practical and cost effective solution
Data center capital costs Facility operating costs
60% costs from mechanical / 3-5X capital costs; energy can be 75%
electrical systems $250
Energy Cost
Cumulative Cost of Operations ($Millions)
Staffing
Mechanical Bldg. Maint. & Mgmt.
$200
20% R E Tax
Instrumentation
& Controls $150
Fees
4%
24%
$100
Power
$50
36%
Fit-Up
0
Costs 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
9% Shell Year
7%
SCALABLE MODULAR DATA CENTER ENTERPRISE MODULAR DATA CENTER PORTABLE MODULAR DATA CENTER HIGH DENSITY ZONE
© 2012 IBM Corporation
18. Flexible Design
Data centers with a long and extended life without disruptive retrofits
despite changing demands from technology is becoming a reality
Capital Costs
Good design is an insurance Provide up to 3x the power density
policy that provide flexibility. with lower up front capital costs
5-10% up front capital costs
$M Pay now Pay later
versus paying ~50% higher
costs plus a disruptive 30
$25M
retrofit later. 25
20 $17M
$15M $14M
• Provision for potential water 15
cooling 10
• Modular UPS design and chiller 5
plant design
0
• Electrical switchgear sized to grow 1x 3x
Power Density Growth (# times)
Incremental costs
© 2012 IBM Corporation
19. Highly efficient data center optimize capital and operating costs over the
10-20 year useful life of the facility
Optimize Capital and Operating Costs
76% 79%
53% 49%
31%
17%
6%
0%
Evaluate capital and Measure energy Follow ASHRAE industry Monitor thermal
operating costs efficiency (PUE) standards conditions
Strategic Basic
19 © 2012 IBM Corporation
20. Poll Question #3
What is your primary data center's Power Usage Effectiveness (PUE)
rating?
A. >3.5
B. 2.5 to 3.5
C. 1.5 to 2.5
D. 1.0 to 1.5
E. We don't measure PUE
21. Optimize Costs
A singular focus on capital costs can lead to sub-optimized data
centers
Capital Cost Analysis Operating Cost Analysis
Show 5x difference in technology alternatives Show 2.5x difference for vendor alternatives
$2 $3 Technology 1, Vendor A
Technology 1, Vendor C
2.5
1.5 Technology 3, Vendor G
2
$ Millions
$ Millions
1 1.5
1
.5
.5
0 0
1 2 3 4 5 6 7 8 9 10
Technology #1 Technology #2 Technology #3
Year
Reduce data center total costs 15-30% over its useful life
Improve decisions looking at vendor and technology tradeoffs
© 2012 IBM Corporation
Source: IBM Estimates
22. Optimize Costs
Monitoring and management tools provide on-going visibility to
improve operational management and operational costs
Insight from Analytics On-going Monitoring
Power savings from improved cooling 5x improvement in PUE based on
efficiencies monitoring from start-up
PUE
6.0
Service AC Units 5.0
16.5%
4.0
3.0
Cable Cutouts Turn off
13.5% CRACs 2.0
50%
1.0
Perforated Tiles 0.0
20% Startup Month 3 Month 6 Full Load
Use real-time monitoring for day-to-day visibility
Analytics provide insight to go beyond the “low hanging” fruit
© 2012 IBM Corporation
23. ―No regrets‖ actions to improve your next data center facility
―Right-size‖ capacity and availability
Find underutilized assets and make changes
Use analytics to define capacity and availability to meet
the business needs
Use real time management systems to stay focused
continued improvement
Design for flexibility
Design any new infrastructure investments for scalability
to support changes in demand and technology
Allow for scalability in smaller increments over time
Design for additions to be non-disruptive
Plan for water cooling to support higher density servers
Optimize total costs over the long term
Measure efficiency on a real time basis
Ensure trade-offs between capital and operating costs are
included in all business cases
Implement actions that will provide short payback in all
situations
© 2012 IBM Corporation
24. A trusted partner knows one size does not fit all. They will
help you determine the right solution to meet your needs
Design / build 200-300 world-class data centers each year around the world
Experts help design an efficient, scalable and reliable solution to meet your business needs
End to end solution provider from strategy to design to move in ready
Determine Model
outcomes Design a Increase
the world-class
before automation
business data center
you
needs
implement
Map the Model space Design an Improve
business needs and capacity efficient, operational
to data center during design scalable and efficiency and
technical process reliable reduce outages
requirements solution to due to human
meet your error
business needs
© 2012 IBM Corporation
25. For more information, please visit
Register for the data center study paper: www.ibm.com/data-center/study
Data center services: www.ibm.com/services/smarterdatacenter
© 2012 IBM Corporation
26. Questions?
Please submit your questions now
Virginia Agee
IDC Research Director
Data Center Trends & Strategies
David Aubuchon
IBM Data Center Services Executive
27. Resources
To view this or other events on demand, please visit:
http://www.theitservicessite.com/webinar_archives.asp
For more information, please visit:
www.theitservicessite.com
IBM Server optimization and integration
IBM Data Center Study
http://www.ibm.com/data-center/study
IBM Smarter Data Center
http://www.ibm.com/services/smarterdatacenter