Date post: | 29-Nov-2014 |
Category: |
Technology |
Upload: | herb-hernandez |
View: | 1,638 times |
Download: | 1 times |
© 2012 IBM Corporation
David Aubuchon IBM Data Center Services Executive
July 17, 2012
© 2012 IBM Corporation
Sector Geography
It represents 308 clients in 7 countries and 25 industries
Organization Size
North America 31%
Latin America
10%
Europe 29%
Asia Pacific 30%
Public 21%
Distribution 12%
Financial Services 13%
Industrial 26%
Other 18%
Comms 10%
1,000 – 9,999 employees 59%
>10,000 employees
13%
0 – 999 employees 28%
The study addressed server, storage, network, facility and data center operations to learn what is being implemented today
58% plan significant data center investments in next 2 years
68% are the first to adopt or rapid technology adopters
© 2012 IBM Corporation
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
is spent on:
35%
65%
Basic
data
centers
Strategic
data
centers
New projects
Maintain existing
infrastructure
47%
Improving data center operational efficiency allows spending to go to:
53%
• Improving public safety or
providing more citizen services
• Transforming core banking and
payment systems
• Implementing smart grids or
improving generation performance
• Allocating funds to other business
priorities
New projects
© 2012 IBM Corporation
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
Characteristics of highly efficient data centers
© 2012 IBM Corporation
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
25%
77%
33%41%
0%
14%
2%
22%
5
Basic Strategic
Right Size Data Center Facilities
Highly efficient data centers right size by matching business needs to current and future capability
Forecast space
(10-20 years)
Forecast
power demand
Implement high and low
density zones
Have fully redundant
systems
© 2012 IBM Corporation
Rightsizing can yield significant benefits. An IBM data center in
Lexington gained 8x IT capacity and avoided $50M capital cost
Diagnose
Establish baseline energy use
20x more servers than needed
Consolidate and Virtualize
10x increase in asset utilization
Compute
Resources
8x growth in IT capacity
2x growth in number of servers
Improved Operations
10% improvement from less over cooling
Best
Practices
Right size infrastructure to virtualized IT workloads
Visibility, Automation,
Control
15-20% power savings from actual usage
Use analytic tools and software monitoring to right-size the facility infrastructure after server virtualization projects
Right Size
© 2012 IBM Corporation
Analytical tools can be used to right-size capacity providing a solution tailored to meet your business growth
Right Size
Business unit of
hospitality organization
in low-growth market
Distribution company in
stable agricultural market
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
Banking and finance
customer in volatile
market
IBM created a modular
expansion plan for the data
center by linking capital and
operational expenses to
business growth.
IBM recommended
retaining their existing
data center with power
and cooling upgrades,
reducing capital expense
by US$13 million.
IBM provided metrics for
a staged release of physical
space, aiding lease
negotiations and saving
US$400,000 per year.
* Based on IBM client engagements; individual results can vary.
675 percent power increase
230 percent space increase
150 percent power increase
No change in space No change in power
60 percent space decrease
Business
plan
Capacity
forecast
Design
direction
Year Year Year
© 2012 IBM Corporation
Analysis shows the impact of your technology lifecycle providing improved plans to maximize capacity and flexibility for change
Today 50% of server density is < 5 kw/Rack or less
2020 60% of server density is > 10-20 kw/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
0%
5%
10%
15%
20%
25%
30%
35%
40%
45%
0 5 10 15 20 25 30 35
h
Kilowatts/rack
0%
10%
20%
30%
40%
50%
60%
0 5 10 15 20 25 30 35
h
Kilowatts/rack
Based on IBM estimates from client engagements
Right Size
© 2012 IBM Corporation
Availability analysis can be used to identify the cost and impact trade-offs for your organization to frame a complex decision
Right Size
"N" "N+1" "2N" "N" "N+1" "2N"
3.618 3.864 3.865
Additional annual expenses2 Reduction in annual outages2
"N"
"N+1"
"2N"
Cost per outage to justify
additional expenses2
US$498K
US$765K
US$980K
Generator redundancy1
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
Optimal solution Determine the most
efficient design
Cost Generate data center
designs alternatives
Value Forecast for the business
benefit of each alternative
US$400M US$1M
© 2012 IBM Corporation
86%
48%
68%
47%43%
27%
53%
9%
10
Basic Strategic
Design for Flexibility
Highly efficient data centers recognize that change occurs naturally and have designed more flexibility into their approach
First and fast technology
adoption
Percentage of servers
virtualized
Build in smaller
increments
Upgrade Mechanical /
Electrical equipment
without disruption
© 2012 IBM Corporation
Capital and operational cost analysis shows data centers with on demand capability is a practical and cost effective solution
Data center capital costs 60% costs from mechanical /
electrical systems
Shell
7%
Mechanical
20%
Fees
24%
Fit-Up
Costs
9%
Instrumentation
& Controls
4%
Power
36%
Facility operating costs 3-5X capital costs; energy can be 75%
Cum
ula
tive C
ost of O
pera
tions (
$M
illio
ns)
Year
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Energy Cost
Staffing
Bldg. Maint. & Mgmt.
R E Tax
$100
$150
$200
$250
$50
0
Cum
ula
tive C
ost of O
pera
tions (
$M
illio
ns)
Year
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
Energy Cost
Staffing
Bldg. Maint. & Mgmt.
R E Tax
Energy Cost
Staffing
Bldg. Maint. & Mgmt.
R E Tax
$100
$150
$200
$250
$50
0
SCALABLE MODULAR DATA CENTER ENTERPRISE MODULAR DATA CENTER PORTABLE MODULAR DATA CENTER HIGH DENSITY ZONE
Flexible Design
© 2012 IBM Corporation
Data centers with a long and extended life without disruptive retrofits despite changing demands from technology is becoming a reality
0
5
10
15
20
25
30
1x 3x
• Provision for potential water cooling
• Modular UPS design and chiller plant design
• Electrical switchgear sized to grow
$M
Power Density Growth (# times)
Incremental costs
Capital Costs Provide up to 3x the power density
with lower up front capital costs
Pay now Pay later
Good design is an insurance
policy that provide flexibility.
5-10% up front capital costs
versus paying ~50% higher
costs plus a disruptive
retrofit later.
$17M $15M $14M
$25M
Flexible Design
© 2012 IBM Corporation
76% 79%
49%
31%
53%
6%
17%
0%
13
Basic Strategic
Optimize Capital and Operating Costs
Highly efficient data center optimize capital and operating costs over the 10-20 year useful life of the facility
Evaluate capital and
operating costs
Measure energy
efficiency (PUE)
Follow ASHRAE industry
standards
Monitor thermal
conditions
© 2012 IBM Corporation
A singular focus on capital costs can lead to sub-optimized data centers
Source: IBM Estimates
Technology #1 Technology #2 Technology #3
2
1.5
1
.5
0
Year
1 2 3 4 5 6 7 8 9 10
2.5
$3
$ M
illio
ns
Technology 1, Vendor A $2
1.5
1
.5
$ M
illio
ns
Technology 1, Vendor C
Technology 3, Vendor G
Capital Cost Analysis Show 5x difference in technology alternatives
Operating Cost Analysis Show 2.5x difference for vendor alternatives
0
Reduce data center total costs 15-30% over its useful life Improve decisions looking at vendor and technology tradeoffs
Optimize Costs
© 2012 IBM Corporation
Monitoring and management tools provide on-going visibility to improve operational management and operational costs
Perforated Tiles
20%
Cable Cutouts
13.5%
Service AC Units
16.5%
Turn off
CRACs
50%
0.0
1.0
2.0
3.0
4.0
5.0
6.0
Startup Month 3 Month 6 Full Load
Insight from Analytics Power savings from improved cooling
efficiencies
On-going Monitoring 5x improvement in PUE based on
monitoring from start-up
Use real-time monitoring for day-to-day visibility Analytics provide insight to go beyond the “low hanging” fruit
PUE
Optimize Costs
© 2012 IBM Corporation
“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
“No regrets” actions to improve your next data center facility
© 2012 IBM Corporation
A trusted partner knows one size does not fit all. They will help you determine the right solution to meet your needs
Determine the
business
needs
Map the business needs to data center technical requirements
Model outcomes
before you
implement
Model space and capacity during design process
Design a world-class data center
Design an efficient, scalable and reliable solution to meet your business needs
Increase automation
Improve operational efficiency and reduce outages due to human error
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
© 2012 IBM Corporation
Register for the data center study paper: www.ibm.com/data-center/study
Data center services: www.ibm.com/services/smarterdatacenter
For more information, please visit