1. Data Center Consolidation & Energy Efficiency in Federal Facilities

2. Most pressures on data centers affect efficiency
Increasing energy cost
Increasing availability expectations
Regulatory requirements
Server consolidation
Rapid changes in IT technology
Increasing power density
Dynamic power variation
Uncertain
long-term plans for capacity or density
Moves/adds/changes
Data center
Planning / Operation
In addition to the desire to understand and control the costs, there are also some substantial challenges that many companies are facing today in the data center:
Rapid changes in IT technology
I have a number of data center operators who tell me stories like “They told me what was going to be in this facility; I built it; only one third of what they told me is actually what ended up here. The other two thirds are things that weren’t even on the radar screen when I started this project. How am I supposed to design for that?… you know, I was trying to design to the circuit level – the branch circuit level in this facility – and they can’t even tell me a year ahead what’s going to be there and what’s going plug into it.”
High density
This is one of the biggest subjects of critical importance to people today because it’s so destabilizing as far as data center design goes. The new densities we’re talking about, as we all know, are an order of magnitude greater than the densities in typical data centers today.
High operating and service cost:
People are looking at those bills and saying “ What can I do to get these under control or reduce them over time?”
Regulatory requirements:
…forcing unusual structural changes, partitioning of equipment, isolating it physically and providing certain levels of physical-access security around the equipment.
In response to all of these issues, APC contends that a radical change is required in the way we approach data center design. We need to take a holistic view of this problem. The old way – trying to make products better on a product-by-product basis -- just isn’t going to get us there.
What we need to do is…
Standardize process and architecture
Eliminate one-time engineering
Reduce today’s plan-specify-engineer-purchase-install-test-commission cycle time by 90%
Pre-engineer, pre-configure, pre-manufacture, and pre-test NCPI using modular components
Create systems that adapt to changing requirements
Systematically drive out the main cause of downtime: Human error
Integrate power, cooling, and rack into a single management architecture
Empower IT professionals to specify, operate, and manage NCPI
APC has the world’s largest R&D budget focused on this problem
So what can we do about all this and what is the entitlement for improved performance and reduced cost?
We must change the way we design, install, operate, manage, and maintain data centers

3. Why act?
$$$ Reasons
More than 50% of power going into a typical data center goes to the power and cooling systems – NOT to the IT loads
The typical 1MW (IT load) data center is continuously wasting about 400kW or 2,000 tons of coal per year due to poor design (DCiE = 50%, instead of best-practice 70%)
Every kW saved in a data center saves about $1,000 per year
Every kW saved in a data center reduces carbon dioxide emissions by 5 tons per year
Every kW saved in a data center has a carbon reduction equivalent to eliminating about 1 car from the road
A 1% improvement in data center infrastructure efficiency (DCiE) corresponds to approximately 2% reduction in electrical bills 66
White paper

4. Why act?
Directives
EO (Federal Leadership in Environmental, Energy, and Economic Performance)
Federal CIO Memo “Federal Data Center Consolidation Initiative”
Presidential Memo “Disposing of Unneeded Federal Real Estate”

5. Three ways we are helping our customers increase efficiency
Optimize existing data centers (Individual Improvements or Full Consolidation)
Efficiency assessment services
Energy management software tools
Specify and design efficient data centers
High-efficiency reference designs
Detailed efficiency data on products and complete systems
Offer the most efficient data center solutions
Industry-leading efficiency for power and cooling equipment
Row-based cooling
High-efficiency scalable UPS

6. Helping customers optimize existing data centers
Basic application guidance from our experienced sales engineers
Validating design concepts
Basic sizing and specification
Translating clear needs into configured solutions
Professional data center assessment services
Data center efficiency & carbon impact assessment
Power and cooling assessment (capacities and reliability)
DOE Certified Energy Practitioner
Efficiency improvement projects
Implement efficiency improvements
Upgrade inefficient equipment
Install efficiency monitoring equipment
Upgrade zones to high density/efficiency “pods”

7. Typical findings of an electrical efficiency assessment
Benchmark vs. comparable data centers – overall and by subsystem
Operations problems impacting efficiency
Configuration problems impacting efficiency
Short-term quick efficiency improvement options
Long-term opportunities

8. What is the metric? PUE - Power Usage Effectiveness
Power entering a Data Center divided by IT Load Power (Total Facility Power / IT Power)
Expressed as a ratio with efficiency improving as the quotient decreases towards 1
Created by the Green Grid
Adopted by the Federal Government and data center industry groups as the official metric for DC Efficiency (EPA/DOE)
DCiE - Data Center Infrastructure Efficiency
IT Load Power divided by power entering a Data Center (the reciprocal of PUE)
Expressed as a percentage that improves as it approaches 100%
Logical: The higher the percentage the more efficient the Data Center

9. Data center efficiency as a function of IT load
OPERATING LOAD % of capacity being used right now
Efficiency varies dramatically with load, so you can’t specify it without specifying the load

10. Where are YOUR data centers?
A realistic data center infrastructure efficiency plan for the industry
The average will get better The worst will get MUCH better
90%
80%
80%
DCiE
Annual data center infrastructure efficiency
of data centers
BEST 10%
72%
DCiE
70%
65%
65%
DCiE
data center
60%
AVERAGE
(DCiE)
47%
of data centers
50%
WORST 10%
Where are YOUR data centers?
40%
30%
30%
20%
2008
2013
2018

11. Drivers of infrastructure efficiency gains
(Baseline: Average of existing installed base)
IMPROVEMENT
Device Gain
DCiE Gain
$$ saved over 15 years in a 1MW data center**
Move from room cooling to dynamic row/rack cooling
70% 8%
$5,900,000
Cooling economizers
38% 4%
$2,500,000
Right-sizing through modular power and cooling equipment
$2,400,000
Higher UPS efficiency
$1,900,000
415/240 V transformerless power distribution (NAM)*
2.5%
$1,500,000
Dynamic control of cooling plant (VFD fans, pumps, chillers)
25%
$1,200,000
TOTAL to get industry from 47% to 72% DCiE
$14,700,000
*No benefit outside of NAM; Transformer-based PDUs typically in NAM only
**$$ values based on $.15 per kwhr electric cost, starting DCiE of 47%, ave density 8KW/rack

12. Is there an efficiency vs density tradeoff?NO
They work together
High density and high efficiency are related
Many people think high density makes efficiency worse
Only true when trying to push existing data centers to cool high density
High density enables high efficiency in a properly designed new data center or a pod in an existing data center
Shorter pipe lengths Less pump power
Shorter air flow lengths
Less air mixing
Higher return air temperatures Less humidification/dehumidification power and chiller power
Many people assume that by going to high density, they will pay a penalty when it comes to energy efficiency. That is NOT true.
In fact, high density and high efficiency are very closely related.
Implementing high density pods in a data center actually enables higher efficiency, if the right design approach is applied.. And why is this.. Things like:
Shorter pipe lengths which leads to Less pump power
Shorter air flow lengths & Less air mixing leads to less fan power
Higher return air temperatures Less humidification/dehumidification power and chiller power
Less fan power
The highest efficiency data centers
will be high density data centers

13. Achieving high efficiency in high-density data centers
Products we offer as part of our solutions 1
Row-based cooling
Ultra-high efficiency UPS
Capacity and energy management software for power, cooling, and racks
Scalable power and cooling equipment
High-efficiency AC power distribution 2 3 4 5
126
White paper

14. In-row cooling Room cooling
Row-based cooling
Inherently higher power density capability than room designs
Fan power reduced by 50%
Needless dehumidification / re-humidification is eliminated
Need for high-bay areas and raised floors is reduced or eliminated (particularly for small installations)
Cooling capacity can “follow” IT loads that move due to virtualization and server power management
In-row cooling
Room cooling
130
White paper
Cooling efficiency = useful cooling power / (power consumed + useful cooling power)

15. InRow Cooling Unit Examples

16. InRow Cooling Architecture
Close coupled heat source and heat removal
Raised floor not required for air distribution

17. InRow with Containment
Standard servers: 20F delta T
Hot aisle
90F – 95F
Separation of hot and cold aisles creates higher return air temperatures
70F supply air temp

18. Cooling Efficiency Case Study
Cooling Infrastructure Power Consumption
Cooling Component
IRAH
CRAH
CAHU
Units
AHU Fan Power
30.6
88.0
83.2
kW/Hr
Chilled Water Pump Power
10.2
11.0
11.1
Chiller Power
83.9
94.7
94.2
Condenser Pump Power
18.5
Cooling Tower Power
16.2
18.3
18.2
Total Cooling Power Consumed
159.3
230.5
225.1
Annual Cooling
Operating Cost
139,527
201,878
197,211
$ USD
750 kW cooling load
Water cooled chillers
In-row is 30% more efficient
A comparison of the energy consumption of each of the cooling systems is shown on this slide.
Overall the IRAH system is the most efficient of the three cases studied. The largest contribution to the energy savings is the fan power savings. This savings also has a cascading effect resulting in less load on the chiller and cooling tower resulting is less power consumption on these components of the cooling system as can be see here. In addition the 100% sensible cooling operation of the IRAH system reduces the load needed on the chiller because 100% of the unit’s capacity is used to cool the heat load and is not consumed by removing moisture from the space.

19. Ultra-high efficiency Symmetra UPS
High-efficiency UPS
UPS efficiency has large effect on data center efficiency because total IT load goes through it
Improvement over typical from 80% to 96% efficiency at 20% load corresponds to loss reduction of 80%
Electricity savings can pay for the UPS in many cases
Side effect of high efficiency design: UPS is nearly silent
Ultra-high efficiency Symmetra UPS
Typical legacy UPS
LBNL worst-case UPS
% Efficiency

20. Scalable infrastructure minimizes waste
Minimize the inefficiency of oversizing during consolidation and re-growth and be prepared for higher densities to come
Oversizing waste
Since virtualizing can significantly reduce load, oversizing is an important efficiency issue in a virtualized data center. Even without virtualization, oversizing has long been a primary contributor to data center inefficiency. Server consolidation and server power management, by reducing load even more, will shift efficiency further toward the low end of the efficiency curve, if power and cooling systems stay the same. While the electric bill will indeed go down because of the lower IT load and less air conditioning needed to cool it, the proportion of utility power that reaches the IT loads – in other words, efficiency – will drop, which signifies wasted power that could be conserved to further reduce energy consumption.
Virtualization is a new chance to take advantage of scalable infrastructure. Power and cooling devices that can scale in capacity will reduce fixed losses and increase efficiency. Scalable architecture will facilitate not only downsizing to follow IT consolidation, but also subsequent regrowth to follow expansion of the now-virtualized IT load.
Rack density
Scalable power and cooling results in better DCiE

21. What is a “Scalable” and Modular UPS?
500kW of High-Efficiency Scalable Power Protection
375kW
350kW
325kW
400kW
450kW
475kW
300kW
425kW
275kW
125kW
100kW
75kW
50kW
25kW
150kW
250kW
175kW
225kW
200kW
Highly animated slide that can be used to show the simplicity, scalability & modularity of the Symmetra PX250_500kW

22. Scalable power and cooling
Over half the power and cooling losses in a data center are fixed and do not vary with the IT load (so efficiency degrades as IT load declines)
Oversizing is therefore a primary contributor to inefficiency
Virtualization and server power management will make this problem worse, if power and cooling are sized to peak IT loads that are not typical
New power and cooling devices can scale in capacity, reducing fixed losses
% Efficiency
Scalable - modular - when needed
All at once
Power and cooling installation
%IT load

23. Higher voltage AC power distribution (NAM Only)
Distributes voltage to loads at international standard 400/230 instead of conventional US 208/120
Eliminates the need for PDU transformers and associated losses
Reduces copper requirements
Overall efficiency increase of over 5% for a high availability data center
By-product is that 4,000 lbs of copper and 30,000 lbs of PDUs are saved per MW, reducing floor loading and saving space
128
White paper

24. Capacity and energy management
Poor utilization of capacity is a primary cause of inefficiency
Software can identify available capacity (even by rack) and help prevent creation of stranded capacity
Side effect is you can fit more IT equipment in the power and cooling “envelope” of the data center
Energy management can identify efficiency improvement opportunities
Central software
with Capacity Manager
150
White paper

25. Tools we provide to help manage data center carbon footprint
Efficiency Assessment and Carbon Assessment services
Designer
Standard data center specifications
Highest efficiency power and cooling equipment
ISX Central data center management software
Detailed power and cooling device efficiency data

26. How can APC Federal Systems & Schneider Electric help?
Provide High Efficiency Solutions & Architectures
Row-based Cooling
Scalable UPS
High Density
Perform Assessments of your data center
Current PUE
Recommended Actions to Improve PUE
DOE Certified Energy Practitioner
Help Design a High Efficiency Environment for High Density / Virtualization

27. Trade-Off Tools Carbon Calculator - http://www.apc.com/tool/?tt=7
Impact of changes in data center efficiency on energy costs and carbon footprint
Energy Efficiency Calculator -
Impact of alternative power and cooling approaches on energy costs
Capital Cost Calculator -
Impact of physical infrastructure design changes on capital costs
Virtualization Energy Cost Calculator -
Impact of server virtualization and data center design choices on energy and space savings

28. …More Trade-Off Tools
Power Sizing Calculator -
Impact of server and storage configurations on IT load capacity and required utility input power
InRow™ Containment Selector -
Impact of preferences and constraints on the recommended containment approach
AC vs. DC Calculator -
Impact on data center efficiency of various AC and DC power distribution architectures

29. Go Done Go to latest online version of this tool
Click to resume presentation mode
Done

30. Go Done Go to latest online version of this tool
Click to resume presentation mode
Done