ENERGY STAR Data Center Infrastructure Rating Development

Data Collection Form

Data Collection Guidelines

§          The data collection effort will focus on stand-alone data centers and data centers in larger buildings (e.g. office buildings)
that have submetering capability to isolate data center operations.
§          Report values that were valid for at least 12 consecutive months within a period from March 20, 2008 to June 1, 2009.
§          The facility must have been in normal operation for at least 12 months.
§          If data elements change over time (e.g. the number of racks, UPS utilization), provide the average of that element over the
12 month data collection period.
§          If you have any specific questions about the items listed below, or difficulty assessing data during this data collection period,
please send an email to ENERGYSTARdatacenters@icfi.com.

Data Elements List

The following table outlines the data elements of interest; each section of the table corresponds to a tab in this MS Excel Data
Template. This data template was created for your convenience, but we will accept the data in any form, provided all of
the listed information is included.
Data Element Name Description/Question
Tab 1 – Building Information
Data Center Name or ID Please provide a unique name (or number) for each data center in your sample.
Zip Code Please provide the five-digit zip code for each data center's location.

Please select among the following: 1 = Traditional Enterprise; 2 = On-Demand Enterprise; 3 =

Type of Data Center
Telecom; 4 = High Performance Computing Center (Scientific); 5 = Hosting; 6 = Internet; 7 = Hybrid

Please select between the following: 1 = Stand-alone, 2 = Enclosed in a larger building (i.e. office
Type of Building building or other space type). NOTE: If the data center is enclosed within another building space,
it must be completely submetered in order to proceed with this data collection.

Building Square Footage Provide the gross or total square footage of the building.

Provide the square footage of the data center space that includes the rack equipment, service
clearance and circulation, control console area, power distribution, and local air conditioning that is
Data Center Square Footage
encapsulated by the proper protective walls. Exclude administrative offices, storage, loading docks,
and other non-essential space that does not directly support the operation of the data center.

Building Earned the ENERGY STAR? Has the building earned the ENERGY STAR? (Yes or No)
Building LEED Certified? Has the building received the US Green Building Council LEED certification? (Yes or No)
Tab 2 – Data Center Operations
For any element that may change over the course of the data collection period (e.g. number of racks, UPS utilization), please provide an
average figure for the 12 months of data.

Using the Uptime Institute's classification of four tiers of reliability (Tier I, II, III, IV), please provide the
level of the data center's reliability that describes the amount of required, active, redundant
infrastructure. When multiple tier levels exist, the reliability tier level should represent the majority, or
the highest percentage, of the data center's total load. When estimating tier levels, the fractional
Reliability (Tier Level)
estimates (i.e. 3+ or 3.5) should be rounded DOWN to the whole number for purposes of the data
input.
(See Uptime's Web site for more information on the tier classifications:
http://www.upsite.com/TUIpages/whitepapers/tuitiers.html)

Number of Racks Please provide the number of racks in the data center that are typically in operation.

Based on your data center's tier level and corresponding UPS configuration, please enter the percent
utilization of your UPS, which is the amount of critical load compared with the redundant capacity of
UPS Utilization the system. Please use the examples in the FAQs to see the calculations associated with each tier
level and UPS configuration. If you have multiple systems and multiple configurations, please use the
pro-rated average of all systems weighted by load and their individual system utilizations.

Total Capacity for Cooling To help determine the Chiller Utilization, provide the data center's total capacity for cooling in tons.
Annual Chiller Runtime To help determine the Chiller Utilization, please provide the annual runtime for the chillers in hours.
Average Chiller Demand To help determine the Chiller Utilization, enter the average demand for the chillers in KW.
Data Element Name Description/Question
Tab 3 – IT Measurement
Please enter the annual total IT plug energy, as measured in KWh, from the output of a UPS meter,
for energy going into the computer room (not crac units, etc.). This can be entered as one value for
the year or as a series of meter/energy readings (i.e. monthly or periodic) that cover the entire
Total IT Plug Energy from UPS Meter
calendar year. It is preferable that these measurements are read at approximately the same period
in KWh
and frequency as the utility meter for the building.

Start Date For each energy use value, enter a start and end date. You can provide an entry for each
End Date meter/energy reading or one entry for the entire year.
Please enter the annual total IT plug energy, as measured in KWh, from the input of a PDU meter, for
energy going into the computer room (not crac units, etc.). This can be entered as one value for the
year or as a series of meter/energy readings (i.e. monthly or periodic) that cover the entire calendar
Total IT Plug Energy from PDU Meter year. It is preferrable that these measurements are read at approximately the same period and
in KWh frequency as the utility meter for the building. If both the UPS and PDU meters are provided, an
(If Available) estimate of the difference between the total UPS and the sum of the PDU's will be estimated as the
non-critical UPS load that might be serving the office, mechanical and non-data center loads.

Start Date For each energy use value, enter a start and end date. You can provide an entry for each
End Date meter/energy reading or one entry for the entire year.
Tab 4 – Electricity Data

Provide the annual electricity consumption for the building (if a stand-alone data center) or the submetered data center. You can provide this
as one value for the year or as a series of meter/energy bill entries that cover the entire year.
Meter ID Enter a distinguishing name or number for each meter.
Start Date For each energy use value, enter a start and end date. You can provide an entry for each bill or one
End Date entry for the entire year.

Electricity Consumption in KWh Enter the electricity consumption in KWh corresponding to the start and end date.
Data Element Name Description/Question
Tab 5 – Natural Gas Data
Provide the annual natural gas consumption for the building (if a stand-alone data center) or the submetered data center. You can provide
this as one value for the year or as a series of meter/energy bill entries that cover the entire year.
Meter ID Enter a distinguishing name or number for each meter.
Start Date For each energy use value, enter a start and end date. You can provide an entry for each bill or one
End Date entry for the entire year.

Natural Gas Consumption Enter the natural gas consumption corresponding to the start and end date.
Energy Units Please enter the corresponding units for the natural gas consumption figure (i.e. therms, ccf, Mcf).
Tab 6 – Other Energy Sources

For each fuel type and other sources of energy (i.e. diesel fuel, steam, chilled water, solar) provide the annual energy consumption for the
building (if a stand-alone data center) or the submetered data center. You can provide this as one value for the year or as a series of
meter/energy bill entries that cover the entire year. NOTE: Stand-by generator diesel fuel used less than 200 hours per year should not
be included.
Fuel Type/Other Energy Source Please enter the type of energy that the following data represents.
Meter ID Enter a distinguishing name or number for each meter.
Start Date For each energy use value, enter a start and end date. You can provide an entry for each bill or one
End Date entry for the entire year.

Energy Consumption Enter the energy consumption corresponding to the start and end date.
Please enter the corresponding units for the energy consumption figure (i.e. gallons, therms, Kbtu,
Energy Units
Mbtu, lbs, ton hours).
Data Element Name Description/Question
Tab 7 – Optional Elements (To be used for future study purposes)
Year Constructed Enter the 4 digit year the building was built.
HVAC "Economizer Cycle" Control Does the building utilize an HVAC "economizer cycle" control strategy? Yes or no?
Strategy

Water-side or Air-side Economizer If yes on "economizer cycle" is this on the water-side or air-side? Please select between the following:
Cycle? 1 = water-side, 2 = air-side

Does the building incorporate peak

Does the building incorporate peak shaving, demand response, thermal storage or co-generation in its
shaving, demand response, thermal
operations? Please select all operations/strategies that apply.
storage or co-generation in its
operations?
What is the current "average" IT processing load as a percentage of the total server processor
Percentage of Average IT Utilization capacity? Enter this percentage as one average calculation for the entire year.

What is the current "peak" IT processing load as a percentage of the total server processor capacity?
Percentage of Peak IT Utilization Enter this percentage as one average calculation for the entire year.

Please select the type of mechanical system used for cooling the data center: 1 = Direct Expansion; 2
Mechanical System Type
= Chilled Water; 3 = Condenser; 4 = Other
Liquid-Cooled IT Equipment/Racks Please check the box if the data center employs liquid cooling for IT equipment and/or racks.

Please check off all of the following air management techniques that are in operation at the data
center:
Blanking Plates in Racks
Hot Aisle/Cold Aisle Equipment
CFD Modeling
Air Management Techniques High Delta "T" Cooling
Air Separation Chambers
Racks with Air Management Control
Static Pressure Control
Static Pressure Boundaries
Dynamically Managed Air-Flow

Please check off all of the following operational practices employed at the data center: New ASHRAE
Operating Set Points
Humidity Set Points; Dynamic Temperature Change; Automatic Chiller Water Reset
 Frequently Asked Questions (FAQ)

Review the additional information for select questions below. If you need further clarification, email your questions to: ENERGYSTARDatacenters@icfi.com

Tab 1 - Building Information

Type of Data Center

1 - Traditional Enterprise: Processes data requirements for data center owner

2 - On Demand Enterprise: Processes and expands data processing capacity, when needed, for numerous customers (this would include batch processing and
manufacturing type data processing facilities)
3 - Telecom: Telecom Switching Center
4 - High Performance Computing: Scientific and high density data processing
5 - Hosting: Sells data processing services to numerous Customers
6 - Internet: Provides high capacity processing for large numbers of web clients
7 - Hybrid: Combination of two or more of the above
Tab 2 – Data Center Operations

UPS Utilization

There are various configurations of UPS support. Some of the redundancy of the UPS system design will be indicated by the Tier Level already collected in the data
elements. Efficiency is also affected by the load on the UPS system (a lightly loaded UPS is less efficient than a system operation in the peak efficiency above 90% -
depending on the system manufacture). The formal definition of the % of UPS capacity is the amount of critical load compared with the redundant capacity of the system.
For illustrative purposes, examples of the utilization calculation could be any of the following:

Example # 1 Tier One Stand-Alone UPS (No redundancy): The percent of utilization would be calculated by noting a fully loaded UPS should not exceed 85% of the rated
capacity - for transients harmonic and temporary loads with equipment swap. Therefore the percent utilization calculation in this example is a 500 kVA unit has 425 kVA of
capacity (85%) and a 150 kVA IT load on this unit would represent a 35.3% loaded unit. (150 / 500 x 0.85 = 35.3%)

Example # 2 Tier Two N+1 Redundant UPS: The percent of utilization in this example would be noted as a three module 500 kVA UPS system (three modules of 500 +
500 + 500 = 1000 kVA (plus a redundant module)) since all modules share load and 1500 kVA will be on line a full 850 kVA of capacity of redundant system. A 150 kVA IT
load in this system would NOT be 10% of 1500 kVA installed but 17.6% of system utilization (150 / (1500 - 500 (redundant module) X 0.85 = 17.6%). The "normal-mode"
UPS loading of each module is 150 / 1500 = 10%, but the system is utilized to 17.6%.

Example # 3 Tier Three N+N UPS modules: The percent utilization would be calculated be noting the redundant UPS (N) system will not need a de-ration factor of 85%
since the system will share load between the "A & B" side UPS systems. In this example the systems would be configured in two 500 kVA modulate (1000 kVA - "A") + two
additional 500 kVA modules ( 1000 kVA - "B") for a total installed 2000 kVA of UPS modules. This configuration can support 850 kVA redundant load while being fully
maintained on the "A" side, while the "B" side is supporting the load (concurrent maintainability). The load calculation for 150 kVA of IT load is ( 150 / (2000 - 1000
(redundant N system)) x .85 = 17.6%). The "normal-mode" UPS load shared across all modules would be 7.5% each but the system utilization in the redundant
configuration would still be 17.6% loaded.

Example # 4 Tier Four N+1 + N+1 UPS System: The difference between Tier Three and Tier Four configurations is the redundant module to offer redundancy when the
system is in the maintenance mode. If the total load was being supported by the "A" side while the "B" side is being maintained, the "A-side +1" module will still be in
reserve for a module failure. In this example the modules installed would be three 500 kVA modules (N+1) for the "A" side and three 500 Kva modules for the "B" side
UPS and a system redundant capacity of 1000 kVA. Since each system has reserve when supporting the total load, the 85% de-rating factor is not necessary. Therefore
the utilization load for this configuration is [ 150 / (3000 - (1500 (redundant system) + 500 (redundant second system module))] = (150 / 1000 ) = 15.0%

Chiller Utilization
These three elements (total capacity for cooling, annual chiller runtime, and average chiller demand) will be factored into an equation to determine the total IT load being
cooled compared to the chiller capacity.
Tab 3 – IT Measurement

Total IT Plug Energy

In general the calculation is attempting to capture the energy used by the IT equipment (this will include the IT equipment fan systems and all other devices: KVM, cabinet
monitoring and cabinet circulation fans) that are connected into the plug strip serving the computer room. If the UPS system is only providing load to the computer
equipment in the data center, and NOT providing Computer room fans, pumps, telecommunication closets and desktop services outside the data center, then the UPS
readings – over time can be recorded. The best case in this circumstance is to read a KWH meter at approximately the same period and frequency as the utility meter for
the building. If a KWH meter is not available on the UPS then the instantaneous meter reading can be measured on a periodic basis to predict how load on the UPS
changes with the data center processing load. A comparison with the known processing peak and the lowest utilization would be good times to record instantaneous meter
readings to capture the range of UPS output. If it is greater than 5%, a more detailed (greater frequency) of recording would be used to estimate the total KWH provided
by the UPS over time.
Tab 6 – Other Energy Sources

Generally these systems provide maintenance start-up testing and battery re-fresh run-time exercise cycles that do not sustain actual data center operations for extended
periods (defined as less than 200 hours per year). Since the on-line status is not significant you can ignore the make-up diesel fuel as a captured energy source.
Tab 7 - Optional Elements

Does the building utilize an HVAC "economizer cycle" control strategy?

“Economizer Cycle” is a mechanical device that uses any type of system that eliminates the mechanical refrigeration cycle by using outside atmosphere to cool the interior
spaces that are normally cooled by an energy intensive refrigerant system. These would include free cooling with economy coils using cooled water from the cooling tower
or dry coolers, replacement air used directly from the outside (with exhaust air being flushed to the outside), or a heat exchanger process that transfers heat to the
atmosphere without the use of a mechanical input to change a refrigerant state.

If yes on "economizer cycle" is this on the water-side or air-side?

A “Water-side” economizer cycle uses water as the key fluid in the transfer of heat to the atmosphere and does not introduce a large quantity of outside air directly into
cooled interior spaces. Usually coils and/or heat exchangers are used with these types of systems.

An “Air-side” economizer cycle generally uses the induction of outside air (when conditions of that air can cool) for the direct conditioning of the space. There is a need to
monitor humidity and contaminates when this type of cooling is provided with attention to filtering and the moisture content of the air.

Does the building incorporate peak shaving, demand response, thermal storage or co-generation in its operations?
Peak shaving is a technique that stores heating or cooling generated in an alternate (cheaper cost) time of day to be used when the utility system is experiencing a high
demand on their system and charges customers for energy use at a higher cost during that period. The shaving provides the utility with higher system utilization and less
capacity peaks – they typically provide customers with a reduced rate of cost for this benefit. Systems can include chilled water storage, ice storage, heated material, solar
heated storage and other processes. The utility rate reduction may lower total utility cost, but total energy usage is generally displaced but not reduced with these
systems.

Demand response (partial) and co-generation (continuous) processes provide removal of utility load from the utility when their system is stressed. For the ability to
temporarily remove or self-generate power, the utility will typically provide a reduced utility rate. These processes may or may not save total energy used.

The ENERGY STAR data collection effort is interested in large energy input to the data center. If the peak shaving, demand response and storage processes do not save
energy and are primarily used for utility cost reduction, the effect on an energy rating will be minimal on the calculated rating. Co-generation however would by its
continuous nature use a large quantity of energy and therefore would need to be included in the “other energy” tab 6&7 collection.
Percentage of Average IT Utilization

Typically the data center monitors the processor, network and storage processes that occur in the data center. There are peaks in production that when reached, tend to
determine if additional processing equipment is required to support the data center applications. The active measurement of the IT equipment is usually automated and
monitored. What it being asked in this data element is not the peak, but the average estimate over time of the IT equipment’s utilization (compared to the processor being
100% utilized). It is a value that currently will not affect the index but is the first step in data capture that will allow correlation the actual IT processing work with efficiency.
Although there is no standard to measure IT work being processed with energy input, this one measurement of the processor activity verses a dormant state will provide
an estimation of the potential for energy saving features, if activated, that could achieve even better efficiencies.

Percentage of Peak IT Utilization

If the console operator can run their capacity or processor monitoring programs, the peak utilization is used for alerting management when more IT equipment is
necessary to sustain business application processing. Since this is measured in most data centers, the historical capture should be accurate. The reason for this data
element is to provide an understanding of the maximum equipment that will be necessary for peak periods and then, depending on the variability of the enterprise
processing needs, be dormant during the non-peak times. By understanding the peak and average utilization characteristics of the data center, analysis with the tracking
of input energy and the extent of dormant equipment sitting in reserve for the peak. If these processes are tracked for the data center, the use of devices that can slow
down processor clock speed and turn off redundant processors can then be estimated. These concepts are the start of understanding IT work efficiency. For where the
data is captured, ask your console operator or check the measurement in the IT utility program available in the processor equipment.
 Tab 1 - Building Information
Building Earned
Data Center Building Square Data Center the ENERGY Building LEED
Name / ID Zip Code Type of Data Center Type of Building Footage Square Footage STAR? Rated?
Example #1 21045 1 - Traditional Enterprise 1 - Stand-Alone 36,000 21,600 No No
 Tab 2 - Data Center Operations
Total Capacity for Annual Chiller Average Chiller
Data Center Reliability Number of Racks Cooling Runtime Demand
Name / ID (Tier Level) UPS Utilization (%) (tons) (hours) (KW)
Example #1 Tier II 420 60% 950 8760 550
 Tab 3 - IT Measurement

Total IT Plug
Total IT Plug Energy from PDU
Data Center Energy from UPS Start Date End Date Meter (KWh) Start Date End Date
Name / ID Meter (KWh) (MM/DD/YYYY) (MM/DD/YYYY) If Available (MM/DD/YYYY) (MM/DD/YYYY)
Example #1 1,824,270 02/21/2008 03/20/2008 1,806,028 02/21/2008 03/20/2008
Example #1 1,814,000 01/21/2008 02/20/2008 1,795,860 01/21/2008 02/20/2008
Example #1 1,723,300 12/21/2007 01/20/2008 1,706,067 12/21/2007 01/20/2008
Example #1 1,723,278 11/21/2007 12/20/2007 1,706,045 11/21/2007 12/20/2007
Example #1 1,723,222 10/21/2007 11/20/2007 1,705,990 10/21/2007 11/20/2007
Example #1 1,723,168 09/21/2007 10/20/2007 1,705,936 09/21/2007 10/20/2007
Example #1 1,723,111 08/21/2007 09/20/2007 1,705,880 08/21/2007 09/20/2007
Example #1 1,723,056 07/21/2007 08/20/2007 1,705,825 07/21/2007 08/20/2007
Example #1 1,723,000 06/21/2007 07/20/2007 1,705,770 06/21/2007 07/20/2007
Example #1 1,722,944 05/21/2007 06/20/2007 1,705,715 05/21/2007 06/20/2007
Example #1 1,722,889 04/21/2007 05/20/2007 1,705,660 04/21/2007 05/20/2007
Example #1 1,722,833 03/21/2007 04/20/2007 1,705,605 03/21/2007 04/20/2007
 Tab 4 - Electricity Data
Electricity
Data Center Start Date End Date Consumption
Name / ID Meter ID (MM/DD/YYYY) (MM/DD/YYYY) (KWh)
Example #1 Electric 45389 02/21/2008 03/20/2008 3,283,686
Example #1 Electric 45389 01/21/2008 02/20/2008 3,265,200
Example #1 Electric 45389 12/21/2007 01/20/2008 3,102,000
Example #1 Electric 45389 11/21/2007 12/20/2007 3,101,900
Example #1 Electric 45389 10/21/2007 11/20/2007 3,101,800
Example #1 Electric 45389 09/21/2007 10/20/2007 3,101,700
Example #1 Electric 45389 08/21/2007 09/20/2007 3,101,600
Example #1 Electric 45389 07/21/2007 08/20/2007 3,101,500
Example #1 Electric 45389 06/21/2007 07/20/2007 3,101,400
Example #1 Electric 45389 05/21/2007 06/20/2007 3,101,300
Example #1 Electric 45389 04/21/2007 05/20/2007 3,101,200
Example #1 Electric 45389 03/21/2007 04/20/2007 3,101,100
 Tab 5 - Natural Gas Data
Data Center Start Date End Date Natural Gas
Name / ID Meter ID (MM/DD/YYYY) (MM/DD/YYYY) Consumption Energy Units
Example #1 Gas 2624 02/05/2008 03/04/2008 7,715 therms
Example #1 Gas 2624 01/05/2008 02/04/2008 8,000 therms
Example #1 Gas 2624 12/05/2007 01/04/2008 2,700 therms
Example #1 Gas 2624 11/05/2007 12/04/2007 2,500 therms
Example #1 Gas 2624 10/05/2007 11/04/2007 2,300 therms
Example #1 Gas 2624 09/05/2007 10/04/2007 2,200 therms
Example #1 Gas 2624 08/05/2007 09/04/2007 1,800 therms
Example #1 Gas 2624 07/05/2007 08/04/2007 1,750 therms
Example #1 Gas 2624 06/05/2007 07/04/2007 1,900 therms
Example #1 Gas 2624 05/05/2007 06/04/2007 2,100 therms
Example #1 Gas 2624 04/05/2007 05/04/2007 2,400 therms
Example #1 Gas 2624 03/05/2007 04/04/2007 2,600 therms
 Tab 6 - Other Energy Sources
Data Center Fuel Type/ Start Date End Date Energy
Name / ID Other Energy Source Meter ID (MM/DD/YYYY) (MM/DD/YYYY) Consumption Energy Units
Example #1 N/A
 Tab 7 - Optional Elements
Air Management Techniques Operating Set Points

in

r
w
Ra T
s

se ile
.
tes

.
Pts
ure

e
nt/ d I
ck

nt

Flo
uip old

mp
nt r i r
n
T"

rie ur

yS E

Re Ch
me

be atio

Co h A
me ole

ng
Pl a

dit HRA
ol

ge l l y
da ss
o l ss
ng a "
HVAC

ir-

ge Te
Eq e/ C

et

t
ter atic
eli

s
na ica
dA
uip -Co

mt wit

ntr Pre

un Pre
am a r
oli elt
"Economizer Water-side or Does the building incorporate

rs
ck ng

an ic
l

od

S
Ais t Ais

Ch Sep

Ma nam

Ch nam
Co gh-D

Wa tom
Cycle" Air-side peak shaving, demand Percentage of Percentage of

Hu A
Ra nki
Eq uid

Mg cks

Co tic

Bo c
DM
s

w
t

mi
Data Center Year Control Economizer response, thermal storage, or Average IT Peak IT Mechanical System

le
Bla

Sta

Sta
Liq

Air
Ho

Au
Ne
Ra

Dy

Dy
Hi
CF
Name / ID Constructed Strategy? Cycle? co-generation in its operations? Utilization Utilization Type

✘ ✘ ✘

Example #1 1998 No 63% 76% 2 = Chilled Water

 Name / ID
Data Center
Year
Constructed
HVAC

Cycle"
Control
Strategy? Cycle?
Air-side
Economizer
peak shaving, demand
response, thermal storage, or
"Economizer Water-side or Does the building incorporate

co-generation in its operations? Utilization

Average IT Peak IT
Utilization
Percentage of Percentage of

Type
Mechanical System

Liq
Eq uid
uip -Co
me ole
nt/ d I
Bla Ra T
ck
Ra nki
ck ng s
s Pl a
tes
Ho in
Ais t Ais
le l
E q e/ C
CF uip old
DM me
nt
od
eli
ng
Hi
Co gh-D
oli elt
ng a "
T"
Air
Ch Sep
am a r
be atio
rs n
Ra
Mg cks
mt wit
Co h A
nt r i r
Sta ol
Co tic
ntr Pre
o l ss
ure
Sta
Bo c ti
un Pre
da ss
rie ur
Dy s e
Ma nam
na ica
ge l l y
dA
ir-
Ne Flo
Hu Aw w
mi S
dit HRA
yS E
Dy et
Pts
Ch nam
an ic .
ge Te
mp
Au .
Wa tom
ter atic
Re Ch
se ile
t r