Cloud Energy
and Emissions
Research Model

CLEER Model Technical Documentation

Version 2
June 24, 2013

Developed by Lawrence Berkeley National Laboratory and Northwestern University
Jump to descriptions:

Data center operations
Client IT device operations
Network systems operations
Business appliance operations
Residential appliance operations
Data center embodied
Data center IT embodied
Client IT embodied
Network systems embodied
Transportation activities
Other materials embodied

Variable definitions

Introduction

The aim of the CLEER Model is to provide a comprehensive and user-friendly framework for assessing the net energy and greenhouse gas (GHG) emissions of cloud computing systems compared to the existing digital and physical systems that they might replace.   As such, the CLEER Model was designed to be both transparent and simple to encourage use and further development by the research community. 

This page provides a concise overview of the mathematical framework employed by the CLEER Model in its net energy and emissions calculations.

Taxonomy

The CLEER Model uses a bottom-up framework to analyze major societal end uses of energy within an analysis system boundary defined by the user.  Each end use of energy is comprised of different energy use elements that are also defined by the user.  Each end use is represented by a sub-model in the CLEER Model framework. 

Currently, the CLEER Model contains the following end use sub-models.  The elements within each sub-model are described in the sections that follow.
  1. Data center operational energy use
  2. Business client IT device operational energy use
  3. Residential client IT device operational energy use
  4. Data transmission systems (i.e., network systems) operational energy use
  5. Commercial building appliance operational energy use (not available through public use model)
  6. Residential building appliance operational energy use (not available through public use model)
  7. Embodied energy and emissions of data center building materials
  8. Embodied energy and emissions of data center IT devices
  9. Embodied energy and emissions of business client IT devices
  10. Embodied energy and emissions of residential client IT devices
  11. Embodied energy and emissions of network devices
  12. Transportation systems operational energy use (not available through public use model)
  13. Embodied energy and emissions of other materials and products (not available through public use model)
The data center sub-model is further divided into different six different data center types: (1) server closets, (2) server rooms, (3) localized data centers, (4) mid-tier data centers, (5) enterprise-class data centers, and (6) cloud data centers. 

The CLEER Model allows for a comparison of present-day systems for providing a digital or physical service to hypothetical cloud-based systems that would provide an equivalent service.  Email is an example of a digital service that might be hosted in local data centers in the present day, but could be shifted to cloud data centers in a cloud-based scenario.   Music is an example of a service that might presently be provided to some consumers by physical systems, for example, compact discs, that could be shifted to the cloud in the form of streaming or downloaded digital music files.

Mathematical notation

In the equations that follow, we use the subscript i to denote an end use, the subscript j to denote a location or facility, and the subscript k to denote an end use element.

The following general relation is used to calculate the net difference in primary energy use between present-day and cloud-based systems:


Equation 1
Equation 1

where Eic refers to the annual primary energy of end use i in the cloud-based system and EiP refers to the annual primary energy of end use i in the present-day system.   The variable m refers to the total number of end uses that the user includes in the analysis system boundary.  As indicated above, the CLEER Model currently allows for inclusion of up to 13 different end uses.  The many energy use elements that comprise each end use are described in the following sections.

The CLEER Model considers both primary energy use and direct energy use.  Primary energy use is indicated by an upper case E and direct energy use is indicated by a lower case e in our mathematical notation.  Examples of direct fuel use include the onsite electricity use of servers or the gasoline use of passenger cars.  Primary energy use accounts for the heating values of the fuels needed to generate electricity at the power plant as well as the embodied energy associated with fuel extraction and refining prior to combustion for all fuels. Furthermore, all embodied energy values in the CLEER Model (e.g., the manufacturing and end of life energy associated with a server) are expressed in units of primary energy.

The following general relation is used to calculate the annual primary energy use associated with an end use i in consideration of both direct energy use elements and embodied energy use elements:

Equation 2
Equation 2
where eijk is the annual direct fuel use of element k in facility/location j for end use i, pjk is the primary energy conversion factor for the fuel used by element k in facility/location j, and E'ijk is the annual embodied primary energy associated with element k in facility/location j of end use i. The variables n and q refer to the total number of facilities/locations and end use elements, respectively, that the user includes in the analysis of a given end use.  It is possible to specify primary energy conversion factors for end use elements by facility/location, given that electricity mix or fuel properties might vary across end uses and elements (e.g., due to different regional electricity mixes or fuel heating values).

The following general relation is used to calculate the net difference in GHG emissions between present-day and cloud-based systems:

Equation 3
Equation 3
where Gic refers to the annual GHG emissions of end use i in the cloud-based system and GiP refers to the annual GHG emissions of end use i in the present-day system.   The variable m refers to the total number of end uses that the user includes in the analysis system boundary.  All GHG emissions are expressed in units of carbon dioxide equivalents (CO2e) in the CLEER Model.

The CLEER Model includes the GHG emissions associated with direct fuel use (e.g., GHG emissions arising from the electricity use of servers) and GHG emissions that are embodied in the materials and devices that comprise the system (e.g., GHG emissions associated with the manufacture of a server).    The following general relation is used to calculate the GHG emissions associated with an end use:

Equation 4
Equation 4

where eijk is the annual direct fuel use of element k in facility/location j for end use i, gjk is the GHG emission factor for the fuel used by element k in facility/location j, and G'ijk is the annual embodied GHG emissions associated with element k in facility/location j of end use i. The variables n and q refer to the total number of facilities/locations and end use elements, respectively, that the user includes in the analysis of a given end use.  It is possible to specify GHG emission factors for end use elements by facility/location, given that electricity mix or fuel properties might vary across end uses and elements (e.g., due to different regional electricity mixes or fuel carbon contents).

The remaining sections describe the mathematical frameworks for each end use sub-model using the notation described above. For convenience, the end use sub-models are listed using the i index value provided in the Taxonomy section above. For brevity, only the equations for calculating primary energy use are presented.  The equations for calculating GHG emissions are identical with two exceptions: (1) gjk is used in lieu of pjk to convert direct energy use to GHG emissions as opposed to primary energy use; and (2) the capital G is used in lieu of the capital E in all equations to denote GHG emissions as opposed to primary energy use.


Back to Introduction

Data center operational energy use sub-model (i=1)

Data center operational energy use is modeled in bottom-up fashion by considering the direct energy use of data center IT devices and the direct energy use associated with data center infrastructure equipment in different data center facility types.  Six different data center facility types are defined in the model to represent differences in devices, infrastructure equipment, and operating practices across different data centers:

Data center facility types (j)
  1. Server closets
  2. Server rooms
  3. Localized data centers
  4. Mid-tier data centers
  5. Enterprise-class data centers
  6. Cloud data centers
The CLEER Model currently includes the following data center IT device and infrastructure equipment elements within each data center type:

Data center IT devices (k)
  1. Volume servers
  2. Mid-range servers
  3. High-end servers
  4. External hard disk drive (HDD) storage
  5. Network devices
Infrastructure equipment (k)
  1. Transformers
  2. Uninterruptible power supplies (UPS)
  3. Cooling
  4. Lighting
The annual direct energy use of a given data center IT device is modeled as follows:

Equation 5
Equation 5
where e1jk is the direct electricity use of IT device k (e.g., volume servers) in data center type j, N1jk is the total number of installed IT devices k in data center type j, W1jk is the average power use of IT device k in data center type j (in watts), and h1jk is the annual operating hours (in hours per year) of IT device k in data center type j.

Data center center infrastructure equipment energy use is captured using an expression for power utilization effectiveness (PUE) when calculating the annual primary energy use of data centers as follows:

Equation 6
Equation 6
where E1 is the annual primary energy use of data center operations (i = 1), the first term in parentheses is the PUE of data center type j, and the second term in parentheses is the sum of all IT device primary energy use in data center type j.  Users can either enter the PUE value for each data center type or the direct energy use of each infrastructure element (k = 5-9) in Equation 6 and the CLEER Model will calculate the PUE.  Both options are available to accommodate user preference.


Back to Introduction

Client IT device operational energy use sub-models (i=2 for businesses, i=3 for residences)

Client IT devices refer to electronic devices used to access a digital service such as email, streaming video, or websites for e-commerce.  The CLEER Model currently includes client IT devices that are representative of major devices used by both business users and consumers.  Although separate sub-models are offered for business client IT devices and residential client IT devices, they are described together in this section because the sub-models are identical from a mathematical perspective.  We use i=2 in the equations below for convenience. The following 14 client IT devices are included as elements in the sub-models:

Client IT devices (k)
  1. Desktop PCs
  2. Notebook PCs
  3. Netbooks
  4. Flat panel computer monitors (17")
  5. Tablet computers
  6. Smart phones
  7. Digital music players
  8. DVD players
  9. Stereo systems (CD)
  10. Set top boxes
  11. Flat panel TV <40"
  12. Flat panel TV >40"
  13. Video gaming console
  14. Other (user defined)
Client IT device direct energy use is modeled as follows:

Equation 7
Equation 7
where e2jk is the direct electricity use of client device k (e.g., a tablet PC) in facility/location j and N2jk is the number of client devices specified by the user, which depends on the analysis scope.  The mode variables W2jk and h2jk refer to the power use (in watts) and total device operating hours (in hours per year) in "on," "sleep/idle," and "off" modes, respectively, for client device k in facility location j
The sum of operating hours across all modes must equal the number of hours in a (non-leap) year (8760). 

The user can also specify what fraction of the annual "on" hours x2jk of client device k in facility/location j that is dedicated to the service being analyzed (e.g., email).  In this way, Equation 7 allocates the same fraction of "sleep/idle" and "off" mode energy use to the service as it does "on" mode energy use.  The rationale for this approach is that
most devices spend many hours in "sleep/idle" and "off" modes but still draw power; thus, we allocate this non-productive power to the service based on its share of annual "on" mode use for a given client IT device.

The annual primary energy use of client IT devices is calculated as follows:

Equation 8
Equation 8

where E2 is the annual primary energy use of client IT devices (i=2 for businesses, i=3 for residences).


Back to Introduction

Network systems operational energy use sub-model (i=4)

The energy use of transmitting data over networks is modeled in aggregate fashion for different network segments.  Users must define the segments that comprise a particular pathway for transmitting data from data centers to client devices and specify the percent of total annual data traffic that is sent over the selected pathway(s).  The CLEER Model will then calculate cumulative network energy use across pathways by adding up the energy use of each segment.  The following 12 network segment elements are included in the network sub-model based on definitions from Bagila et al. (2009, 2011):

Network segments (k)
  1. Wavelength division multiplexing (WDM) segment (terrestrial and submarine)
  2. Core network segment
  3. Metro network segment
  4. Video distribution network (VDN) segment
  5. Asymmetric digital subscriber line (ADSL) access
  6. Hybrid fiber coaxial (HFC) access
  7. Passive optical network (PON) access
  8. Optical point to point (PTP) access
  9. Fiber to the node (FTTN) access
  10. WiMAX access
  11. UMTS access
  12. LTE access
The above 12 network segments are included in the CLEER Model to provide sufficient flexibility to model a number of different segment combinations. Default values for each have been obtained from the literature, but users are encouraged to enter their own values for the energy use of each segment given that network technologies (and data available on those technologies) can change rapidly.

The annual primary energy use of data transmission is modeled as follows:

Equation 9
Equation 9

where E4 is the annual primary energy use of data transmission, A4 is the annual amount of data transmitted (in bits) for the service being analyzed (e.g., email), y4jk is the fraction of those bits that are transmitted across network segment k in location j, e4jk is the direct electricity intensity of the network segment (in joules per bit) in location j, and pjk is the primary energy conversion factor for network segment k in location j.  Users should note that the CLEER Model currently uses the U.S. average primary energy conversion factor for electricity as a default for all network segments k in the Equation 9, given that tracking the specific pathway of any given data packet is difficult.  However, users can change the electricity mix if desired to better reflect the primary energy intensity of local access network segments for which the grid mix may be known. 

Baliga, J., R. Ayre, K. Hinton, W.V. Sorin, and R.S. Tucker (2009). Energy Consumption in Optical IP Networks. Journal of Lightwave Technology, Volume 7, Number 13.
Baliga, J., R. Ayre, K. Hinton, and R.S. Tucker (2011). Energy Consumption in Wired and Wireless Access Networks. IEEE Communications Magazine. June 2011.



Back to Introduction

Commercial building appliance operational energy use sub-model (i=5)

Building appliances are included in the CLEER Model for consideration of how non-IT building energy use might change under cloud-based services.  For example, telework might drive greater use of residential appliances but lesser use of commercial appliances if there are fewer workers and/or less required floor space in the office.  The following 13 commercial appliances are included as elements in the commercial appliance sub-model, based on data from the United States Department of Energy's 2011 Buildings Energy Data Book:

Commercial building appliances - electrical (k)
  1. Lighting
  2. Space Heating
  3. Space Cooling
  4. Ventilation
  5. Refrigeration
  6. Water Heating
  7. Office electronics
  8. Cooking
  9. Other (user defined)
Commercial building appliances - natural gas (k)
  1. Space heating
  2. Water heating
  3. Cooking
  4. Other (user defined)
The annual primary energy use of commercial appliances is calculated as follows:

Equation 10
Equation 10

where E5 is the annual primary energy use of commercial appliances, e5jk is the energy intensity of appliance k in location j (energy per square foot of floor space per year), F5jk is the commercial floor space (in square feet) dedicated to the service being analyzed,
and pjk is the primary energy conversion factor for appliance k in location j.  Appliance energy intensities are expressed per unit of floor space for two reasons: (1) it is the conventional energy intensity unit in commercial building energy analysis; and (2) the best available building energy data for the United States are expressed using this energy intensity unit.  The user must specify the appropriate floor space assumptions for each service being analyzed.


Back to Introduction

Residential building appliance operational energy use sub-model (i=6)

Building appliances are included in the CLEER Model for consideration of how non-IT building energy use might change under cloud-based services.  For example, telework might drive greater use of residential appliances but lesser use of commercial appliances if there are fewer workers and/or less required floor space in the office.  The following 28 residential appliances are included as elements in the residential appliance sub-model, based on data from the United States Department of Energy's 2011 Buildings Energy Data Book:
 
Residential building appliances - electrical (k)
  1. Coffee Maker
  2. Dishwasher
  3. Microwave Oven
  4. Toaster Oven
  5. Refrigerator-Freezer
  6. Freezer
  7. 18-W Compact Fluorescent
  8. 60-W Incandescent Lamp
  9. 100-W Incandescent Lamp
  10. Torchiere Lamp-Halogen
  11. Hair Dryer
  12. Waterbed Heater
  13. Clothes Dryer
  14. Clothes Washer
  15. Dehumidifier
  16. Furnace Fan
  17. Ceiling Fan
  18. Center air conditioning
  19. Space Heater
  20. Water Heater-Family of 4
  21. Water Heater-Family of 2
  22. Portable Spa
  23. Other (user defined)
Residential building appliances - natural gas (k)
  1. Space heating
  2. Water heating
  3. Dryer
  4. Range/oven
  5. Other (user defined)
The annual primary energy use of residential appliances is calculated as follows:

Equation 11

Equation 11
where E6 is the annual primary energy use of residential appliances, e6jk is the power use of appliance k in location j, H6jk is the number of households associated with the service being analyzed in location j, h5jk is the annual hours of use per household associated with the service being analyzed for appliance k in location j, and pjk is the primary energy conversion factor for appliance k in location j.

Back to Introduction

Embodied energy and emissions of data center building materials sub-model (i=7)

Embodied energy and emissions are included in the CLEER Model for major materials and devices that comprise present-day and cloud-based systems, both of which include physical materials and IT devices.  The inclusion of embodied energy and emissions allows for consideration of the manufacturing and end-of-life stages for materials and devices that might lead to significant indirect energy and emissions footprints.  The following 9 construction materials and devices are included in the data center building materials sub-model:

Data center building materials (k)
  1. Structural steel
  2. Structural concrete
  3. Extruded polystyrene insulation
  4. Steel electrical conduit
  5. Copper
  6. Steel cooling pipes
  7. Cooling equipment
  8. Transformers
  9. UPS
Annual embodied primary energy use is modeled as follows:

Equation 12
Equation 12
where E7' is the annual embodied primary energy use associated with data center facilities, F7j is the floor space (in square meters) allocated to the data center IT equipment associated with the service being analyzed, I7jk is the materials intensity (kg per square meter) of building material k, EM7jk is the manufacturing primary energy (in energy per kg) of building material k for data center type j, ER7jk is the recycling primary energy "credit" (energy per kg) associated with recycling building material k for data center type j, ED7jk is the primary energy (energy per kg) associated with landfill disposal of material k for data center type j, z7jk is the percentage of material k that is recycled for data center type j, and L7j is the lifespan of the data center facility (in years).


Back to Introduction

Embodied energy and emissions of data center IT devices sub-model (i=8)
Embodied energy and emissions are included in the CLEER Model for data center IT devices to allow the user to consider the manufacturing and end-of-life stages for major data center IT equipment.  The following 5 IT devices are included in the embodied energy and emissions of data center IT devices sub-model:

Data center IT devices (k)
  1. Volume servers
  2. Mid-range servers
  3. High-end servers
  4. External hard disk drive (HDD) storage
  5. Network devices
Annual embodied primary energy use is modeled as follows:

Equation 13

Equation 13
where E8' is the annual embodied primary energy use associated with data center IT devices, N1jk is the number of data center IT device k associated with the service being analyzed (which is specified in sub-model 1), EM8jk is the manufacturing primary energy (in energy per device) of data center IT device k for data center type j, ER8jk is the recycling primary energy "credit" (energy per device) associated with recycling data center IT device k for data center type j, ED8jk is the primary energy (energy per device) associated with landfill disposal of data center IT device k for data center type j, z8jk is the percentage of data center IT device k that is recycled for data center type j, and L8jk is the lifespan of data center IT device k (in years) in data center type j.

Back to Introduction

Embodied energy and emissions of client IT devices sub-model (i=9 for business, i=10 for residences)
Client IT devices refer to electronic devices used to access a digital service such as email, streaming video, or websites for e-commerce.  The CLEER Model currently includes client IT devices that are representative of major devices used by both business users and consumers. Embodied energy and emissions are included for client IT devices to allow the user to consider the manufacturing and end-of-life stages for client IT devices in the system being analyzed.
 
Although separate embodied energy and emissions sub-models are offered for business client IT devices and residential client IT devices, they are described together in this section because the sub-models are identical from a mathematical perspective.  We use i=9 in the equations below for convenience. The following 14 client IT devices are included as elements in the embodied energy and emissions of client IT devices sub-models:

Client IT devices (k)
  1. Desktop PCs
  2. Notebook PCs
  3. Netbooks
  4. Flat panel computer monitors (17")
  5. Tablet computers
  6. Smart phones
  7. Digital music players
  8. DVD players
  9. Stereo systems (CD)
  10. Set top boxes
  11. Flat panel TV <40"
  12. Flat panel TV >40"
  13. Video gaming console
  14. Other (user defined)

Annual embodied primary energy use is modeled as follows:

Equation 14

Equation 14
where E9' is the annual embodied primary energy use associated with client IT devices, N2jk is the number of client IT device k associated with the service being analyzed in location j (which is specified in sub-model 2), x2jk is the fraction of the annual "on" hours of client device k in location j that is dedicated to the service being analyzed (which is specified in sub-model 2), EM9jk is the manufacturing primary energy (in energy per device) of client IT device k in location j, ER9jk is the recycling primary energy "credit" (energy per device) associated with recycling data center IT device k in location j, ED9jk is the primary energy (energy per device) associated with landfill disposal of client IT device k in location j, z9jk is the percentage of clientIT device k that is recycled in location j, and L9jk is the lifespan of client IT device k (in years) in location j.

Back to Introduction

Embodied energy and emissions of network devices sub-model (i=11)
Due to a lack of data on the embodied energy and emissions associated with the many network devices that comprise different network segments, the CLEER Model currently calculates the embodied energy and emissions associated with each segment as a fraction of the direct operational energy use of each segment. 
The following 12 network segment elements are included in the network sub-model based on the definitions in sub-model 4:

Network segments (k)
  1. Wavelength division multiplexing (WDM) segment (terrestrial and submarine)
  2. Core network segment
  3. Metro network segment
  4. Video distribution network (VDN) segment
  5. Asymmetric digital subscriber line (ADSL) access
  6. Hybrid fiber coaxial (HFC) access
  7. Passive optical network (PON) access
  8. Optical point to point (PTP) access
  9. Fiber to the node (FTTN) access
  10. WiMAX access
  11. UMTS access
  12. LTE access

Annual embodied primary energy use is modeled as follows:

Equation 15
Equation 15

where E'10 is the annual embodied primary energy of network data transmission, and E'10k is the primary energy use of network segment k, which is expressed in units of embodied primary energy per unit of direct operational energy of network segment k.  The remainder of the equation is identical to Equation 9, which calculates the annual direct operational energy use of network segment k.


Back to Introduction

Transportation systems energy use sub-model (i=12)

The transportation sub-model calculates the direct fuel use of common modes of freight and passenger transport.  Consideration of transportation activities can be relevant when analyzing the replacement of present-day physical goods transport and personal travel by cloud-based services.  The following 28 transportation modes are currently included as elements in the transportation sub-model:

Freight transport (k)
  1. Diesel fueled barge
  2. Diesel fueled combination truck
  3. Diesel fueled locomotive
  4. Diesel fueled ocean freighter
  5. Diesel fueled single unit truck
  6. Residual fueled ocean freighter
  7. Residual oil fueled barge
  8. Cargo plane

Passenger vehicles (k)
  1. Two seater (Smart ForTwo Coupe)
  2. Subcompact (Chevrolet Spark)
  3. Compact (Toyota Corolla)
  4. Compact hybrid (Toyota Prius)
  5. Mid size sedan (Honda Accord)
  6. Large sedan (Dodge Charger)
  7. Large hybrid (Ford C-MAX)
  8. 2WD pickup truck (Toyota Tacoma)
  9. 4WD pickup truck (Ford F150)
  10. Sport utility vehicle (Ford Escape)
  11. Minivan (Chrysler Town and Country)
  12. Motorcycle

Commercial air travel (k)
  1. Commercial airplane (domestic)
  2. Commercial airplane (international)

Public transit (k)
  1. Intercity rail (diesel)
  2. Commuter rail (diesel)
  3. Transit bus (diesel)
  4. Intercity rail (electric)
  5. Commuter rail (electric)
  6. Transit rail (electric)

The annual primary energy use of transportation activities is modeled as follows:

Equation 16
Equation 16
where E12 is the annual primary energy use of transportation activities associated with the service being analyzed, T12jk is the transportation services demand for mode k in location j (expressed in annual ton-kilometers for freight modes and annual passenger miles for passenger modes), e12jk is the fuel intensity of mode k in location j (expressed in units of energy per ton-kilometer for freight modes and energy per passenger mile for passenger modes), and pjk is the primary energy conversion factor for fuel used by mode k in location j


Back to Introduction

Embodied energy and emissions of other materials and products sub-model (i=13)

The CLEER Model allows for consideration of the embodied energy and emissions associated with other materials and products that might be consumed in present-day or cloud-based systems which aren't covered by the other embodied energy and emissions sub-models.  For example, one may wish to consider the embodied energy associated with the manufacture of DVDs and their cases when comparing video delivery in streaming fashion as opposed to the purchase and viewing of physical DVDs.  Currently this sub-model allows for user-specified materials and values of embodied energy and emissions on a per unit mass basis.  In the future, this sub-model may be populated with specific materials and products that are commonly analyzed when considering cloud-based alternatives. 

The annual embodied primary energy of a user-defined material or product k is modeled as follows:

Equation 17
Equation 17

where E13' is the annual embodied primary energy use associated with other materials and products consumed in the system being analyzed, M13jk is the annual mass of material or product k associated with the service being analyzed, EM13jk is the annual manufacturing primary energy (energy per kg per year) of material or product k in location j, ER13jk is the recycling primary energy "credit" (energy per kg per year) associated with recycling material or product k in location j, ED13jk is the primary energy (energy per kg per year) associated with landfill disposal of material or product k for in location j, and z13jk is the mass percentage of material or product k that is recycled in location j.

Back to Introduction

Variable definitions

Data Center Type IDC taxonomy. Based on data center size and IT and infrastructure characteristics. See Masanet et al. 2011 for details.
Location Location estimates carbon intensity of data center operation. Choose EU country or USA. If USA, choose state or national average.
Carbon intensity of electricity source(g CO2/ kWh source) Based on carbon emissions from primary energy souce of regional electrity (i.e., coal, natural gas, hydro, etc.)
Primary Energy Accounts for source energy (i.e., coal, natural gas, hydro, etc.) losses during electricty conversion and transmission.
Volume Servers IDC taxonomy. Systems with an average sales value < $25,000 (e.g., x86-based systems).
Mid-Range Servers IDC taxonomy. Systems with an average sales value between $25,000 - $250,000.
High-End Servers IDC taxonomy. Systems with an average sales value > $250,000 (e.g., Mainframe RISC Unix systems).
External HDD storage External storage hard disk drives
Data center network devices Installed network devices (i.e., network ports)
PUE Power Use Effectiveness. Ratio of (Total Data Center Energy)/(IT Energy)
WDM Wavelength Division Multiplexing
Internet Core Core Network (gateway to neighboring core nodes)
Metro/Edge Interface between core and access networks
VDN Video Distribution Network
ADSL Asymmetric Digital Subcriber Line
HFC Hybrid Fiber Coaxial
PON Passive Optical Network
PTP Point-to-Point optical
FTTN Fiber to the Node
WiMAX Worldwide Interoperability for Microwave Access (wireless access)
UMTS Universal Mobile Telecommunications System (wireless access)
LTE Long Term Evolution (wireless access)