Energy Efficiency for Telecommunication Equipment: Methodology for Measurement and
Reporting for Router and Ethernet Switch Products
Contents
1
Scope ..................................................................................................................................................... 3
2
Definitions ............................................................................................................................................. 3
3
Normative References .......................................................................................................................... 5
4
Equipment Category Description .......................................................................................................... 6
5
Metric Definition ................................................................................................................................... 6
6
5.1
Introduction .................................................................................................................................. 6
5.2
TEER Metric Definition (representative) ....................................................................................... 7
5.3
TEER Metric Definition (modular) ................................................................................................. 8
5.4
TEER Evaluation............................................................................................................................. 9
Test Procedure ...................................................................................................................................... 9
6.1
Measurement considerations ....................................................................................................... 9
6.1.1
Power measurements ........................................................................................................... 9
6.1.2
Environmental Considerations .............................................................................................. 9
6.1.3
Electrical Considerations ..................................................................................................... 10
6.1.4
Power measurement equipment considerations ............................................................... 11
6.1.5
Power source considerations .............................................................................................. 11
6.1.6
Equipment stabilization considerations.............................................................................. 11
6.1.7
Measurement duration ....................................................................................................... 11
6.1.8
Test configurations.............................................................................................................. 12
6.2
Equipment Configuration ............................................................................................................ 12
6.3
TEER measurements — Modular Method .................................................................................. 12
6.4
Traffic generation/Operational Conditions................................................................................. 14
6.4.1
Traffic topology ................................................................................................................... 14
6.4.2
Use of traffic generators ..................................................................................................... 14
6.5
Measurement procedure ............................................................................................................ 14
6.5.1
Step 1: Qualification ............................................................................................................ 14
6.5.2
Step 2: Full Load .................................................................................................................. 15
1
7
6.5.3
Step 3: Utilization (u2) ........................................................................................................ 15
6.5.4
Step 4: Idle Load .................................................................................................................. 15
Reporting and Documentation ........................................................................................................... 16
7.1
General requirements ................................................................................................................. 16
7.2
Reporting format ........................................................................................................................ 16
Annex A: Classification Examples ................................................................................................................ 18
Annex B : Supplementary Metric ................................................................................................................ 23
Annex C : IMIX Traffic .................................................................................................................................. 25
Annex D: Alternative metrics for routers and switches .............................................................................. 28
Annex E : Modular measurement method for routers and switches ......................................................... 31
Annex F : Sample Reporting Format ........................................................................................................... 32
2
1 Scope
This document provides a set of requirements and guidelines for calculating the
Telecommunication Energy Efficiency Ratio (TEER) of IP routers and switch. The document
also provides standardized definitions of operational data rates and condition to be used when
calculating theTEER of any given configuration. This document is intended to be used by
network operators, equipment manufacturersand re-sellers as a standard method for determining
the energy efficiency of Router and Ethernet Switch products. By comparing the TEER reports
of multiple platforms that meet a common set of requirements, a communications network
operator can select the equipment that best meets their energy efficiency targets.
2
Definitions
1.1 Downlink ports: Group of port facing downstream (access/user side port).
1.2 Uplink port(s): Group of port facing upstream (network/upstream side port).
1.3 IMIX traffic:A stateless traffic profile that contains a mixture of frame sizes similar to
a composition observed in the Internet (Internet traffic mix).
1.4 Maximum Demonstrated Throughput: Highest achievable system throughput at NonDrop Rate(bps).
1.5 Non-Drop Rate: Non-Drop Rate is the observed system throughput atwhich nopacket
drop are recorded (NDR).
1.6 Port Throughput: Rate of traffic (in bps) passing through a port on a sustain basis in
either in either direction, including minimally needed line overhead.
1.7 Port Utilization: Port Throughput expressed as percentage of theoretical maximum.
1.8 System Throughput: Sum of throughput on all system port in the egress direction
(bps)Example: Maximum throughput for 1G E is 1Gbps.
1.9 System Utilization: System throughput expressed as percentage of Maximum
Demonstrated Throughput.
1.10 Tester: Packet generator/analyzer platform.
1.11 Traffic profile: Description of the packet load sent through equipment under test.
1.12 Active mode: It is the operational mode where all ports (WAN and LAN) are
connected.
1.13 Idle mode:
Operational mode with no user data traffic (it is not zero traffic as
service and protocol supporting traffic are present), being used although it is ready to
be used.
1.14 Sleep mode: The state that happens after the device detects no user activity for a
certain period of time and reduces energy consumption. No user facing LAN ports are
connected. The WAN port may be not active. The device will reactivate on detecting
a connection from a user port or device.
3
1.15 Power state:It is a mode of operation with reduced performance and reduced
energy consumption. Power state is a static, not a traffic dependent, mode of
operation. Transition between power states is not instant and may incur a delay,
during which excess traffic might be lost.
1.16 Duty cycle: Duration for each power mode to a specific time period, day, week,
etc.
1.17 Functional unit: The term functional unit is considered a performance
representation of the system under analysis, for example, throughput of the router or
switch measured inGbit/sec. Sometimes the term is used with the same meaning to
indicate useful output or work. ISO 14040 sect 3.20 [Error! Reference source not
found.]
4
3 Normative References
The following standards contain provisions which, through reference in this text, constitute
provisions of this specification.
1. GISFI TR GICT.106, Metrics and Measurement Methods of Telecommunication
Equipments: IP Routers; (Release 1), Sept. 2013
2. GISFI TS GICT.100, Metrics and Measurement Methods for Energy Efficiency: General
Requirements; (Release 2 Draft), Oct. 2013
3. ATIS-0600015.2009, Energy Efficiency For Telecommunication Equipment:
Methodology For Measurement and Reporting - General Requirements
4. ATIS-0600015.03.2009; (07/2009): Energy efficiency for telecommunication equipment:
Methodology for measurement and reporting for router and Ethernet switch products
5. ETSI ES 203 136 V1.0.0 (2013-03) Environmental Engineering (EE); Measurement
methods for energy efficiency of router and switch equipment
6. ITU recommendation: L.1310; (11/2012); Energy efficiency metrics and measurement
for telecommunication equipment
7. ISO 14040:2006, Environmental management − Life cycle assessment − Principles and
framework.
8. ISO/IEC 17025:2005, General requirements for the competence of testing and calibration
laboratories
9. ATIS 0600315.2007; Voltage Levels for DC-Powered Equipment Used in the
Telecommunications Environment; December, 2007.
10. ETSI EN 300 132-2 v2.4.6; Power supply interface at the input to telecommunications
and datacom (ICT) equipment; Part 2: Operated by -48 V direct current (dc).
5
4 Equipment Category Description
This document addresses equipment categorized as Enterprise, Service Provider and
Branch office routers and Ethernet switch products. Small office, CPE, and personal
networking products are outside the scope of this document. The detailed classification of
telecommunication equipment is as per the GISFITS GICT.100 [Error! Reference source not
found.].The present document is applicable to core, edge and access routers as per the
classification. Annexure A of this document provides examples of equipment configurations
intended for reference to assist the user when evaluating products based on ATIS0600015.03.2009[4].
5
Metric Definition
5.1 Introduction
There are fixed and modular configuration systems. A fixed configuration system has one TEER
and follows the test proceduredefined in Clause 6. A modular system consists of a chassis
(shelf) with multiple slots that can be equipped with a variety of cards and/or service modules.
In such application, TEER is a ratio of throughput and energy consumption at different
utilization levels and thus describes relative energy efficiency for a given product [1], [3].
There could be as many TEER reports for a given chassis-based system as there are ways of
configuring it. Furthermore, in order to provide flexibility and scalability, some systems can be
extended by adding (stacking) shelves together.
Depending on the configuration, the number of modules, and the type of those modules,
energy efficiency of a modular packet platform will vary and can be represented with a range of
TEER values.
To reflect this situation, this standard defines two ways for presenting a TEER for a modular
system:
 A TEER for a given family of products, based on a "representative" HW
configuration (as defined by the vendor).
 A TEER for a specific configuration of products.
The TEER value may be either a Certified TEER or a Declared TEER, as defined in ATIS0600015.2009 [Error! Reference source not found.]. For example, a vendor may get a
Certified TEER for a family of products and provide a Declared TEER for a given configuration
to respond to an RFP.
The "representative" method will have a TEER for a fixed configuration (combination of
modules). This metric can be used to compare similar products from different vendors.
The "modular" method allows vendors to establish the energy consumption for each module, so
EE for a desired system hardware configuration (combination of modules) can be calculated by
the user.
NOTE: Providing the module-level data may involve some approximation of
6
throughput and energy usage and therefore can only serve as a guideline.
5.2
TEER Metric Definition (representative)
TEER is defined as a ratio of maximum demonstrated throughput (Td) to weighted power
(energy consumption rate) Pw.
𝑇
𝑇𝐸𝐸𝑅 = 𝑃𝑑
(1)
𝑤
Where:


Td = Maximum Demonstrated Throughput
Pw = Weighted Power (Energy Consumption Rate)
Weighted energy consumption is calculated with the following formula:
𝑃𝑤 = 𝑎 ∗ 𝑃𝑢1 + 𝑏 ∗ 𝑃𝑢2 + 𝑐 ∗ 𝑃𝑢3
(2)
Where:
 Pw= Weighted Power
 (a, b, c) = Weighting for power at each utilization level, where a + b + c=1.0
 (Pu1,Pu2, Pu3) = Power at system utilization level
The traffic profile, weights (a,b,c) and system utilization levels (u1, u2, u3) vary according to
equipment class and position in the network (see Table 1 and 2). The method for
throughput measurement is described in Annex B.
Table 1: Class definitions, TEER calculation parameters, and load profiles for Routing
Products
Class
Representative
Utilization
Access
1-3%
Router
Edge Router 3-6%
Core Router 20-30%
%
of
utilization Weight
forenergymeasurements multipliers
, u1,u2, u3
a, b, c
Traffic
Profile
Simple IMIX
0; 10; 100
a=0.1; b=0.8; c=0.1
0; 10; 100
a=0.15; b=0.75; c=0.1 IPv4/6/MPLS
0; 30; 100
a=0.1; b=0.8; c=0.1
7
(IPv4)
IPv4/6/MPLS
Table 2: Class definitions, TEER calculation parameters and load profiles for Ethernet
Switching Products
Class
Access
High Speed Access
Distribution/Aggregatio
n
Core
Data Center#
Representative
utilization
1-3%
5-8%
10-15%
15-20%
12-18%
A of utilization
forenergy
Weight multipliers
measurements,
u1,u2, u3
0; 10; 100
0; 10; 100
0;10;100
0; 30; 100
0; 30; 100
a, b, c
a=0.1; b=0.8; c=0.1
a=0.1; b=0.8; c=0.1
a=0.15; b=0.75; c=0.1
a=0.15; b=0.75; c=0.1
a=0.1; b=0.8; c=0.1
TrafficProfileSimpl
e IMIX,
Unicast
Ethernet
Ethernet
Ethernet
Ethernet
Ethernet
#
Data Center equipment like switches is not considered within the purview of the current
document.Category of equipment addressed within the provisions of the current document is as
per Clause 4.
Additional metrics that apply to routers and switches that support explicit sleep mode and low
power states are defined in Annex D. These metrics and measurement methods may be utilized
to compare the energy saving performance of access routers and switches that have hardware
support to switch to low power states by turning off relevant parts of the equipment.
5.3 TEER Metric Definition (modular)
TEER for modular packet-based network systems can also be estimated as throughput measured for
components/ modules (Ti) divided by the sum of weighted components/modules power (energy
consumption) (Pwi):
𝑇𝐸𝐸𝑅 =
∑𝑛
𝑖=1 𝑇𝑖
(3)
∑𝑛
𝑖=1 𝑃𝑤𝑖
Where:

Ti= Individual Module Throughput

Pwi= Modular Weighted Energy Consumption
Modular weighted energy consumption Pwi is calculated with the following formula:
𝑃𝑤𝑖 = 𝑎 ∗ 𝑃𝑢1 + 𝑏 ∗ 𝑃𝑢2 + 𝑐 ∗ 𝑃𝑢3
(4)NOTE: Modular energy consumption and throughput are vendor approximations.
When a modular TEER is provided, a representative-configuration TEER shall also be provided
for comparison purposes.
8
5.4
TEER Evaluation
To compare energy efficiencyof products, they shall belong to the same product class. Recommended
product classes are listed in Annex A.
The class description covers the expected applications for EUTs deployed at certain points in the
network. If the system can be deployed in multiple roles, multiple TEER ratings can be
provided.Examples of listing:
1. Medium Core Router, TEER=42 (representative HW/SW configuration follows).
2. Small Edge Router, estimated TEER = 50 (modular configuration based on component/
module ratings).
See Annex A for Classification Tables applicable to IP routers and switches.
Due to a wide variety of features and functions available on the EUT, it is very essential to report
all features and functions active in the test configuration as described in Tables A.1 and A.2.
6 Test Procedure
The general requirements for measuring energy efficiency of telecommunication equipmentare
defined inGISFI TS GICT.100 [Error! Reference source not found.]. In thesub-clause 6.1, the
relevant requirements for measurement and operations of IP routers and switches towards the
determination of TEER have been described.
6.1
Measurement considerations
6.1.1 Power measurements
Measurement accuracy plays a critical role in determining the energy consumed by telecom
equipment at test facilities. Errors due to misconfiguration, insufficient samples can impact the
determination of figure of merit for energy efficiency. The DC power equipment powering the
equipment under test (EUT) should confirm to the electromagnetic noise requirements as
specified by applicable standard [ATIS 0600315.2007 [9] or equivalent BIS specification in
India]. All power measurement equipment used for taking measurements shall be in a current
state of calibration as specified by the applicable requirements [In India, labs accredited by
National Accreditation Board for Testing and Calibration Laboratories or equivalent].
6.1.2 Environmental Considerations
This section provides the applicable environmental considerations for test and measurement of
telecommunication equipment. Environmental conditions for consideration include:
a. Temperature
9
The equipment should be evaluated at an ambient temperature of 25°±3ºC (77oF ±5oF). The
equipment itself should stay online or operate at this air temperature for no less than three hours
prior to the test. No ambient temperature changes are allowed until the test is complete.
b. Humidity
The telecommunication equipment should be evaluated at a relative humidity of 30% to 75%.
c. Barometric pressure
The equipment should be evaluated at site pressure between 860 to 1060 hPa. No targeted
airflows are allowed except for regular ambient room, data center or rack cooling.
6.1.3 Electrical Considerations
This section provides the applicable electrical considerations for test and measurement of
telecommunication equipment. Electrical conditions for consideration include:
a. Voltage supply [DC powered equipment]
DC powered equipment (-48 V DC systems):
Majority of telecommunication equipment (servers, routers, switched, transport etc.) are powered
from central DC power plants. In such systems, the nominal load voltage can be from 50 to 55
VDC at the utilization equipment. For these classes of equipment, the DC voltage powering the
equipment shall be chosen in the range of - 55.5 to -52.5 V (54±1.5 V).
For telecommunications equipment intended to be powered by local DC power obtained from
small AC to DC supplies, testing at -48VDC (± 1%) is recommended.
DC powered equipment (other than -48 V DC systems):
Unless otherwise specified in a supplemental standard to this General Requirements
specification, equipment using nominal DC voltages other than -48 V DC shall be evaluated at ±
2% of the specified voltage.
b. Voltage supply [AC powered equipment]
The input to the equipment (all active feeds) should be the nominal specified voltage ±5% and
the specified frequency ±1%. In case the equipment can work at a different nominal voltage, the
measure shall be executed at one of the nominal voltages. The total harmonic distortion <= 2%
up to and including the 13th harmonic.
Unless otherwise specified in a supplemental standard to this general requirements specification
document, the environmental conditions for the measurements would be as per the applicable
national or international standards [ATIS-0600015.2009 [Error! Reference source not found.],
10
ITU Recommendation L. 1310 [Error! Reference source not found.], ETSI EN 300 132-2
v2.4.6 [10] or equivalent national standard].
6.1.4 Power measurement equipment considerations
It is recommended that integrated power analyzer equipment be utilized for the purpose of
measurements. However, an equivalent setup (voltage and current measurement equipment) with
high sampling frequency and data storage capability is acceptable. Unless otherwise specified in
a supplemental standard to this general requirements specification document, the power
measurement equipment requirements would be as per the applicable national or international
standards [ATIS-0600015.2009[Error! Reference source not found.], ITU Recommendation L.
1310 [Error! Reference source not found.] or equivalent national standard].
Every active power feed should have the power (current) meter installed in the power line with a
desired accuracy not less than ±1% of the actual power level. The power meter should include
correction for power factor (PF) on AC feeds; otherwise, it will be necessary to also record the
power factor in the measurement report. All energy consumption calculations are based on
averaging multiple readings over the course of measurements. Power meters should be able to
produce no less than 100 evenly-spaced readings in every full test cycle duration.
Power measurement instruments (such as voltmeters and ampere meters or power analyzers)
shall have a resolution of 0.5% or better. AC power measurement instruments shall have the
following minimum characteristics:
i.
A minimum digitizing sample rate of 40 kHz.
ii.
Input circuitry with a minimum bandwidth of 80 kHz.
iii.
Capability of accurate readingsof waveforms having a crest factor up to at least 5.
iv.
Power factor correction and reporting.
6.1.5 Power source considerations
Power sources used to provide power to the EUT shall be appropriately over provisioned for any
transient. A minimum of 1.5 times the power rating of the EUT is recommended.
6.1.6 Equipment stabilization considerations
The equipment is to be powered and placed into the relevant operating mode.Allow the
equipment to stabilize in this mode for 15 minutes.
6.1.7 Measurement duration
Measurements are recorded for a minimum specified duration after the EUT reaches a stable
condition of operation. Measure the power for a period of 15 minutes.If the power varies over the
15 minute measurement time interval, an average of the measurement will be calculated.
11
6.1.8 Test configurations
Equipment with multiple power connections (such as those provisioned with redundant power
supplies) shall be configured with all power supply interfaces active and the total power flow
from all these interfaces is computed to obtain the total system power consumption.
Traffic parameters and mode of load generation shall be defined in supplemental standards. As a
general requirement, it is to be ensured that traffic and system is so configured as to exercise all
the required features and functions of the equipment.
The ambient temperature for the measurement is considered to be 25°±3ºC (77oF ±5oF). The
energy consumed by the cooling fans within the EUT may be different at different ambient
temperatures. The information reported in the measurement report should account for the fan
power expected when the equipment operates in a controlled environment of the lab at a
temperature in the range 25°±3ºC (77oF ±5oF). One of the following recommended methods
must be employed to obtain the controlled measurement conditions:
a. Test in a thermally controlled environment with recommended temperature of 27ºC (80.6
o
F)
b. If fans are configurable, they shall be configured with speed settings to simulate the
operating environment of 27ºC (80.6 oF) and barometric pressure 1013.25 hPa at sea
level.
c. If fans are not configurable, a fan speed adjustment must be added to the measured
system power.
6.2 Equipment Configuration
All testing shall be performed on a fully-loaded chassis, as defined by the referenced
application. If there are customer specific applications defining redundancy requirements, they
should be clearly documented in the report.
All ports shall be in an active state and passing or ready to pass traffic.
System software (SW) shall be properly configured prior to the test and all the
necessary HW components installed. HW and SW shall be representative of a production unit.
There is no EUT configuration change allowed any time beyond preparation phase. This
includes (but not limited to) external configuration commands, scripts executing configuration
commands on EUT during testing, etc.Measurements shall be carried out for the appropriate
applicable traffic conditions and mode of generation. Power measurements would be averaged
over the recommended time duration. All energy measurements shall be taken at the main system
power supply unit incorporating all the operational modules.
6.3 TEER measurements — Modular Method
If the vendor chooses to provide the "modular" TEER estimates, it may be required to build more
than one setup if the total number of modules exceeds the number of available slots in a chassis
(or if some modules cannot be used together).
12
In this case, the "base" system configuration is defined as a common system parts, used by all
modules. It may include chassis, fan tray, routing engine, etc.At this time, all other system slots
should be fully populated with "function" modules, not necessary the same, all passing traffic
at the same rate: idle, representative, or maximum NDR.Common system is equipment with no
service card installed and including main processing cards, fan try, power input cards, etc.,
which is used by all the service cards [Error! Reference source not found.].
Each test performed on complete system and then without one module at the time, following
steps in Annex E.The power for each "function" module is the difference between total system
power, with and without this "function" module.
NOTE: Throughput and energy consumption may be affected by interaction between
the system and module under test, so total calculated numbers may be not exactly the
same as in representative test results.
Table 3: Example of Hardware modular system data reporting
Module
Name
Base
system
Module 1
Module 2
Part
number
Maximum
Power (Energy Consumption), W
TEER
u2=-30%
u3=100%
Throughput u1=0%
only)
800-xxxxx- (egress
na
02
800-zzzz- 40 Gbps
200
220
240
N/A
02
800-xxxxx20 Gbps
120
130
150
03
. .
.....
Module N
Total
N/A
500 Gbps
2000W
2100 W
Actual module names may be different for different products.
13
2500W
0-1000
6.4
Traffic generation/Operational Conditions
Traffic topology
If the ports on EUT can be grouped into "network/uplink" and "access/downlink" sides,
according to vendor discretion, then traffic shall be run from every "network" side port to
every "access" side port and vice-versa, thus forming full mesh traffic between two groups.
All streams originated from every port shall be the same capacity.
If all ports on EUT have identical roles, then full mesh traffic with identical capacity streams
between all ports shall be used.
6.4.1
6.4.2 Use of traffic generators
Traffic generators are used to simulate traffic and collect the performance-related results
according to the test conditions. Generators have to be configured for the correct traffic topology
and traffic profile.Traffic of Simple IMIX type as defined in Table C.1 of Annex C shall be
utilized for the measurement purposes.
Uplink Ports
Traffic
Generator
EUT
Downlink
Ports
Figure 1: Example EUT Test Interconnect for two groups of ports
6.5 Measurement procedure
Prior to testing, the EUT shall be configured according to class requirements and offered load
defined in the class requirements (Annex A). Prior to the actual test, the EUT shall be exposed
to environmental conditions outlined in sub-clause 6.1.2 based on GISFI TS GICT.100 [Error!
Reference source not found.].
The procedure consists of four major steps.
6.5.1 Step 1: Qualification
14
The first run determines the maximum load that can be sustained at Non Drop Rate
(NDR).
Any methodology is suitable, including binary search (similar to RFC2544), heuristics,
ATIS-0600015.03.2009
or known maximum load values. There is no time limit for this run. The run is complete after a
maximum (lossless) line rate is determined.
The following three runs should be separated with idle time of 300 seconds .
If the test class requires the EUT to be "primed" with control plane information
(ARP/MAC/route learning, etc.), this shall be completed within the idle time window.
6.5.2 Step 2: Full Load
The second run applies the NDR (identified at step 1) to the EUT for period of 15 min. Power
shall be sampled for the entire period, and average consumption P100 recorded.
6.5.3 Step 3: Utilization (u2)
The third run reduces the line rate to utilization (u2) and runs for another 15 min. Power
shall be sampled for the entire period, and average consumption Pu, recorded. Load reduction is
achieved by reducing the line rate on all configured ports.
Packet loss during any run (if seen) invalidates the measurement and resets testing to the
qualification run to provide a better NDR estimate.
6.5.4 Step 4: Idle Load
Run the EUT idle for another 15 minutes. Power shall be measured for the entire period,
and the average value shall be recorded. Load reduction is achieved by setting line data
rate to 0% on all configured ports.
15
7
7.1
Reporting and Documentation
General requirements
All test reports shall be written according to GISFI TS GICT.100 [Error! Reference source not
found.].Test reports shall comply with the general requirements for testing and calibration
laboratoriesspecified by [Error! Reference source not found.].Additionally, the sub-clause 7.2
lists the basic requirements of the reporting format and Annex F provides a sample reporting
format based on these requirements.
The measurement report must have the following details:
a. Details of supplemental standards applicable to the equipment
b. Time, Date and location of test.
c. Physical and environmental configuration of the test equipment.
d. Model, Serial number, Software and Hardware versions supported on the equipment during
the test.
e. List of features supported by the equipment and those activated during the test.
f. Traffic generation profile and duration of traffic generation.
g. Test and measurement equipment calibration and configuration details.
h. System setup diagram, detailing the electrical and network connections and configurations.
i. Energy or power measurement results for all applicable test profiles as detailed in the
applicable supplemental standards.
j. Start and Stop times for the record of measurements undertaken.
7.2
Reporting format
16
The general requirements for a test report are contained in ISO/IEC 17025 [9]. The following
basic information must accompany the measurement report for energy efficiency of
telecommunication equipment.
a. A Title: Energy Efficiency Measurement Report
b. Name of the equipment under test, Manufacturer detail
c. Equipment category as per this document
d. Applicable standards for energy efficiency measurement
e. Name and address of the measurement laboratory, location where the measurements were
carried out.
f. Unique identification of the measurement report
g. Name and address of the manufacturer
h. Identification of the method used.
i. Description of measurement conditions and unambiguous identification of the equipment
tested.
j. The date of performance of the measurement and generation of report
k. The measurement report with appropriate metrics and units of measurements as described in
the supplemental standards to this document.
l. The name(s), function(s) and signature(s) or equivalent identification of person(s) authorizing
the measurement report.
17
Annex A: Classification Examples
Extract from ATIS-0600015.03.2009 [4]
A.1 Reference Tables
The following tables are intended for reference to assist the user when evaluating products.
Table A.1: Router Classifications
Route
Scale
Class
S,M,L
S
Access
N/A
Logical
Port
Interface Configuration
Scale
Service
Scale
Typically up to T1
worth IMIX traffic
Up to 50 users per
LAN
Up to 500 ACL
entries
Typically up to 4
classes
WAN Optimization:
5
sites/ peers
DLS up to T1
for
Small Branch
Configuratio
n, up
to 50 users
on LAN
Function of
Memory Typically up to half T3/E3 IMIX Traffic
and
CPU; Can IPSec Tunnels: up to
200
go to
ACL Entries: 1000
Internet
M
QoS:
(Full
4-8 Classes
BGP)
WAN
Optimization:
Table
10
with
sites/peers
512MB or
above
18
Configured
Forwarding
Options:
MPLS,
IPv4
Typical Feature Set
Configured Routing
Protocols: BGP, OSPF,
EIGRP, Static Routing
Configured Forwarding
Options: MPLS, IPv4
Function of
Memory
and
CPU; Can
go to
L
Internet
(Full
BGP) Table
with
512MB or
above
Typically up to T3/E3
IMIX Traffic IPSec
Tunnels: up to 500
ACL
Entries: 2000 QoS: 48
Classes or even more
WAN Optimization:
50
sites/peers
Additional
Features:
ALCs,
QoS, Firewall,
IPSec, Voice,
WAN
Optimization,
etc.
Additional Features:
ALCs, QoS, Firewall,
IPSec, Voice, WAN
Optimization, etc.
S
M
BGP: IPv4:
300000
IGP: 10000
Edge
L BGP: IPv4:
1M
IGP: 10000
Core S
IPv4 BGP:
300k
IPv6 BGP:
5k
IGP
Routes:
4000
Multicast
routes: 5k
IPv4 BGP:
IPv6
BGP:
300k
15k
BGP, TMaximum
LDP
customer
Sessions:
VRF Scale: 250
Configured Routing
facing GE
1000
VPNv4: 250k
Protocols: BGP, OSPF,
ports +
Attachmen
Pseudo-wires: 8k
redundant
ts
ISIS, LDP
VPLS Scale: 500
Ccts: 16k uplink 1-GE
TE
ports
Tunnels
BGP,
TMaximum
(head/tail):
LDP
customer
500
Sessions:
VRF Scale: >500
facing GE
Configured Forwarding
1000
VPNv4: 500k
ports +
Optoins: MPLS, IPv4,
Pseudo-wires: >16k Attachmen
redundant
ts
IPv6
VPLS Scale: 1k
Ccts: 16k uplink 1-GE
TE
ports
Tunnels
(head/
Maximum
tail): 1k
10G
IPv4 BGP: 300
Configured Routing
Up
to
16
x
IPv6 BGP: 50
ports +
10G
Protocols: BGP, OSPF,
Subintf: 1000
redundant
TE Tunnels (Mid): 2K
40G core ISIS, LDP, PIM-Multicast
uplinks
Maximum
10G
IPv4 BGP: 500
19
Configured L2VPN
M
24- 72 x
10G
IGP Routes:
8000
Multicast
routes: 10k
IPv6 BGP: 100
Subirttf: 2000
ports
+redundant
Services: VPWS, VPLS,
Inter-working
40G core
uplinks
TE Tunnels (Mid):
5K
Configured Forwarding
IPv4 BGP:
500k
IPv6 BGP:
50k
Options: MPLS PVVE3, IP
Maximum
10G
IPv4 BGP: 1000
L
96- 192 x
10G
IGP Routes:
15000
IPv6 BGP: 200
Subintf: 4000
(GRE), L2TPv3
ports +
redundant
40G core
uplinks
Additional Features:
Multicast
routes: 15k
TE Tunnels (Mid):
10K
ACLs, QoS, Netflow,
EoMPLS, MPLS TE
20
Table A.2: Ethernet Switch Classifications
Class
S
Access
M
L
S
High Speed Access
M
L
Uplink
Number of Count
& Throughput
downlink
Type
ports
Criteria
12 - 50
2-4 Non-Drop rate
1Gbps
1Gbps
= 0.
4 Non-Drop rate
50- 192
1Gbps
1Gbps
=13.
4-8 Non-Drop rate
192- x
1Gbps
1Gbps
= 0.
2 1°Gbps
12- 50
Non-Drop rate
1Gbps
= 0,
4 Non-Drop rate
50- 192
10Gbp
1Gbps
= 0'
s
4- 8
192- x
10Gbp Non-Drop rate
1Gbps
= 0'
s
8- 48
1Gbps
48 - 96
M
1Gbps
96- 192
L
1Gbps
S
Distribution & / or
Aggregation
XL
192- x
1Gbps
8- 16
10Gbps
16- 36
M
10Gbps
36 - 48
L 10Gbps
S
Core
XL
48- x
10Gbps
2
10Gbp
2- 8s
10Gbp
8-16s
10Gbp
s
16- x
10Gbp
s
4
10Gbp
4- 8s
10Gbp
8-12s
10Gbp
s
12- 16
10Gbp
s
Non-Drop rate
= O.
Non-Drop rate
= O.
Non-Drop rate
=

.
Typical Feature Set
IPv4/IPv6 Forwarding, IPv4/1Pv6
Multicast
Snooping, VLAN's, IGMP, MLD,
802.1d/s/w, 802.1q/p, Port
Security/802.1x,
Radius, Mirroring,
802.3AD/LACP, SYSLOG,
IPv4/IPv6 Forwarding, IPv4/IPv6
SNMP1/ 2c/3
Multicast
Snooping, VLAN's, IGMP, MLD,
802.1d/s/w, 802.1q/p, Port
Security/ 802.1x,
Radius, Mirroring,
802.3AD/LACP, SYSLOG,
SNMP1/ 2c/ 3
IPv4/IPv6 Forwarding, OSPF, RIP,
PIM,
OSPFv3, RIPng, Access Control,
IPv4/IPv6
Multicast Snooping, VLAN's,
IGMP, MLD,
802.1d/s/w, 802.1q/p, Port
Non-Drop rate Security/ 802.1x,
= 0.
Radius, Mirroring,
802.3AD/LACP, SYSLOG,
SNMP1/ 2c/ 3
Non-Drop rate
IPv4/1Pv6 Forwarding, OSPF,
= 0.
Non-Drop rate RIP, PIM,
OSPFv3, RIPng, Access Control,
=0
Non-Drop rate IPv4/IPv6
Multicast Snooping, VLAN's,
= O.
IGMP, MLD,
802.1d/ s/ w, 802.1q/p, Port
Non-Drop rate Security/802.1x,
Radius, Mirroring,
= 0.
802.3AD/LACP, SYSLOG,
SNIVIP1/ 2c/ 3
21
Data Center
S 12 - 48
1Gbps
2
10Gbp
s
Non-Drop rate
= 0.
22
IPv4/IPv6 Forwarding, IPv4/IPv6
Multicast
Snooping, VLAN's, IGMP, MLD,
802.1d/s/w, 802.1q/p, Port
Security/802.1x,
Radius, Mirroring,
802.3AD/LACP, SYSLOG,
SNMP1/ 2c/ 3
Annex B: Methods for effective throughput computation
Extract from ATIS-0600015.03.2009 [4]
B.1 Methods for effective throughput computation
B.1.1 Method 1
In first method, the measured throughput is decomposed into a packet-per-second rate and
packet size corresponding to the load. If packet sizes are variable, the average proportions are to
be computed. Next, all applicable minimum overheads are added to compute the effective wire
–rate, at which the EUT performed.


Example 1: The EUT is a packet platform that can drive ten 10Gbps Ethernet ports at
7,291,702 frames per second each with 64B Ethernet frames without loss. According to
the tester, this corresponds to 7,291,702 packet-per-second rate per each port.
Td= 10 x 7,291,702 x 8 x (64+1+7+12) = 49.000237440 Gbps (accounting for Ethernet
start of frame, preamble and minimum interpacket gap).
B.1.2 Method 2
In second method, the traffic generator itself can report port utilization (percent of theoretical
maximum). In this method, the well-known line rates for selected transport interface type are
multiplied by port utilization to calculate the final data rate.


Example 2: The EUT is a VPLS edge platform with 10Gbps Ethernet port (LAN PHY)
on access side (toward Ethernet CPEs ) and 10Gbps Ethernet ports on the network side
(toward MPLS core network). The EUT can forward the incoming L2 frame (256 bytes
each ) toward MPLS core egress interface utilization of 100percent. However, because of
the 1:1 matching of access and network sides, the access side can only be utilized at
99.22 percent to allow lossless application of the (minimally) 4-byte MPLS L2 VPN
header required for packets on the network side. Same limitation is seen in the opposite
direction, where the incoming network-side packet can only fill the access-side interface
at 99.22 percent after the headers are stripped.
Well-known data rate for 10Gbps Ethernet IEEE 802.3ae is 10,000 Mbit/s
Td= 10 x10.000 x 1 +10 x 10.000 x0.9922 = 19.922 Gbps
B.1.3 Other notes
23
Ideal timeout inserted by EUT to compensate for asymmetric test pattern are not counted for
(see Example 2). Also, note that Td should not account for any optional overheads not required
by the traffic profile.


Example 3: The EUT is an Ethernet switch that can operate at 100 percent line utilization
when configured for 802.1q packet encapsulation (VLAN headers applied). The same
switch can only operate at sub-line rate speed when not configured for VLAN
encapsulation (for the same L3 packet payload).
The measurement results from the second case should be used, unless the test profile for
this particular equipment class specifically requires VLAN encapsulation to be present.
24
Annex C: IMIX Traffic
Extract from ATIS-0600015.03.2009 [4]
The reference used for understanding the distribution of packet lengths on the real Internet comes
fromdata collected by the Measurement & Operations Analysis Team of the National Library for
AppliedNetwork Research (NLANR) project during the month of February, 2001 under the
National ScienceFoundation Cooperative Agreement No. ANI-9807479, and the National
Laboratory for AppliedNetwork Research. (The raw data can be found on the NLANR web site
at<http://moat.nlanr.net/Datacube/ >.)
Briefly summarized, a total of 342 million packets were sampled and recorded at the Merit
Networkmonitor site during this period. The average packet size was 402.7 bytes, with the
following packetsizes and types occurring most frequently:
♦ 40 bytes: TCP packets with header flags but no payload (20 bytes of IP header and 20 bytes of
TCP header), typically sent at the start of a new TCP session. Approximately 35% of Internet
packets measured during this period were exactly 40 bytes long. Because these packets are small,
they represent only 3.5% of the traffic.
♦ 576 bytes: TCP packets from old implementations that use this Maximum Segment Size
(MSS).These packets account for about 11.5% of the packets and 16.5% of Internet traffic.
♦ 1500 bytes: Packets corresponding to the Maximum Transmission Unit (MTU) size of an
Ethernetconnection. Most data transferred on the Internet consists of full–size Ethernet frames,
accounting for about 10% of the packets but 37% of the traffic.
Several other packet sizes occurred more frequently than normal (where normal is defined as
morethan 0.5% of all packets). In order of frequency, they are 52, 1420, 44, 48, 60, 628, 552, 64,
56, and 1408bytes.
About 1.2% of all packets were smaller than 40 bytes (e.g., 28, 32 bytes).
They represent a small percentage of overall traffic (0.1%), but a router will nevertheless need to
forward these very small packets.
A realistic mixture of packet sizes can be approximated by a set of packet lengths (three or more)
thatrepresent the common modal lengths, plus an even distribution of every other packet size.
(Thenumber of modal lengths determines the accuracy of the packet mixture -- i.e., its
correlation to a set ofreal measurements.)
Three traffic models are described below: Simple IMIX, Complete IMIX, and Accurate IMIX.
The following Simple IMIX packet size mixtures shall be used:
Table C.1: Simple IMIX
Packet Size (Bytes)
40
576
Proportion of Total
7 parts
4 parts
25
Bandwidth (Load)
6.856%
56.415%
1500
1 part
36.729%
Some router manufacturers commonly use this mixture as a "quick and dirty" approximation of
theInternet packet mixture.
The Simple IMIX has an average packet size of 340.3 bytes and a correlation value of 0.892
whencompared to realistic Internet traffic.
Table C.2: Complete IMIX (Informative)
Packet Size (Bytes)
40
576
1500
40-1500 (range)
Proportion of Total
55.0%
15.0%
12.0%
20.0%
Bandwidth (Load)
5.15%
20.25%
42.20%
32.40%
This mixture retains the simplicity of the Simple IMIX but includes an additional set of packets,
representing all other packet sizes. This set includes a random mix of packet lengths in a
flatdistribution (equal probability of each size), ensuring that some non-zero number of packets
of everysize are offered to the EUT.
The Complete IMIX has an average packet size of 427.0 bytes and a correlation value of 0.985
whencompared to realistic Internet traffic.
Table C.3: Accurate IMIX (Informative)
Packet Size (Bytes)
28
40
44
48
52
552
576
628
1420
1500
40-80 (range)
80-576 (range)
576-1500 (range)
Proportion of Total
1.20%
35.50%
2.00%
2.00%
3.50%
0.80%
11.50%
1.00%
3.00%
10.00%
10.80%
11.80%
6.90%
26
Bandwidth (Load)
0.08%
3.51%
0.22%
0.24%
0.45%
1.10%
16.40%
1.50%
10.50%
37.10%
1.60%
9.60%
17.70%
This mixture has an average packet size of 404.5 bytes and has a correlation value of 0.999 when
compared to realistic Internet traffic.
Table C.4: Additional IMIX Information
Reference
Title
http://pma.nlanr.net/Datacube/Data/TXS/PLen/20010901 Sample internet traffic can be found
/999302444.PLen
at the URL
http://www.spirent.com/documents/4079.pdf
Spirent, Test Methodology Journal,
IMIX (Internet Mix) Journal, March
2006
http://www.ixiacom.com/library/test_plans/display?skey IXIA Library: Test Plans, Broadband
=testing_pppox
PPPoX and L2TP Testing
27
Annex D: Alternative metrics for routers and switches with low power states
Extract from ITU recommendation: L.1310 [6]
This section reports some alternative metrics available for routers and switches. These additional
metrics cover the explicit power states supported by the hardware are based on [Error!
Reference source not found.].
D.1
Routers and switches supporting sleep mode
This metric is applicable only to routers and switches that can go on sleep mode.
The proposed metric is:
TEER = Ti/Pi
[Gbps/W]
(D-1)
Where:
Pi = c*Pmax+ b*Ptypical+ a*Pidle + d*Psleep
[W]
(D-2)
Tiis the weighted throughput
Ti = c*Tmax + b*Ttypical + a*Tidle
(D-3)
(Tmax, Ttypical,Tidle) = Throughput measured at respective utilization levels
Where:
Pmax is the power at maximum traffic load in real time; here maximum traffic load is defined as
maximum non drop rate, equivalent to 100% load (u3 in ATIS-0600015.03.2009 [Error!
Reference source not found.])
Ptypical is the power at typical traffic load in real time; here typical traffic load is defined as 30%
load or 10% load which is dependent on the different equipment types (u2 in ATIS0600015.03.2009 [Error! Reference source not found.])
Pidle is the power in idle state in real time; here idle state is defined in 0% load (u1 in ATIS0600015.03.2009 [Error! Reference source not found.])
Psleepis the power in sleep mode in non-real time, applicable only for equipment that offers sleep
mode.
c is the weighting multiplier for maximum state in real time,
b is the weighting multiplier typical traffic load in real time,
a is the weighting multiplier for idle traffic load in real time,
28
d is the weighting multiplier for sleep mode in non-real time,
a+b+c+d=1
The parameters a, b, c, d values, considering the traffic distribution during the day, are defined in
Table D.1 for routers and in Table D.2 for switches equipment.
Values for a, b, c, and d shall be substantiated with data; sleepmode can be used in a limited
number of networking devices, but only if nothing is attached to the device network connection.
For this group of routers/switches, expected traffic is close to idle.
Table D.1 − Weight factor definition for router equipment
Class
Representative
utilization
Access router 1-3%
with
sleep
mode support
% of utilization for Weight multipliers
energy measurements, u1,
a, b, c, d
u2, u3
0; 10; 100
a=0.15,
d=0.45
b=0.25,
c=0.15,
Table D.2 − Weight factor definition for switches equipment
Class
Representative
utilization
% of utilization for energy Weight multipliers
measurements, u1, u2, u3
a, b, c, d
Access switch
with sleep mode
support
1-3%
0; 10; 100
D.2
a=0.15, b=0.25,
c=0.15, d=0.45
Measurement methodology
Measurement methodology shall be in line with ATIS-0600015.03.2009 [Error! Reference
source not found.], and the general requirements as per GISFI TS GICT.100 [Error! Reference
source not found.], and the sleep mode.
Power measurement for sleep mode
With each of the equipment’s ports operating in sleep mode for 20 minutes, record the average
input power over 15 minutes.
29
D.3
Routers and switches supporting explicit power states
To evaluate EENRT, we define three measurement points that may correspond to different power
states of EUT:

S0 - full performance

S1 - 30% performance

S2 - 10% performance
We also define sample duty cycle as a fraction of time during which the planned traffic levels are
applicable. Level 0 will be used for 55% of duty period, Level 2 for 25% and Level 3 for 20% of
duty period.
EENRT = (0.55TS0 +0.25TS1 +0.2TS2) / (0.55PS0 +0.25PS1 +0.2PS2)
Where:
TS0, TS1, TS2 is the throughput in the three measurement points
PS0, PS1, PS2 is the power in the three measurement points
30
Gbps/W]
(D-3)
Annex E: Modular measurement method for routers and switches
Based on GISFI TR on IP Routers GISFI TR GICT.106 [1]
Step 1: Configure the system with all service cards to obtain a full chassis and set the system at
the maximum NDR traffic load defined in Clause 6.5.1 within the cards, and test the power used
PT.
Step 2: Remove one service card and test the power used of this configuration P1.
Step 3: Calculate the power used by the service card:
PCard1 = (PT - P1)
(E-1)
Step 4: Repeat steps 1 and 2 with other traffic load
Step 5: According to above steps, test the power used by other service cards, PCard2, PCard3, PCardn .
Step 6: Calculate the power used by the common system; this power is determined by subtracting
the power used by different service cards present in the system from the power measured in the
initial configuration during the step 1 at maximum traffic P1.
𝑃𝑐𝑜𝑚𝑚 = 𝑃𝑇 − ∑𝑛1 𝑃𝑐𝑎𝑟𝑑𝑖
(E-2)
𝑃𝑐𝑎𝑟𝑑𝑖 is the power used by service𝑐𝑎𝑟𝑑𝑖 identified during the Step 3.
Step 7: The weighted power of the card is calculated using the following formula:
𝑃𝑤𝑐𝑎𝑟𝑑𝑖 = 𝑎 ∗ 𝑃𝑢1 + 𝑏 ∗ 𝑃𝑢2 + 𝑐 ∗ 𝑃𝑢3
(E-3)
𝑃𝑤𝑐𝑎𝑟𝑑𝑖 is the weighted power of card 1…n
𝑃𝑢𝑗 is the power of the estimated configuration with traffic load determined by Table 1 and 2
under Clause 5
𝑎, 𝑏, 𝑐are the weighted factor depending on traffic load and equipment type see Clause 5
The estimated TEER is :
𝑇𝐸𝐸𝑅 =
∑𝑛
𝑖=1 𝑇𝑖
(E-4)
𝑃𝑐𝑜𝑚𝑚 + ∑𝑛
𝑖=1 𝑃𝑐𝑎𝑟𝑑𝑖
Where:
𝑇𝑖 is the throughput of any individual service card𝑖used to determine the𝑃𝑐𝑎𝑟𝑑𝑖 .
31
Annex F: Sample Reporting Format
Extract from GISFI TS GICT.100 [2]
Table F.1: Sample measurement report format. [Report must be inclusive but not limited to the
relevant mentioned fields]
Equipment under test
Name of EUT
Manufacturer of EUT
Unique Identification of EUT
Equipment Category
Applicable standards for energy
efficiency of the EUT
Date of Testing
Location of Testing
Date of Report Publication
Hardware
Description:
Dimensions:
Length
Breadth
Configuration mounting:
Fan compensation power notes:
Modification required:
Software
Firmware version:
Boot loader version:
32
Height
Operating system details:
Configuration notes and software tuning parameters
Configuration Notes:
Software tuning parameters:
Enabled features:
Disabled features:
Annotated picture of equipment under test
<INSERT PICTURE/BLOCK DIAGRAM OF THE SETUP>
Description of setup
Environmental conditions
Criteria
Specified conditions
Test duration
Ambient temperature
Relative humidity
Atmospheric Pressure
Feed voltage
Supply power ratio
33
Test conditions
Energy Analyzer
Hardware vendor
Model
Serial number
Calibration lab
Accredited by
Calibration level
Calibration date
Calibration validity period
Support Equipment (s)
Hardware vendor
Model
Serial number
Calibration lab
Calibration date
Power feed sizing
Minimum
Maximum
Total power feeds
34
Feed redundancy
Calculations
Any other notes
Contact Information for this report
Name
Designation
Contact Address
Contact Telephone
Email ID
Website
Place
Date
Signature
35
Contributors to this document
At the time of release of this document, the TEC Core Committee on Green Passport
Implementation Plan, which was responsible for it development, hadthe following roster:
Ram Krishna
DDG(FLA)
TEC
AnupamGupt
DDG(GP)
TEC
Laxmi
Director (FLA)
TEC
Members of the Core Committee
Organization Represented
Name of the representative
NEC / GISFI
Ritesh Kumar Kalle
NEC / GISFI
Anand R. Prasad
CISCO / GISFI
ArvindMathur
I2TB/ GISFI
Krishna Sirohi
VNL
Ericsson
NSN
ETSI
Spirent
Vodafone
Airtel
Reliance
AUSPI
36