ETNO Paris 2006-12-14 Telecom NW Power Consumption

1
ETNO Paris2006-12-14 Telecom NW Power
Consumption

• Tomas Edler
• Business Unit Access
• Ericsson AB
• tomas.edler_at_ericsson.comSourceJens.malmodin_at_er
  icsson.com

2
Content Digital Power Part omitted

• LCA, Telecom systems Life Cycle Assessment
• LCA method
• Fixed Mobile NW results
• Way Forward upcoming LCA on Fixed NW
• Power Consumption and Efficiency of Teleocm
  Equipment
• Code of Conduct versus standardization.
• ETSI EE/EEPS contributions. Power efficiency as a
  base for benchmarking of performance
  improvements.DSL efficiency figures based on
  simulations.
• Example. Traffic and site models applied to EC
  CoC DSL equipment power targets. Operator and
  Subscriber Power costs.

3
LCA Method
4
LCA - Life Cycle Assessment
Operation (e.g. coal based electricity production)
Ericsson Kista (xxx)
Product Life Cycle
End-of-life treatment (This is a worst case
example from China)
Manufacturing (e.g. copper mining and smelting)
TransportsTravel
5
LCA explained on one slide...
An LCA is a very iterative process

• Physical description - studied product/system,
  system boundaries, functional unit - sets the LCA
  start parameters.
• LCI, Life Cycle Inventory - Starts with your
  product/system, ends with a lot of data for
  resource consumptions and emissions from all life
  cycle phases and activities associated with your
  product/system.
• LCIA, LC Impact Assessment - Takes on where the
  LCI stops. Translates LCI end data into
  environmental impact data.
• LCSEA, LC Stressor-Effects Assessment - A more
  advanced LCIA, sets further requirements on LCI
  data.
• Results interpretation - The LCA results can be
  summed accross environmental impact categories,
  life cycle phases, parts of the product/system.
  An LCA gives complex results and a lot of time
  must be spent on understanding and presentation.

The truth
(An LCA is a very resource hungry method... But
to come near the truth is better than just
guess. And life cycle thinking is the key, with
that, you can come a long way.)
X
Your results
6
1. General LCA Model ? Total Site View
Emissions to air (ex CO2, NOX)
Input of energy and energy carriers
Use of land resources
Emissions to ground (ex RHW)
Transports
Input of products

• Model examples
• (1 model built from
• several suppliers
• 2 ICs
• PCBs
• 9 component fam.
• LCDs
• Batteries
• PBA processes
• 3 diff. mech. proc.

Site model
Output of product(s)
Input of ancillary products
Use of natural resources (ex Crude Oil)
Output of waste (incl. recycling)
Emissions to water (ex Cu)
7
2. Life Cycle Inventory - LCI
World average 3G system, annual
operation Original green field system end 2001
New design
42 GWh (mid 2005)
28 GWh (end 2006)

• Other important
• parameters
• Traffic model
• Geography
• CapCov
• Climate
• Site types

8
3. LCIA, Life Cycle Impact Assessment Climate
Change (CC) CO2-eqv
(RBS end 2001)
kg CO2-eqv. per subscriber and year
Operation
Manufacturing
Traffic Model New RBSs
RBS mid 2005
Terminal
End-of-life treatment
RBS end 2006 Estimate
Service Equip. Terminal
Comercials Service Equip. Transmission
Terminal
Reduction of operator per subscriber
Site material important
Transports no. 1 at E.
Recycling
9
4. LCSEA (Stressor-Effect Assessment)
Seven regions studied Sweden, UK, EU, Japan, US
average, US Texas and Taiwan. A unique ECF
(Environmental Characterization Factor) is
calculated for each region and each environmental
impact category
..a World average is also calculated for each
environmental impact category.
10
5. LCA Result Interpretation3 viewsGlobal,
Branches Ericsson
11
Energy, Material, LandTelecoms share - Global
Energy (CO2) Phone books Node sites
User equipment Fixed Cables Antenna towers
Energy ? Material Land CO2-eqv. resources u
se World Telecom 0.4 0.05 0.2 (50ppm) 3G
(LCA, 50) 0.2 0.03 0.1 (20ppm) Total
telecom revenues 1 500 billion , 4 compared
to World GDP (at 36 000 billion ). Operators 1
100 billion . ) Add on scenario of 50
penetration 3,2 billion subscribers
According to WWF analysis Fossil energy
biological footprint
Ericsson original LCA results in ppm
12
, kWh and CO2 Overview
13
Ericsson View Use Phase CO2Produced products
2006 and life time in operation 10 years,
unless otherwise stated
Mton CO2
For each typical product annual volume typical
electricity consumption est. life time (10 yrs.
87660 h) 0.6 kg CO2/kWh
-7, Grey bar Scenario same production volumes
but with only the older product generation.
Supply chain CO2 4 Mton raw materials, manufactu
ring
Ericsson direct CO2 0,8 Mton transports, sites,
travel
Refers to manufacturing of chip sets and not use
-27
15 yrs.
Dotted line whole terminal
(3 yrs.)
Other life cycle phases for comparison, Ericsson
total
Ericsson total use phase CO2 2006 21,5
Mton Reduction for 2006 (WCDMA GSM) -2,2 Mton
CO2 (-10)
14
5. LCA Result InterpretationCO2 Use Phase
Energy5 examples.ICT, Telecom BB
15
Fixed Wireless kWh/sub, lineOperator overview
1) Offices included in original figure have been
withdrawn by -15 (based on other operators) 2)
Besides offices, a split between wireless and
fixed have also been made based on subscriber
averages
16
ICT/Broadband Energy Map W/subW / Heavy User.
Based on US figures.
1 PC at home, 1 at the office (Trp Net
impact) A worst case scenario but with average
use figures 2006, there was 1 000 million PCs,
160 million broadband connections and LAN-PCs is
estimated to about 200 million
Transport 8 W 4W/PC
The Net 20 W LANs, Intranets, Internet 10W/PC
PC 28 W
PC 28 W
TOTAL 97.6 W 855 kWh, equal to 55 of an oil
barrel per year or
425 kg CO2 / year
17
Telecom ICT Energy MapW/Average User
At home At the Operator At
the office User Equipment Access
Network Transport Network The Net
Work PC
GSM RAN 2 W incl. Mobile Core /subscriber
WCDMA RAN 4 W incl. Mobile Core /subscriber
PC 28 W
Wireless phone 5 W (IP-phone 4 W)
18
Study of Broadband in Japan
CO2-equivalents

• Approx. 275 kWh operation electricity per BAP and
  year (Broadband Access Point)
• Equal to 2 of electricity in Japan, equal to
  0,7 of energy (2005, 10 MBAP)
• 200 kg CO2 per BAPyear
• Modemnetwork always on...
• The PC in the study are the internet part of a
  PC (?)
• Operators activities also included (offices)

?
BB operation fraction
Two other studies of the Internet in the US
give results of total energy consumption of about
2-3 of US total.
19
PSTN Energy Map (based on LCA)
Energy Consumption
20
CO2, Energy Mtrl Reduction over time3
examples
21
PSTN Energy (based on LCA)
22
Total CO2 Wireless EvolutionCO2 / subscriber and
year, based on LCA results
1st generation mobile systems NMT, AMPS
200
2nd generation D-AMPS, GSM (90 kg)
100 kg CO2 / subscriber and year
First 3G systems (55 kg)
New 3G system (37 kg)
50
RBS 884/2000 (45 kg)
3G 2006 est. (29,5 kg)
GSM Hi-cap RBS (33 kg)
LTE?
1985 1990 1995
2000 2005
2010
Terminal breakpoint? (gt RBS?)
New GSM (25 kg) - full traffic model
25 kg CO2 equals driving 125 km - 1 hour on the
motorway, or 45 kWh global electricity (LCA
600 g/kWh)
23
Less material and energy per function, Switching
Center (MSC)
Typical 500 W / cabinet
1992 2400 E
2006 32000 E
2002 18000 E
Typical 1000 W / cabinet
24
LCA Findings

• Extensive database with a number of
  models/studies
• Manufacturing of electronics
• ICs and PCBs needs a lot of energy and chemicals
  during production (dont let their low weight
  share fool you)
• Cooling need - for our products in use and in
  offices
• Transports Travel
• Building overhead (Offices, Malls,
  Warehouses)
• These buildings has not gone through the same
  dematerialization process as our products
• Mobile terminals and PCs
• Its not the use phase, rather the manufacturing
  phase that require most of the energy. One
  important reason is of course the short
  commercial lifetime.

25
LCA and the Future

• Conduct a full scale LCA of broadband, with
  particular emphasis of total energy consumption.
• Operator cooperation imperative
• Update the fixed and wireless studies with new
  models and data
• In the end, study ICTs impact on society now and
  potential impact in the future, on a per service
  base

26
Content

• LCA, Telecom systems Life Cycle Assessment
• LCA method
• Fixed Mobil NW results
• Way Forward upcoming LCA on Fixed NW
• Power Consumption and Efficiency of Telecom
  Equipment
• Code of Conduct versus standardization.
• ETSI EE/EEPS contributions. Power efficiency as a
  base for benchmarking of performance
  improvements.DSL efficiency figures based on
  simulations.
• Example. Traffic and site models applied to EC
  CoC DSL equipment power targets. Operator and
  Subscriber Power costs.

27
EC Code of Conduct, Ericsson view

• EC initiative, Code of Conduct for BB equipment.
• EC view - Pace of industry driven improvements
  too slow
• Non transparent process.
• Measurement conditions to be defined.
• Proving compliance to CoC is demanding much
  more than technical compliance like reporting,
  supporting initiatives etc
• Relation to other EC initiatives like EUP -
  unclear
• ETSI standardization
• Democratic process open to all parties
• Technical competence and experience as needed to
  provide accurate specifications. Taking
  end-to-end considerations into account.
• Speed is an issue but aggressive time plan for
  the Broadband power consumption TS.
• Compliance process is efficient.
• Ericsson view
• Recommending ETSI standardization. Ericsson will
  contribute to ETSI EE/EEPS WIs for an early
  and accurate standard.
• Ericsson will comply to operator requirements, EC
  CoC as well as ETSI standards.

28
Energy EfficiencyStandardization

• ETSI proposals
• Energy efficiency modelling
• Reference models, interfaces
• Operational conditions, traffic models and energy
  efficiency
• Energy eff. Examples DSL simulated values
• Example BB
• Estimated traffic models, operational modes
• Using models on CoC BB target figures.
• Substantial savings if low power modes were
  operational.
• Way Forward

29
Energy EfficiencyStandardization

• ETSI proposals
• Energy efficiency modelling
• Reference models, interfaces
• Operational conditions, traffic models and energy
  efficiency
• Energy eff. Examples DSL simulated values
• Example BB
• Estimated traffic models, operational modes
• Using models on CoC BB target figures.
• Substantial savings if low power modes were
  operational.
• Way Forward

30
ETSI proposals TR BB power consumption

• Need to compare different products on fair
  grounds base for bench-marking
• Move from power consumption to energy efficiency
• A number of terms and models proposed
• Define the useful output.
• Distance and bitrate important factors.Example
  Bitrate Mbps x distance km
• Define consumption
• Equipment/site reference models. Site view based
  consumption
• Climate models
• Traffic/User models
• Energy efficiency proposalNPC Normalized Power
  Consumption per subscriber line

NPC Avg Power Consumption/Useful
OutputmW/Mbpskm. Typical best value 40 - 50.
31
Power Efficiency model, Telecom NW
ACInput
Network model
Node A
. .
Node B N
UsefulOutput
Node B 1
FieldData

NW Energy EfficiencyModel
Traffic Model
Useful Entity
Climate Model
Site Model
Product Data - Equipment
32
DSLAM Site
Enclosure
Split-ter
DC2
DC3
DC1
DSLAM
AC1
AC2
Recti-fier
. ..
-48V
. ..
. ..
ClimateUnit
POTS/ISDN
ACInput
. ..
Battery
LineInput /OutputSignal
3pp/Aux Eq
Site factor DC3 to AC1 1,7 (Rectifier Climate
AC part)
33
RBS Site
FeederCable
An- tenna
Enclosure
DC1
DC2
DC3
AC1
AC2
Recti-fier
RadioBaseStation
-48V
. ..
ClimateUnit
ACInput
Battery
Out-putSignal
RF1
RF2
3pp/Aux Eq
Site factor DC3 to AC1 1,6 (Rectifier Climate
AC part)
34
DSLAM Performance / Energy Efficiency
Bitrate UL/DL /Power cons
Bitrate ULDL/NPC
VDSL2T2
ADSL2T1
35
Energy EfficiencyStandardization

• ETSI proposals
• Energy efficiency modelling
• Reference models, interfaces
• Operational conditions, traffic models and energy
  efficiency
• Energy eff. Examples DSL simulated values
• Example BB
• Estimated traffic models, operational modes
• Reference models for NW Site AC consumption.
• CoC target figures used
• Using traffic models on CoC target to estimate
  savings.
• Substantial savings on L0 mode power reduction.
• ? Additional, substantial savings if low power
  modes were operational.
• Way Forward

36
DSLAM Operational Modes user traffic models
L0
Today
L2
lt 5 min interrupt
L3
5 - 30 min interrupt
Stair- Way Forward?
gt 30 min interrupt
37
EC CoC Power Req
DSLAM Site
Enclosure
Split-ter
DC2
DC1
DC3
DSLAM
AC1
AC2
Recti-fier
. ..
-48V
. ..
. ..
ClimateUnit
POTS/ISDN
ACInput
. ..
Battery
LineInput /OutputSignal
3pp/Aux Eq
Site factor from DC3 to AC1 1,7 (Rectifier AC
Climate)
38
DSL-Modem Operational Modes user traffic models
ON
Today
lt 30 min interrupt
Std By
Transition
OFF
gt 30 min interrupt
Stair- Way Forward?
Manual
39
Modem End User
DC
DSL-Modem
ACInput
Recti-fier

Split-ter
Phone
LineInput /OutputSignal
EC CoCReqs
40
EC CoCPower Consumption Target Values, 0607 Rev.
41
DSLAM Site AC costsAverage traffic model. L0
only and L0-L3 modes.
Euro/Sub, Y 9kWh/Euro
42
DSL-Modem AC costsAverage traffic modelOn only
and ON/StdBy modes
Euro/Sub, Y
9kWh/Euro
43
Way Forward

• Move on - reducing power consumption of BB
  equipment at L0 modes. Provide an ETSI
  Specification/Standard.
• However - L2-L3 CoC specs has no impact today!
• DSLAM L2-L3 and Modem Stand By not operative.
• Make DSLAM L2-L3 and Modem StdBy modes operative.
  Do what is needed
• Field testing, disturbance simulations and
  measurements,
• Standardization
• Implementation
• Improve power consumption operational models
• include site models, traffic models adding low
  power mode power impact - when the modes are
  operational(Ev nämna stör-studier som
  finns/föreslagits av operatörer)
• Your feedback? I appreciate your comments.

44
(No Transcript)
45
2005 Relative Results / Normalization
0
100
Raw materials - EoLT
0,060
Manuf.
Operation
0,55
0,54
0,39
0,42
0,35
0,072
0,18
No data available (no world average) 0,011
World EoLT- Scenario
No data available (no world average) 0,15
0,00046
0,0032
Operator
100
0
1
of total impact (per capita)
EoLT and Supply Chain aspects most uncertain
46
Materials, Energy and GDP
Change 1935-2005 (/capita) Food
35 Materials 170 Energy 250 GDP
400
We need an energy consumption reduction as strong
as this (or even stronger), or a complete new
energy system...

• World 2005 compared to 1935 Food production x4,
  Material production x8, Energy production x10.5,
  GDP x14.5Population x3 (2.2 ? 6.5 billion
  people)
• Materials where slowly decoupled from GDP growth
  in the late 60s and energy in the mid 70s

47
, kWh and CO2 Overview
(Revenue and GDP can be compared but are not the
same thing)
1 2 3
1) PSTN and broadband consumes much more energy
then wireless. 2) PCs, Internet datacom
office equipment consumes much more then just
telecom. 3) TV and physical media (distribution
production of paper products, CDs, DVDs)
consumes even more then just ICT. 4) About 50
of Buildings energy / direct CO2 comes from
households, making households share to about
25. ) Air travel, raw materials, semiconductors
and food production have their indirect GHG
emissions (e.g. CH4, NOX) figure in
paranthesis. Because of this, all figures becomes
a little bit lower, but only energy is
recalculated.
48
Telecom/ICT and the Watts
Laptop / PDA B of 1 - 100 W
28 W
PC / Laptop A of 30 - 300 W
PSTN phone(s) 0 - 5 W
Mobile terminal 0.25 - 0.6 - 2.5 W
PBX terminal 0 - 1 - 3 W
Home network 5 - 9 - 20 W
2 W
GSM RBS 0.8 - 1.8 - 3 W
WCDMA RBS 1.6 - 2.5 - 10 W
PBX 0.5 1.1 W
Telestation 4.5 W (Telia SWE) PSTN, ISDN BB
(xDSL, Fiber)
Mobile core network 0.2 W
Core network / Transport network 1 W (Telia
SWE)
Can be distributed according to traffic volume
-9W- (both)
LAN / Intranet
Internet / Data centers W Z of 5.4 W
US studies 42 W ICT total electricity or
embodied energy 50 liter oil / year (includes
manufacturing)
49
Direct CO2 - Employee View

• Fewer manufacturing sites (outsourcing)
• Less air travel!
• Car travel includes commuting
• Electronic equipment future candidate (manuf.)

Ericsson direct-bar (from slide before)

• With transports I/O
• Greater product volumes
• Fewer finalassembly sites
• Outsourcing