Transcript Energy Use Implications of ICT Hardware (Andrius Plepys)

Energy Use Implications of ICT
Hardware
by Andrius Plepys
NATO SCIENCE PROGRAMME
in conjunction with the Carnegie Bosch Institute
ADVANCED RESEARCH WORKSHOP:
Life Cycle Analysis for Assessing Energy and Environmental
Implications of Information Technology
Budapest, Hungary
September 1-3, 2003
Why the issue?





Dynamism of ICT sector
Productivity and structural impacts
Role in sustainable development
Climate change policies
Energy security
Energy and the New Economy
 Decoupling between GDP growth and
energy consumption often attributed to ICT
sector
Period
Reduction rate of
energy intensity
1960-96 (“old Economy”)
1.3 %
1973-86 (oil crisis)
2.6 %
1996-99 (the New Economy)
3.4 %
1996-01 (N.E. incl. IT stock crash)
2.8 %
Source
(EIA 2001)
(Laitner 2000)
(Laitner 2002)
Electricity crisis – a hoax or a reality?
New York, August 15, 2003
Internet to blame for the blackouts?
El. consumption dynamics in Silicon Valley and California, 1990-2000
Electricity consumption increase
Silicon Valley
Rest of California
Residential consumption
19.6 %
18.1 %
Non-residential consumption
15.4 %
15.1 %
Total
16.5 %
16.0 %
derived from California Energy Commission’s data (2002)
 Interests of power suppliers (coal industry)
 Poor planning and artificial price increase?
 From Mills to LBNL
 National estimates of AEC
ICT-related electricity consumption as % of national AEC
Country
USA
Japan
Germany
2000
2010
Sources
high
2.7
2–4
ADL (2002)
low
2–8
2 – 50
medium
3.3
4
low
4.3
30
4
n.a.
Aebischer* (2003)
0.5 – 1.7
6
Barthel/Turk (2001)
< 3-4%
<5-6% ?
Reliability
medium
low
Bottom line
several studies
NTT/FRIC (2002)
ISTEC (2000)
However…
 Absolute consumption will increase
 Future predictions are fuzzy
 Reportedly large energy saving potential
Electricity consumption by
component in non-residential sector
UPSs
6%
Other*
10%
Monitors &
displays
22%
Printers
6%
Data
networks
7%
Telecomm
networks
7%
Copiers
10%
PCs &
workstns.
20%
Servers
12%
Source: Roth et al. in ADL (2002)
The power of power management
 CPU – idling >90% of the time
 Hardware actively used <25% of the time (Webber, 2001)
 PM already saves 25%, but additional 15% could be
saved by optimal set up (US EPA, 2002)
 Largest saving potential in offices:

desktop computers/workstations

CRT monitors

Copiers & printers (Kawamoto et al., 2001)
The two legs of power management
Technology
solutions
Behavioural
solutions
- software (BIOSOS)
- awareness
- products (CRTLCD)
- knowledge
- components (CPU)
- informed choice
Relevancy of the issues - DC example
 High power reliability costs dearly
 Overestimated needs
 Consumption insignificant on national scale, but a
large share of ICT infrastructure
 HVAC – largest consumer DC’s energy
 Saving 20-40% technically feasible today

HVAC optimisation (airwater, CHP, to)

night switching
 Economic barriers (large build-up, risk aversion)
The impacts of trends
 Wireless communications
 Mobile devices
 LCD displays
 ICT diffusion into other products
 Optic fibre – broadband – data traffic
 The “last mile” limitations
 Voice and data n-work convergence
 E-services
Shortcomings
Methodological and data issues:



ICT definition and system boundaries
Allocation procedures
Data:



Behavioural data (!)
Power rating
Stock data and return rates
Reflections
 Electricity consumption – not significant today, but
future is uncertain
 Growth rate and saving potential makes it
important for continuous research
 Supply side: energy efficiency not always a design
priority (often a trade-off with costs)
 Demand side: marginal role of energy costs to
encourage savings (hardware costs, performance,
ergonomics before environmental considerations)
Reflections
 Technology can take care of some
efficiency improvements
 Behavioural changes are needed to fully
exploit the potential savings
 Market “failure”?
A role for policy makers?
 Economic instruments (e.g. green taxes)
 Informational voluntary instruments (performance
standards, labelling initiatives)
 Governmental procurement for more energy
efficient equipment
 more research on policy role