strategic research agenda - ict-energy · domestic wifi router, mobile phone, led bulb smartphone,...
TRANSCRIPT
Strategic Research Agenda
Prof Douglas Paul
Director: James Watt Nanofabrication Centre University of Glasgow, U.K.
What is a Strategic Research Agenda?
Roadmap and links to other appropriate roadmaps
> 30 page document with executive summary
Strategic objectives / recommendations to drive common research objectives
Comparison of different approaches
Overview / brief introduction of field (laymans version - everyone can understand)
Ultimately used by Brussels to aid what should / could be funded in the future
EC 2020 CO2 Emission Target
106
107
1990 1995 2000 2005 2010 2015 2020
CO
2 (
Mil
lio
n k
g p
er y
ea
r)
Year
World
China
EU28
USA
2020
20% target
ICT Global Energy Consumption
0
1
2
3
4
5
2000 2005 2010 2015 2020
% o
f Tot
al A
nnua
l Ele
ctric
ity P
rodu
ctio
n
Year
Total ICT energy
End-user devicesCommunications
Data centres
This excludes TV, media, publishing, games, power switches, domestic & industrial ICT devices
=> 10% to 15% of electricity & ~ 5 % CO2 emissions
1.0
1.5
2.0
2.5
3.0
2000 2005 2010 2015 2020 2025 2030IC
T C
O2e E
mis
sio
n %
Year
Microprocessor Power Density
1.0
10
100
1970 1980 1990 2000 2010 2020
200
400Po
wer
den
sity
(W/c
m2 )
Year
PWR nuclearreactor core
0.5
PC MPUs
Portable MPUs
Intel prediction
Microprocessors & ICT Devices
10-2
100
102
104
106
108
1940 1960 1980 2000 2020 2040
To
tal
po
we
r c
on
su
mp
tio
n (
W)
Year
HPC
PC MPUs
Portable MPUs
105
106
107
108
109
1010
1980 1990 2000 2010 2020 2030
Num
ber
Year
World populationMobile phone sales
PC sales Tabletsales
Smartphonesales
TV sales
without history
Devices Energy Consumption
TOPOL SETs & SATs
Internet and Telephone Traffic
10-3
10-1
101
103
105
107
1990 2000 2010 2020 2030
Inte
rnet
Tra
ffic
(PB
/mon
th)
Year
Total
Fixed
Managed
Mobile
Ciscoforecasts
Interconnect ~ 0.1 fJ to 100 pJ / bit
Photonics (long distance) ~ 1 µJ / bit
Wireless ~ 10 nJ / bit.m
Photonics (rack) ~ 40 pJ / bit
Wire limit ~ 3 fJ / bit.m
ICT Energy Harvesting Requirements
10 MW
1 MW
10 nW
100 nW
1 µW
10 µW
100 µW
1 mW
10 mW
100 mW
1 W
10 W
100 W
1 kW
10 kW
100 kW
Power consumption
Sustainable energy generation
32 kHz quartzoscillator
standby
Electronic watch,calculator
RFID tag
Hearing aid, CO2 detector
Miniature FM radio
Bluetooth transceiver
MP3 player
Domestic wifi router, mobile phone, LED bulb
Smartphone, Tablet, Laptop MPU, low-energy bulb
Desktop computer
High performance workstation, small server
HPC supercomputer, large HPC cloud storage
Vibrational energy harvester (50–100 Hz)10 cm2 indoor PV
1 m2 outdoor Si PV
7 mm2 thermoelectric ΔT = 50 ˚C
10 mm2 thermoelectric – body heat
1 m2 outdoor 50% efficient CPV
1 km2 SiPV solarfarmDomestic heating system
Small hydro scheme
ENIAC valve computer
HSPA mobile wifi substationMobile phone network base station
Laptop computer, thin client computer
Software Energy Performance / Opportunities
EC 2020 Society Agenda
Digital Agenda for Europe:
By 2013, access to all for internet with ≥ 30 Mbps
By 2020, 50% of people with access to internet with ≥ 100 Mbps
Smart Growth:
Increase internet of things (ICT devices) to improve GDP
Whilst meeting climate change targets:
Reduce CO2 emissions by 20% from 1990 levels !!!!!!
Develop economic growth based on knowledge & innovation
ICT Energy Objectives
High performance and low cost are major ICT drivers
Energy is only optimised when it prevents high performance (scaling, multi-core, photonics at rack level, etc….)
The Vision of the ICT Energy Co-ordinated Action is to direct research towards sustainable energy solutions for ICT devices and systems
Can a heuristic approach produce larger savings than individual energy improvements at each level of the stack?
How to produce energy sustainable ICT systems?
or the cost is excessive (HPC)
System Stacks: Autonomous Sensor
www.sensorsystems.org.uk
Page 2 of 5
www.sensorsystems.org.uk
and maintain interoperability across the range of activities and requirements in a complex system, it is proposed that a reference model be developed. Using the approach taken by the successful OSI model, a comparable model for a hierarchical sensor system is proposed.
A sensor systems reference model As in the case of communication systems, a sensor system is now more complex than can be effectively produced by a single, vertically integrated organisation. Consequently, the component parts and functions of sensor system need to be delineated by function and interface in order that concurrent system development can take place efficiently. The derivation of the model is based upon the OSI model but adapted for the specialist requirements of a complex sensor system. By examining the fundamental requirements of a sensor system, it can be established that all sensor systems carry out the same functions. In all cases, there is a requirement to detect a parametric change in the measureand, and after suitable conversion of the data into information, to pass that information to a ‘decision making authority’. Importantly, although the sensor system may carry out signal pre-processing or filtering, it does not make any decisions; this is beyond the remit of a sensor system. In the case of complex networks, this decision authority is often a human being, however it may be a machine system such as a automotive engine management system.
Pow
er a
nd m
anag
emen
t Sensor Element
Transductance & pre-processing
Communications and networking
Data repository
Analysis & post processing
Visualisation & presentation
Measurand
Superior decision system (human, biological, electronic, mechanical, chemical) acting upon the sensor system output
Dat
aIn
form
atio
n
Pow
er a
nd m
anag
emen
t Sensor Element
Transductance & pre-processing
Communications and networking
Data repository
Analysis & post processing
Visualisation & presentation
Measurand
Superior decision system (human, biological, electronic, mechanical, chemical) acting upon the sensor system output
Dat
aIn
form
atio
n
Figure 2 Sensor Systems Hierarchy
Using this generic approach, the model can be applied to a simple sensor system such as a thermometer or weather vane, or a highly complex network such as a battlefield management system or traffic management system. In order to understand the logic in How comparable is a supercomputer system stack?
Layout
RTL (u architecture)
Synthesis
Functional Blocks
Gate
Compiler
Program
Algorithms
Application
ISA
OS
The ICT System Stack
Questions
What is the biggest challenge to reduce energy consumption in ICT devices?
What is the most promising opportunity to reduce energy consumption in ICT devices?
Which part of the system stack do you belong to?
What are the fundamental limits for reducing energy consumption in this level of the stack? (Are they known?)
What strengths does Europe have to produce sustainable ICT systems?
What weaknesses does Europe have to produce sustainable ICT systems?
Wireless Standards vs Fundamental Limit
10-22
10-20
10-18
10-16
10-14
10-12
10-10
10-8
10-6
10-4
108
109
1010
1011
1012
1013
1014
1015
10-6
10-5
10-4
10-3
10-2
10-1
100
En
erg
y (
J/b
it o
ve
r 1
0 m
)
Frequency (Hz)
Wavelength (m)
single
photon
limit
3G
Bluetooth
Zigbee
802.11n wifi
GSM
Opportunity Challenges
Contact details
Contact: Prof Douglas Paul [email protected]
Tel:- +44 141 330 5219
http://userweb.eng.gla.ac.uk/douglas.paul/index.html
Draft available – e-mail: [email protected]
Input Deadline: Friday 9th October 2015
Questions
What is the biggest challenge to reduce energy consumption in ICT devices?
What is the most promising opportunity to reduce energy consumption in ICT devices?
Which part of the system stack do you belong to?
What are the fundamental limits for reducing energy consumption in this level of the stack? (Are they known?)
What strengths does Europe have to produce sustainable ICT systems?
What weaknesses does Europe have to produce sustainable ICT systems?