apidays LIVE Paris - Responding to the New Normal with APIs for Business, People and Society
December 8, 9 & 10, 2020
Reconcile the European data strategy with Carbon Neutrality Objectives of the Green deal
Stefano Nativi, Big Data Lead Scientist at DG Joint Research Centre of the European Commission
Handwritten Text Recognition for manuscripts and early printed texts
apidays LIVE Paris - Reconcile the European data strategy with Carbon Neutrality Objectives of the Green deal by Stefano Nativi
1. Reconcile the European Data Strategy
with Carbon Neutrality Objectives
of the Green Deal
10th Dec 2020
Stefano Nativi (JRC Big Data lead scientist)
co-authors: Alex Kotsev, Monica Posada, and
Lorenzino Vaccari
2. EU Data Strategy and Green Deal strategy
Carbon neutrality by 2050
3. • Surviving factors for Innovative technologies (AI/ML)
The Computing, Data, and Energy nexus
• In the next future, HPC and AI/ML technologies will be
largely fused and users will not be able to distinguish them
4. • After decades, CPU
power growth is
decelerating
• Limiting factors are
energy consumption
and transistors
miniaturization
AI drivers: the HPC, Data, Energy nexus
• the amount of data
generated yearly is
meaningfully
accelerating
Data growth will largely outpace
foreseeable improvements in
computational power –Energy is a limiting
factor
5. • In 10 years’ time, the development of the
Internet of Things would lead to some
200 billion devices being interconnected,
i.e. 25 times more than the number of
humans
• As a result, the amount of data the world
produces will double every 12 hours
IoT 2.0 and Digital Twins (present revolutions)
• In 2020
• 40 Zettabytes of data (every
person generates 1.7MB per
second)
• 14B Mobile devices
• 31B IoT Devices (127 being added
every second)
• 5,000+ AI companies
• 200M active VR users
6. • generates up to 2% of the global CO2 emissions, a
number on par to the aviation sector contribution
• data centres are estimated to have the fastest growing carbon
footprint across the whole ICT sector.
• consumes approximately 7% of the global electricity,
and it is forecasted that the share will rise up to 13%
by 2030
• data centres are estimated to account for 1.4% of the global
electricity consumption and the compound annual growth rate
(CAGR) between 2007 and 2012 has been estimated as 4.4
• Real-time video streaming, online gaming as well as
mobile devices (5G, IoT, etc.) already account for 60%
of all data traffic, and it is expected that this will rise to
80% by the end of 2020.
ICT sector footprint
7. Energy consumption for AI/BD sector
source: IEEE STC on Sustainable Computing
An AWS zone infrastructure
consumes between 100 and
160 MW power
• 69 Availability Zones (within 22
geographic regions around the
world) + 1 local zone
• 13 more zones are planned
AI Training
8. • Accelerated the Cyber-physical interaction
paradigm, generating more data
• Reaching in few months the Internet traffic
growth expected in one year
• New users and society activities moved on
the Web
COVID effect (1st wave)
9. • Reduce the energy consumption of processing units, servers, and
software (e.g. new chips, data centers cooling systems, SW
optimizations, etc.)
• Power consumption vs usage time
• Reduce data movement energy consumption by moving intelligence to
the edge of the network
• Power consumption vs usage time
• Ensure more sustainable lifecycle process for ICT products and services
Interesting sustainability areas/approaches
10. • Synergetic implementation of European Green Deal and
European Data Strategy
• To unlock the full benefits of the digital transformation and
data-driven innovation to support the ecological transition
• The datafication process can support in many ways the transition
towards a carbon neutral (and smarter) society
• To apply the Green Deal carbon neutral strategies to the ICT
production and usage processes
• ICT enabling technologies (e.g. AI, blockchain, IoT, etc.) must be
green as well
Next Generation EU: a great Opportunity
11. • It seems infeasible to address climate change agenda without the massive adoption of
Digital Transformation in our Society
• Coordination of datafication and digitization processes will facilitate:
1. industrial innovation towards a greener production (e.g. IoT 2.0, Internet of transformation,
Smart industry, Blockchain, Digital Twins, Future Internet, 5G/6G)
2. a greener ICT (Edge computing, green Data Centers, adiabatic microelectronics, ...)
3. support to the objectives of increasing European technological competitiveness and open
autonomy (data and technology sovereignty, Green Deal data space, Digital Twins, ..)
4. the monitoring and reporting of policy effect on environmental issues such as the emissions
(monitoring and reporting digital infrastructure and tools)
Rationale
12. • 2018 Best Practice Guidelines for the EU Code of
Conduct on Data Centre Energy Efficiency: Version
9.1.0 (M. Acton, P. Bertoldi et al., 2018)
• European Commission DG JRC
• A voluntary initiative created in 2008 in response to
the increasing energy consumption in data centres
and the need to reduce the related environmental,
economic and energy supply security impacts
• Anticipating the standardization effort
• https://ec.europa.eu/jrc/en/publication/2018-best-practice-
guidelines-eu-code-conduct-data-centre-energy-
efficiency-version-910
EU Code of Conduct on Data Centers
13. 3
5 6
15
27
68
73
40
27
4
0
10
20
30
40
50
60
70
80
2.8-3.0 2.6-2.8 2.4-2.6 2.2-2.4 2.0-2.2 1.8-2.0 1.6-1.8 1.4-1.6 1.2-1.4 1.0-1.2
NumberofDataCentres
PUE ranges
Data Centres per PUE range
(Source: Paolo Bertoldi, JRC)
PUE = Power usage effectiveness
1 + energy not used for IT
14. • Formalize best practices, provide technology guidelines, and assessment criteria
• ITU-T Focus Group on Environmental Efficiency for Artificial Intelligence and other
Emerging Technologies (FG-AI4EE) –Co-chaired by JRC
• Identify the standardization needs to develop a sustainable approach to AI and other
emerging technologies including (e.g. ML/DL, AR/VR, IoT, Digital Twins, industry 5.0,
cloud/edge computing, nanotechnology, 5G)
• WG1: Requirements of AI and other Emerging Technologies to Ensure Environmental Efficiency
• WG2: Assessment and Measurement of the Environmental Efficiency of AI and Emerging
Technologies
• WG3: Implementation Guidelines of AI and Emerging Technologies for Environmental Efficiency
Standardization role
https://www.itu.int/en/ITU-T/focusgroups/ai4ee/Pages/default.aspx
15. • JRC leads a Technical Report on “Computer Processing, data
management and energy perspective”
• Define a set of energy efficiency criteria for AI and
innovative technology system classes
• IoT 2.0, Digital Twins, Cloud-based smart applications
• Efficiency in Data storage, transfer, and processing
ITU-T TR: Computer Processing, data
management and energy perspective
16. • take the lead in supporting the transition to a less energy intensive AI sector
• Examples: AI and new tools/KPI dealing with ICT lifecycle sustainability
• Work on introducing/facilitating certification scheme for green ICT
components, infrastructures and practices
• Examples: characterization and communication of the environmental impact of digital technologies
–including SW code design, development, and execution.
• Consider in-built (by design) constraints to power data infrastructures within
a carbon-free energy scenario –i.e. Green Deal scenario
• Examples: innovative economic models including environmental and social costs; Green Deal
strategies implementation.
Opportunities 1/2
17. • To reduce the energy consumption of processing units, servers and associated
infrastructures
• Examples: putting energy consumption and data access at the core of AI/Data computing innovation (in memory computing)
–pushing also the European technology open autonomy.
• To minimise the energy required for moving data without compromising the quality of value-
added products
• Examples: moving the intelligence from the centre to the edge of the network; distributed intelligence management.
• To foster cross-sector and multi-stakeholder relationships and leverage on data-driven
innovation in support of the transition to a low-carbon economy
• For example: mobilizing multiple stakeholders and transforming existing practices across all economic sectors; the reuse of
vast amounts of heterogeneous data; innovative approaches on sustainable digitalization along with appropriate
technologies and standards should be intertwined together.
Opportunities 2/2
MOST OF FACILITIES HAVE ACHIEVED A PUE OF 1.6 – 1.8
FOLLOWED BY THE 1.8– 2.0 RANGE.
FOUR SITES THAT ARE IN THE RANGE OF 1.0-1.2 – UK AND NORDIC COUNTRIES
Update/add/delete parts of the copy right notice where appropriate.
More information: https://myintracomm.ec.europa.eu/corp/intellectual-property/Documents/2019_Reuse-guidelines%28CC-BY%29.pdf