Discussions

Energy sustainable ICT

  • Luca Gammaitoni profile
    Luca Gammaitoni
    29 April 2016 - updated 4 years ago
  • 6 comments
    Total votes: 15

The big picture

The reliance of society on the use of Information Communications Technology (ICT) devices and systems is increasing with over 4% of all electricity consumption and 2% of all CO2 emissions now the result of ICT use. If entertainment, telephones, TV and media which are now being translated onto ICT devices and systems are added then these consumption numbers approximately double. FP7-ICT and Horizon2020 have highlighted ICT as a key engine of growth, with the use of ICT to improve energy efficiency by managing energy demand and use. The energy consumption and carbon dioxide emission from the expanding ICT use with present drivers, however, is unsustainable and will impact heavily on future climate change.

Transistors are approaching the minimum amount of energy per switch and the access resistance of electrical interconnects is a key issue for future scaling of energy in all ICT hardware. The failure of the International Technology Roadmap for Semiconductors after 2013 provides evidence of the end of scaling and also of Moore’s law, the economic law driving future ICT hardware systems. Future scaling of transistors is therefore unlikely to result in any significant reduction in system energy consumption and power. This is further exasperated by the product volume requirements for return on investment at the leading edge CMOS technology nodes. Radical new devices, interconnect solutions and system architectures are required if reductions in power per device are to be achieved in the future. Technologies that can be delivered on legacy foundries would also significantly reduce the financial costs of adoption. As Europe is the leader in low power design and consumer embedded systems, this is an enormous opportunity for the EC to lead as applications with old wired devices move towards portable solutions.

 

The work needed

What it is highly needed is a focused activity aimed at identifying methods, techniques and directions for making ICT devices and systems sustainable with regard to energy.

Two different strategies are required: short term and long term.

Short term strategy aimed at gaining fast, incremental progresses in energy efficiency should be focused on improving the role of software in ICT systems. A key issue is the identification of tools or models to allow energy transparency throughout the system stack and optimisation to reduce energy consumption.

Long term strategy aimed at gaining orders of magnitude in energy efficiency should be focused on finding new information processing devices and new information carriers. At the same time, with regards to mobile devices, it is necessary to bridge the gap between how much energy is required and how much is provided with energy harvesting techniques for powering small autonomous devices.

 

The opportunity

Whilst the direct contribution to CO2 emission from ICT devices might be small, ICT devices have the potential to contribute significantly to the reduction of CO2 in transportation, heating (through smart building control) and manufacturing (autonomous sensors). ICT devices in the form of autonomous sensors have the potential to significantly improve the quality of life of society both in Europe and across the whole world.  For autonomous systems, significant improvements in energy harvesting and energy storage at the small scale would also provide disruptive solutions to the use of smart sensors for a host of applications in personalised healthcare, environmental monitoring, industrial monitoring, security and transportation. Such applications have the potential for significant improvements in the quality of life and reductions in energy consumption and CO2 emissions.

In Europe, there is a wide community (identified as ICT Energy community) interested in these topics, covering ULSI, microarchitectures, high performance computing (HPC), energy harvesting, thermoelectrics, energy storage, ICT system design, embedded systems, efficient electronics, static analysis and the physics of computation.

While in US and Japan exist initiatives in the form of research centres and funding programs (see e.g. the E3S, Center for Energy Efficient Electronics Science and the “Next Generation Nano-Micro Devices and Systems for Energy-Efficient Information Processing Division” of IIIEE as a part of social cooperation program with IBM Japan), Europe is still missing a focused effort in this field.

On May 6th  the ICT-Energy consortium is organizing a consultation workshop on energy consumption in ICT with the aim of addressing the needs, rationales and means to develop innovative approaches in this field, and verifying the coverage of the next European research programme compared with industrial needs, emerging technological trends and new scientific insights. We expect further indication to emerge from this workshop in order to inform and advice the European Commission on the strategic directions needed for efficient ICT research, development and innovation.

The ICT-Energy consortium includes the following laboratories and associated FET projects:

UNIVERSITY OF PERUGIA (Lnadauer, Nanopower, Zeropower), ROSKILDE UNIVERSITET (Entra), RUPRECHT-KARLS-UNIVERSITAET HEIDELBERG (Exa2Green),  BARCELONA SUPERCOMPUTING CENTER - CENTRO NACIONAL DE SUPERCOMPUTACION (ParaDime), ECOLE POLYTECHNIQUE FEDERALE DE LAUSANNE (Phidias), AALBORG UNIVERSITET (Sensation), HITACHI EUROPE LIMITED (Tolop), UNIVERSITY OF BRISTOL (EACO Workshop), UNIVERSITY OF GLASGOW (GreeSilicon), UNIVERSITY COLLEGE CORK (SiNAPS).

This contribution is part of the ICT-Energy Strategic Research Agenda (http://www.ict-energy.eu/sites/ict-energy.eu/files/ICTEnergySRA5.3.pdf).

Tags: 
FET Proactive consultationICTICT-Energysustainable energyenergytransistorenergy harvestingautonomous deviceenergy efficiency

Comments

Log in to post comments
Luca Gammaitoni's picture
Luca Gammaitoni,
19/05/2016 13:21

Summary Report

The workshop was organized by the ICT-Energy CSA (www.ict-energy.eu) with the support of the European Commission.

Organising Committee: Luca Gammaitoni, Douglas Paul and Andrea Feltrin (EC)

Participants: Heike Riel (IBM Zurich), Eric Yeatman (Imperial College), Valentina Bacchettini (Wisepower srl), Chiara Benedetta Mezzetti (NiPS Lab), Adrian Cristal Kestelman (BSC), Fernando Gonzalez Zalba (Hitachi Cambridge Lab), Kerstin Eder (Bristol Univ.), Axel Legay (INRIA), John Patrick Gallagher (Roskilde Univ.), Martin Wlotzka (Heidelberg Univ.), Cristina Rusu (ACREU Swedish ACT), Hervè Fanet (Grenoble SEEA), Francesc Moll Echeto (UPC), Luca Benini (ETH and Bologna Univ.), Emre Ozer (ARM) Francesco Orfei (NiPS Lab), Gabriel Abadal (UAB).

Rapporteur: Giorgos Fagas

Area: Future ICT-Energy Concepts for Energy Efficiency & Sustainability

Challenge: The reliance of our society on ICT is intensifying with the translation of entertainment, telephones, TV and media onto mobile and distributed ICT devices and systems and with the increasing number of interconnected smart objects. There is also a growing demand to convert everyday objects into sources of information to achieve energy and resource savings for the new Internet of Things era. These advances increase the energy footprint of ICT (the electricity consumption and CO2 emissions directly attributed to ICT devices and systems are soon expected to surpass the 10% and 5% marks respectively) and put severe constraints on the energy and power requirements for devices and systems. To make this progress sustainable, it is both necessary to reduce the energy consumed in the total system via energy-efficient software and hardware integrated solutions and to achieve energy autonomy through harvesting and the improved management of energy sources available in the environment.

Scope: Research and innovation activities are required aimed at jointly exploring scientific directions and technological options to establish a solid baseline of knowledge and skills for transformational innovation, and to foster the emergence of a broader innovation ecosystem as well as a fertile ground for its future take-up (e.g., through public engagement processes when relevant, or through formal and informal training and education). More specifically research should focus on developing breakthrough technologies based on interdisciplinary and holistic approach for ultra-low power consumption computing paradigms. Ideally, the proposals should benchmark their power and energy savings at the hardware and software level and aim to produce an energy sustainable system for specific tasks.

We identify the following topics where research activity can be focussed with the aim of producing at least two orders of magnitude improvement in energy efficiency for the target application:

a) Energy efficient computing paradigms and systems – exploration and demonstration of new computing paradigms ans systems (like non-Boolean, analogue, hybrid electrical-mechanical, to mention a few) pushing energy consumption below conventional transistor based circuits. It is essential to apply a holistic approach from devices through circuit architectures as necessary for specific application implementations, aiming at the limits of energy consumption for each layer and as a whole. The possibility to integrate new logic and device architectures and apply novel system integration schemes should be considered. Open source aspects of design tools and computing platforms may be addressed to achieve an innovation ecosystem. Attention should be given also to novel paradigms that exploit the unreliability of ultra-low energy devices.

b) Energy autonomous miniaturised smart objects – Demonstrate novel approaches and applications of ‘sensing, analysing and communicating’ using energy autonomous systems with small form factor (approximately 1 mm3 and below) that go beyond what is presently available. Explore the fundamental limits of energy harvesters, managing hybrid energy harvesting and integrating novel energy storage with high-energy density and high-power capability. The use of novel computation paradigms or transmission protocols with improved signal-to-noise ratio may be considered for “intelligence at the edge” applications.

c) Energy transparent and energy adaptive systems – Explore energy transparency mapping energy-related data and requirements between hardware and software and demonstrate systems (hardware-software) that adapt their energy demands in real time depending on the resources available and performance requirements. Using energy transparency tools, develop new energy-efficient algorithms and systems optimising the tradeoffs of computation and communication and that go beyond the tradition dichotomy between application software and universal computing hardware.

As potential impact we identify the following:
• Reduction of the environmental impact of ICT technologies.
• Establishment of a solid baseline of knowledge and skills for a future ICT technology with a significant reduction in energy consumption.
• Development of energy-efficient autonomous sensor for the Internet of Things era.
• Characterisation of novel device technologies as for beyond the CMOS era.
• Development of new energy-efficient computing paradigms that operate close to the fundamental energy limits.
• Design of problem-specific accelerator modules for orders of magnitude energy reduction.
• Demonstration of the emergence of an ecosystem around future low energy technologies with a clear innovation pathways through outreach to and partnership with high potential actors in research and innovation, and from wider stakeholder/public engagement.
• Structuring of a goal oriented community and true interdisciplinary collaboration.

PDF file available at:http://www.ict-energy.eu/sites/ict-energy.eu/files/ICT-Energy%20Consulta...

Total votes: 1
Gabriel Abadal Berini's picture
Gabriel Abadal ...,
10/05/2016 12:46

Energy harvesting technologies (EHT)- including energy transduction, energy management and energy storage sub-technologies- have achieved a notorious degree of maturity during a period in which the paradigm of autonomous system has been the self-powered wireless sensor networks (WSN). However, the fragile equilibrium between the energy extracted from the ambient by the EH devices and the energy demanded by the systems operation is still critically compromised when dimensions are reduced below the mm-scale. So the “bridge the energy gap” problem is 2-dimensional and it has to be considered in an energy-volume space.
Future extrapolations of WSN such as the Internet of Things (IoT) components will need radical improvements in the EHT performance, that can only come from revolutionary new ideas. Nanotechnology, in general, and Nano-OptoElectroMechanical Systems (NOEMS), in particular, can play an important role in proposing new energy harvesting alternatives still unexplored at the nanoscale. Paraphrasing Prof. Luca Gammaitoni, "There’s plenty of energy at the bottom".

Total votes: 1
John Gallagher's picture
John Gallagher,
04/05/2016 13:45

Energy sustainable ICT is indeed a critical topics and has an important dimension of "energy-aware" software development. Energy is not directly consumed by software, but software has a major impact on the energy efficiency of ICT systems, and it is possible for software to waste a great deal of energy. The developer needs an understanding of the energy implications of software, at the developement stage. Waiting until deployment to measure or estimate energy is probably too late to do much about it. Therefore energy modelling at the software level, and static analysis of software with respect to an energy model are central research topics for ICT Energy.

Total votes: 3
Giorgos Fagas's picture
Giorgos Fagas,
03/05/2016 21:36

The Internet of Things (IoT) is considered the enabler of the fourth industrial revolution as miniaturised self-powered wireless sensors underpin: e-health, wearables, smart agriculture, food safety/security, advanced manufacturing, resource-efficient and safe transport, and in general the well-being of people in urban and rural areas. If energy management in these 'miraculous' ICT devices can be resolved, large energy savings (and reduced CO2 emissions) can be delivered by using ICT for resource efficiency and optimisation. To achieve long lifetime and small form factors as required in applications, emerging autonomous sensors need to maintain ultra-low power duty cycles and incorporate an energy harvesting source, an energy storage device and electronic circuits for power management, sensing and communication into sub-cm scale systems. A continuous challenge for every implementation and deployment is the gap between energy requirements for operation and energy supply from harvesting sources and storage to enable truly autonomous wireless devices. To bridge this gap, apart from lower energy switching technologies step changes are also needed to advance energy harvesting, storage and micropower management solutions for the integrated design and deployment of miniaturised autonomous sensors.

Total votes: 4
M. Fernando Gonzalez Zalba's picture
M. Fernando Gon...,
03/05/2016 12:23

As it is well explained in this summary of ICT-Energy, transistors, the computational unit of all ICT devices are at the verge of its minimum working dimensions. No exponential improvement on power needs is expected if we continue down the miniaturization route.

Research on novel devices with functionalities beyond the simple ON-OFF switch or that work below the Landauer limit is then a necessity for the ICT industry.

I expect this initiative to be a success.

Total votes: 5
Douglas Paul's picture
Douglas Paul,
03/05/2016 18:36

The design industry now sees the end of Moore's law as new technology generations are not providing x2 reductions in power, only 5% or so. Hence there is no significant benefit to move to the next technology generation for many of the portable products that dominate ICT. There is a significant opportunity if new, lower energy switching technologies can be produced. Also autonomous systems that require their own energy harvesting systems are a strong developing market for personalised healthcare, smart building control, environmental and security monitoring. Finding practical energy harvesting technology with a step improvement in power outputs could revolutionised many personal and future sensing markets.

Total votes: 6