Content uploaded by Ralph Hintemann
Author content
All content in this area was uploaded by Ralph Hintemann on Jan 14, 2019
Content may be subject to copyright.
Technology radars for energy-efficient data centers
A transdisciplinary approach to technology identification, analysis and evaluation
Ralph Hintemann
Borderstep Institute for Innovation and Sustainability
Berlin
hintemann@borderstep.de
Simon Hinterholzer
Borderstep Institute for Innovation and Sustainability
Berlin
hinterholzer@borderstep.de
Abstract— Data centers are responsible for a constantly growing
demand for energy and resources. Numerous studies show that
there is still considerable potential for improving the energy and
resource efficiency of data centers. Against this background, it is
of great importance that existing efficiency technologies be applied
to a greater extent and new technologies put into practice. One
instrument to support this development is a technology radar for
energy efficiency technologies in data centers. A technology radar
is an instrument for monitoring technology that supports the
identification, evaluation and observation of relevant technologies
in specific areas. This paper describes how such technology radars
are developed in a transdisciplinary process and presents current
and future efficiency technologies for data centers.
Keywords- data center, energy efficiency, technology radar,
transdisciplinary approach, energy, cooling, power supply, IT
management)
I. INTRODUCTION
The advancing digitalization of economy and society means
that in business, science and public authorities, but also in all
other areas of social life, ever larger amounts of data are being
transported, stored, managed and made available for various
services. Scientific studies on the worldwide energy
requirements of data centers assume a magnitude of between 300
and 400 billion kWh [1]–[4], some studies even expect it to rise
to some 1.000 billion kWh until 2025 [5]. This means that data
centers are responsible for about 1.5% of the world's electric
power consumption. The demand for computing and storage
capacity in data centers is predicted to increase significantly in
the future. Cisco expects global data center IP traffic to triple to
20.6 zettabytes per year between 2016 and 2021, with server
workloads increasing by a factor of 2.3 to 567 million installed
workloads over the same period. In connection with this
development, the global energy requirements of data centers will
continue to rise significantly in the future, despite existing
efficiency advances in IT and data center infrastructure [1], [4],
[6], [7].
To illustrate the increasing energy demands of data centers,
a current development is briefly presented. According to current
calculations, energy requirements due to bitcoin mining alone
are increasing very strongly [8], [9]. Forecasts assume that the
energy requirements of bitcoin mining could reach more than
eight gigawatts in the course of 2018 [10]. This is equivalent to
the average annual power demand for all of Austria or about one-
fifth of annual data center power consumption worldwide [11].
Against the background of the increasing energy demand of
data centers, it is of great importance that energy efficiency
technologies are used in data centers. This applies both to
existing technologies and to new approaches that are still in the
research stage. As numerous studies show (e.g., [4], [12]–[16],
up to 50% of the current energy requirements of data centers
could be saved if available efficiency technologies were used
across the board. However, new energy-efficient and
environmentally-friendly technologies often do not become
established as quickly as is desirable from an environmental
policy perspective [17], [18]. Data center operators, in particular,
often fear that a new technology may have a negative effect on
the availability of the data center.
This is where the technology radars for energy-efficient data
centers presented in this paper come in. In a transdisciplinary
process, current and future efficiency technologies for data
centers are identified, evaluated and communicated within the
data center expert community. The transdisciplinary approach is
particularly well suited to the sustainability-oriented problem
with its high degree of uncertainty and different values [19],
[20].
II. METHODOLOGY
Section 2 explains the methods by which the technology
radars were created. In 2.1 the background of the investigation
is described, in 2.2 the transdisciplinary approach is considered
and in 2.3 the graphical structure of the technology radars is
explained.
A. Project background
The technology radars presented in this article were
developed within the research project "Total Energy
Management for Professional Data Centers" (TEMPRO), which
is funded by the German Federal Ministry of Economics and
Energy. The overriding goal of TEMPRO is holistic increase in
the energy and raw material efficiency of data centers in
Germany, taking into account the upstream and downstream
stages of the value chain. TEMPRO has two complementary
project priorities for achieving this goal:
On the one hand, a basis for evaluating the holistic energy
and raw material efficiency of data centers should be created,
and the necessary information about the energy and raw material
requirements of data centers should be compiled. TEMPRO
analyses and documents the entire life cycle of the infrastructure
elements (information technology, power supply, cooling and air
conditioning) using selected data centers as examples. On this
basis, suitable key figures and processes can be jointly
developed that enable the energy and raw material efficiency of
data centers to be holistically evaluated.
Secondly, new efficiency technologies in data centers, which
lead to considerable energy savings, should be examined, and
confidence in their reliability strengthened. To this end,
information on new technological approaches to improving the
holistic efficiency of energy and raw materials should be
gathered and communicated to the professional public. Selected,
promising energy efficiency technologies are prototypically
developed by the participating companies in cooperation with
research institutes.
B. The Bordersstep transdisciplinary approach to the
development of technology radars
As an independent and non-profit research institution,
Borderstep is active in the field of application-oriented
innovation and entrepreneurship research and is committed to
the guiding principle of sustainable development. The research
projects are characterized by their close relationship to real-
world practice. The majority of the projects are cooperative
projects with partners in the business world, as is the case with
TEMPRO. One focus of the work of the Borderstep Institute is
the identification, analysis and evaluation of new sustainable
technologies and the support of their implementation. For this
purpose, a transdisciplinary approach developed and applied at
the Institute is followed.
Transdisciplinarity is a methodological approach that
combines scientific and practical knowledge. The use of such
transdisciplinary approaches is driven by the need to solve
complex problems in the real world, taking into account the
diversity of scientific and social perspectives on these problems.
Since the end of the 20th century, more and more
transdisciplinary orientations have prevailed in science and the
knowledge society [19].
The transdisciplinary approach developed at the Borderstep
Institute involves experts from science and society in the
identification, analysis and evaluation of new sustainable
technologies in a stepwise process. Transdisciplinary research is
particularly suitable for problems where there is a high degree of
uncertainty and where different values come into play [21], [22].
Transdisciplinary research approaches are especially valuable in
sustainability issues [20], [23]–[25].
Transdisciplinary research primarily addresses three
different kinds of knowledge: systems knowledge, target
knowledge and transformation knowledge. System knowledge
refers to questions about the origin and possible development of
a problem field and about interpretations of problems in the real
world. Target knowledge includes knowledge about the
determination and explanation of practice-oriented targets, and
transformation knowledge refers to the development of
pragmatic means and possibilities for transforming existing
systems [26]. These three types of knowledge are also addressed
in the technology analysis of the Borderstep Institute. On the one
hand, new technologies are analyzed and described in a
combination of knowledge from research and practical
application (system knowledge). Secondly, the technologies are
evaluated with regard to sustainability goals (target knowledge).
Thirdly, the knowledge gained is transferred and the
technologies identified as promising (transformation
knowledge) are promoted in cooperation with business partners.
In this specific case, the transdisciplinary approach was
applied to support the development of technology radars for
energy-efficient data center technologies. A technology radar is
an instrument for technology monitoring, and supports the
identification, evaluation and observation of relevant
technologies [27]. The approach is based on the typical three
phases of a transdisciplinary project: the problem identification
and structuring phase, the problem research phase and the
implementation phase [26], [28].
A total of about 50 data center experts from science and
industry were involved in the process. These included
representatives from industry such as data center operators, data
center service providers (planners and consultants) and data
center suppliers (manufacturers of IT hardware and software,
air-conditioning technology, energy technology, measurement
and control technology). In addition, scientists from six different
disciplines (computer science, air conditioning, economics,
materials science, environmental technology and energy
management) were involved in the process.
Within the scope of problem identification, the topic area of
holistic energy and resource efficiency was first processed and
the fields of technology to be investigated were structured.
Based on analyses of the German data center structure, the
following technology areas were identified as essential for
energy requirements:
Cooling/air conditioning/ventilation
Electric power supply/generation
IT and IT management
In 2015, 59% of the energy requirements of data centers in
Germany were accounted for by IT, 25% by cooling/air
conditioning/ventilation and 14% by electric power supply [29],
[30]. This covers the essential areas in which energy is needed
in the data center. This distinction also takes into account that
competencies in data center operations and in the planning and
implementation of data centers are almost always differentiated
in terms of facility management and IT management. In the field
of facility management, it makes sense to distinguish between
cooling/air conditioning/ventilation and power
supply/generation, because the technologies and systems used
here are largely independent of one another. In the area of IT
hardware and IT management, a distinction could be made
between server, storage and network. However, because there
are more and more solutions, especially for future technologies,
in which the boundaries between these areas become blurred, a
subdivision was dispensed with.
Based on extensive technology research, an interdisciplinary
team of Borderstep scientists developed the first drafts for three
technology radars in the above technology fields within the
problem research phase. The sources used were trade journals
for data center technologies such as dcd, IT Business,
Technology Review and the following online portals: Data
Identify applicable sponsor/s here. (sponsors)
Center Knowledge, Network World from IDG, The Datacenter
Journal, Data Centre Dynamics, Data Center Week, DataCenter-
Insider by Vogel IT-Medien and LANline. Interviews were also
conducted with experts on individual technologies. In addition,
the winners of recent data center awards such as the DCD
Award, the Datacloud Award, the Datacenter Week Award and
the German Data Center Award were analyzed. The analyses
and technology forecasts of market analysts such as IDC and
Gartner, e.g. Gartner Hype Cycles, were also included in the
research.
The technology radars developed in this way were presented
to experts from business and science at an internal TEMPRO
workshop in Berlin on October 11, 2017. A six-week
commentary phase followed, during which the technology
radars were further developed and supplemented. The interim
results were then presented to external experts at an innovation
workshop in Berlin on December 5, 2017. A total of 15 experts
took part in the workshop. These included representatives of
major international IT manufacturers, international
manufacturers of air-conditioning and power supply
infrastructure, data center consultants and representatives of
various scientific disciplines at universities. In the course of the
workshop, the innovation radars were slightly supplemented and
adapted once again. In addition, a joint prioritization of the
particularly important energy efficiency technologies was
carried out.
The implementation phase of the transdisciplinary project
has been under way since March 2018. First results and excerpts
from the technology radars were presented at a non-public
technical workshop at the Future Thinking conference (April
2018). In addition, the technology radars were presented and
discussed at a meeting of the working group "Computer Center"
of the German digital association Bitkom (June 2018). The
members of the working group confirmed the basic contents and
provided some additional information. The international
publication of the results in this paper is another important step.
The technology radars will be made available to large sections
of the expert public later, in a coordinated process with industry
partners through lectures, publications in journals, blog articles
and the like.
C. Basic structure of the Borderstep technology radars
Figure 1 shows the basic structure of a technology radar used
in TEMPRO. With the help of this instrument, energy efficiency
technologies can be presented clearly and communicated
effectively. The technology radar currently distinguishes
between solutions already in use in data centers (Technology A),
solutions for which there is already a niche market (Technology
B), solutions that are in the pilot stage of application
(Technology C) and solutions that are still in the research stage
(Technology D). The technologies are further differentiated
according to whether they are currently already intended for use
in data centers or whether they come from other areas of
application (Technology E or Technology F).
Figure 1. Basic structure of the Borderstep technology radars
III. RESULTS
The technology radars are presented in a very short form in
the following chapter. The chapter is divided into the three
technology fields ‘cooling/air conditioning/ventilation’, ‘energy
supply and generation’ and ‘IT and IT management’. For each
technology radar, three technologies identified by experts as
relevant are briefly presented. Technology radar for cooling/air
conditioning/ventilation in data centers plus a short description
of three relevant technologies
A. Technology radar for cooling, air conditioning and
ventilation in data centers
Figure 2 shows the technology radar for energy efficient
cooling, air conditioning and ventilation technologies in data
centers.
Figure 2. Technology radar for cooling/air conditioning/ventilation
In the area of cooling, air conditioning and ventilation, 21
different energy efficiency technologies were identified and
analysed in detail. For each of the technologies mentioned in the
technology radars, a one-page fact sheet was prepared together
with the experts. The solutions range from established
technologies such as Direct Free Cooling to the research projects
like use of Phase Change Materials (PCM) for refrigeration in
data centers. Cooling solutions from other industries, which have
not yet been used in data centers, were also considered. These
are, for example, cold storage solutions with ice.
In the innovation workshop (December 2017), described in
section II.B, experts voted on the significance of the various
technologies. The top three technologies are highlighted in the
radar above and a short description of each can be found below.
Water cooling of entire servers: Water cooling of entire
servers is a server cooling method where all relevant
components of a server are thermally conduced to a water
drained system. To do this, an individually shaped metal body
connected to water pipes collects all the heat from the
components. As this concept uses only one inlet and one outlet,
the risk of leakage is minimized.
This system can improve energy efficiency by reducing the
mass flow of the coolant as a result of the higher heat capacity
of water (compared with air). In addition, water cooling
improves capacity for waste-heat-recovery enormously as it
binds the heat in a more compact way and allows higher
operational temperatures.
Waste heat recovery: The entire electric energy that supplies
all the data center equipment (e.g., IT, UPS, lighting) is
ultimately transformed into thermal energy. This low-
temperature heat can be recovered for secondary appliances
through the use of various technologies.
A direct method of heat recovery is the use of the heat to
supply the rooms of the data center with heat, for example to
regulate the temperature in offices during the winter. In addition,
some secondary appliances can be supplied with low
temperature heat; for example, greenhouses or swimming pools.
Indirect heat recovery can be carried out with a heat pump
system that supplies a thermal load at a higher temperature. If
the heat needs to be transported over some distance, it is
mandatory to transfer the heat to fluids with much higher heat
capacity than air -- for example water.
To improve effectiveness and temperature level, heat
recovery can be combined with direct fluid cooling of
components for example through immersion cooling or water
cooling of entire systems.
Data center without refrigeration: The operation of
refrigeration systems in data centers depends mainly on two
temperatures, the maximum permissible temperature of the heat
source (IT hardware, UPS) and the temperature of the heat sink.
As long the temperature of the heat sink is lower than the
maximum permissible inlet temperature for the cooling cycle
(plus a ΔT for a heat exchanger) the cooling can be provided
without refrigeration.
B. Technology radar for electric power supply/generation in
data centers
The second field targets new technologies and innovations
that improve the efficiency of the electric power supply in data
centers. Figure 3 shows the technology radar for energy supply
and generation.
A total of 14 different energy efficiency technologies were
identified in the area of power supply/generation. Solutions were
discussed and analyzed in detail that are already being used in
data centers today, but whose use should also be further
promoted, such as modular UPSs. But also solutions that are still
in the state of research, such as the e48V POL technology, in
which the power is supplied with 48V without its own server
power supply to the point of load (POL).
Figure 3. Technology radar for electric power supply/generation
The following technologies (marked yellow) were selected
by the experts in the innovation workshop (December 2017) as
the most significant innovations:
Flexibility for the smart grid: The components of a data
center, especially uninterruptible power supply and a diesel
generator can supply the electricity system with flexible power.
Depending on local technical and market regulations (e.g.,
government, grid operator, energy market) it can be possible and
profitable to use those facilities to provide flexibility in the
electric grid.
Direct current (DC) power supply: Almost every data center
is supplied with AC power from an electric grid. Depending on
the power architecture, this AC power from the grid is
transformed a small number of times through the UPS (AC to
DC to AC) as well as inside the servers to low voltage DC power
(e.g. 12 V, 5 V) to provide energy to the electronic components.
A central DC supply at the input and DC power distribution
reduces the electric conversion losses and makes the overall
electricity supply more efficient.
Software-defined power: New server technologies integrate
all components of local power conversion from servers and racks
into the data center management. This can help to optimize the
whole power supply and adapt it to the current workload. The
adaptation of voltage can also ensure more stable operation in
connection with aging IT equipment [31], [32].
C. Technology radar for IT and IT management in data
centers
The third field of efficiency technologies in data centers
focuses on technologies within IT and IT-management itself. It
includes energy-efficient computing technologies as well as
management tools that improve the overall efficiency of IT
hardware in data centers. Figure 4 shows the technology radar
for this field.
Figure 4. Technology radar for IT and IT management
In the area of IT and IT management, a total of 21 different
technologies were identified and analyzed that are suitable for
further increasing the energy efficiency of data centers in the
future. The range of technologies extends from existing
solutions such as All Flash solutions to quantum computers.
The following technologies were identified as highly
significant for the energy efficiency of data centers in the
innovation workshop in December 2017:
Enterprise resource planning (ERP) for data centers: ERP
optimizes overall operations in a business such as a data center.
Modern implementations use software tool to monitor and adjust
operating processes. Through this constant optimization, a data
center can operate at maximum efficiency in both economic and
also ecological terms.
Analytics for energy efficiency: Modern data centers are
equipped with high numbers of sensors as well es meters that
constantly monitor processes relevant to energy consumption.
Raw data having varying structure is complex and hard to
interpret. Data analytics can improve the value of such sensor
data for making decisions leading to greater energy efficiency.
Neuromorphic computing: Neuromorphic computing is an
approach which involves imitating the function of the central
nervous system within complex circuits through a combination
of digital and analogue components. Most artificial intelligence
(AI) algorithms still only simulate neuronal networks, but these
functions can be directly integrated into silicon. This is called a
“neuromorphic chip.” These chips carry out specific tasks much
faster and with greater energy efficiency compared with
traditional chips, and are able to improve the overall efficiency
of ICT [33].
IV. DISCUSSION AND OUTLOOK
The transdisciplinary method has the potential to accelerate
the diffusion of technologies that improve the overall efficiency
of data centers. The graphical elements of technology radar help
to illustrate the stage of development for each individual
technology. In addition to the technology radars, a standardized
form for analyzing each technology is currently in development.
The form includes a short description of each technology, the
main use of the technology, facilitators and restraints, the
targeted type of data center, a forecast of market availability and,
of course, potential to improve efficiency.
To ensure an up-to-date overview of current trends in
technologies that enhance energy efficiency in data centers it is
necessary to have a comprehensive view which also involves
adjacent technological fields, and a constant feed of information
on new innovations. The technology radars as well as the
standardized forms for analysis will be updated on the basis of
another expert workshop in the summer of 2019.
ACKNOWLEDGMENT
This contribution is a result of the project TEMPRO - Total
Energy Management for professional data centers, which is
funded by the German Federal Ministry of Economics and
Energy. We would like to thank all project partners and
especially the project manager Alexandra Pehlken for the
constructive and excellent cooperation. We would also like to
thank the working group “data center” of the German digital
association Bitkom for their constructive criticism and further
input on the individual technology radars.
REFERENCES
[1] T. Bawden, „Global warming: Data centres to consume three times as
much energy in next decade, experts warn“, The Independent, Jan. 2016.
[2] G. A. Brady, „Energy Efficiency in Data Centres and the Barriers to
Further Improvements: An Interdisciplinary Investigation“, The
University of Leeds, Leeds, 2016.
[3] G. Cook, T. Dowdall, D. Pomerantz, und Y. Wang, „Clicking clean: how
companies are creating the green internet“, Greenpeace Inc., Washington,
DC, 2014.
[4] R. Hintemann und J. Clausen, „Green Cloud? The current and future
development of energy consumption by data centers, networks and end-
user devices“, in Proceedings of ICT for Sustainability 2016, Amsterdam,
The Netherlands, 2016.
[5] A. S. Andrae, „Total Consumer Power Consumption Forecast“, gehalten
auf der Nordic Digital Business Summit, Helsinki, Finland, 2017.
[6] G. Cook u. a., „Clicking Clean: Who is winning the race to build a Green
Internet?“, Greenpeace Int. Amst. Neth., 2017.
[7] R. Hintemann und J. Clausen, „Rechenzentren in Deutschland: Eine
Studie zur Darstellung der wirtschaftlichen Bedeutung und
Wettbewerbssituation. Studie im Auftrag des Bundesverbandes
Informationswirtschaft, Telekommunikation und neue Medien e.V.
(BITKOM)“, Berlin, 2014.
[8] K. J. O’Dwyer und D. Malone, „Bitcoin mining and its energy footprint“,
2014.
[9] A. de Vries, „Bitcoin’s Growing Energy Problem“, Joule, Bd. 2, Nr. 5, S.
801–805, 2018.
[10] Digiconomist, „Bitcoin Energy Consumption Index“, Digiconomist,
2018. [Online]. Verfügbar unter: https://digiconomist.net/bitcoin-energy-
consumption. [Zugegriffen: 26-Sep-2018].
[11] International Energy Agency (IEA), „World Energy Outlook 2009“,
2009.
[12] K. Fichter und R. Hintemann, „Beyond Energy. The Quantities of
Materials Present in the Equipment of Data Centers“, J. Ind. Ecol., Bd. 18,
Nr. 6, S. 846–858, 2014.
[13] J. Gröger und M. Köhn, Leitfaden zur umweltfreundlichen öffentlichen
Beschaffung: Produkte und Dienstleistungen für Rechenzentren und
Serverräume. Umweltbundesamt, 2016.
[14] J. Koomey, „Growth in data center electricity use 2005 to 2010“, Rep.
Anal. Press Complet. Req. N. Y. Times, 2011.
[15] J. G. Koomey, „Worldwide electricity used in data centers“, Environ. Res.
Lett., Bd. 3, Nr. 3, S. 034008, 2008.
[16] A. Shehabi u. a., „United States Data Center Energy Usage Report“,
Ernest Orlando Lawrence Berkeley National Laboratory, Berkeley, CA,
LBNL-1005775, 2016.
[17] J. Clausen, E. Göll, und V. Tappeser, „Sticky Transformation – How path
dependencies in socio-technical regimes are impeding the transformation
to a Green Economy“, gehalten auf der IST-Conference 2016, Wuppertal,
07-Sep-2016.
[18] K. Fichter und J. Clausen, „Diffusion Dynamics of Sustainable Innovation
- Insights on Diffusion Patterns Based on the Analysis of 100 Sustainable
Product and Service Innovations“, J. Innov. Manag., Bd. 4, Nr. 2, S. 30–
67, 2016.
[19] G. H. Hadorn u. a., „The emergence of transdisciplinarity as a form of
research“, in Handbook of transdisciplinary research, Springer, 2008, S.
19–39.
[20] D. J. Lang u. a., „Transdisciplinary research in sustainability science:
practice, principles, and challenges“, Sustain. Sci., Bd. 7, Nr. 1, S. 25–43,
2012.
[21] S. Funtowicz und J. Ravetz, „Values and uncertainties“, in Handbook of
transdisciplinary research, Springer, 2008, S. 361–368.
[22] S. Funtowicz und J. Ravetz, „Post-normal science“, Int. Soc. Ecol. Econ.
Ed Online Encycl. Ecol. Econ. Httpwww Ecoeco Orgpublicaencyc Htm,
2003.
[23] S. Baumgärtner, C. Becker, K. Frank, B. Müller, und M. Quaas, „Relating
the philosophy and practice of ecological economics: The role of
concepts, models, and case studies in inter-and transdisciplinary
sustainability research“, Ecol. Econ., Bd. 67, Nr. 3, S. 384–393, 2008.
[24] P. Brandt u. a., „A review of transdisciplinary research in sustainability
science“, Ecol. Econ., Bd. 92, S. 1–15, 2013.
[25] P. Nanz, O. Renn, und M. Lawrence, „Der transdisziplinäre Ansatz des
Institute for Advanced Sustainability Studies (IASS): Konzept und
Umsetzung“, GAIA-Ecol. Perspect. Sci. Soc., Bd. 26, Nr. 3, S. 293–296,
2017.
[26] C. Pohl und G. H. Hadorn, Principles for designing transdisciplinary
research. oekom Munich, 2007.
[27] R. Rohrbeck, J. Heuer, und H. Arnold, „The technology radar-an
instrument of technology intelligence and innovation strategy“, in
Management of Innovation and Technology, 2006 IEEE International
Conference on, 2006, Bd. 2, S. 978–983.
[28] A. Elzinga, „Participation“, in Handbook of transdisciplinary research,
Springer, 2008, S. 345–359.
[29] R. Hintemann, S. Beucker, J. Clausen, L. Stobbe, M. Proske, und N. F.
Nissen, „Energy efficiency of data centers - A system-oriented analysis of
current development trends“, in EGG 2016+ Proceedings, Berlin, 2016.
[30] L. Stobbe, R. Hintemann, M. Proske, J. Clausen, H. Zedel, und S.
Beucker, „Entwicklung des IKT-bedingten Strombedarfs in Deutschland
- Studie im Auftrag des Bundesministeriums für Wirtschaft und Energie“,
Fraunhofer IZM und Borderstep Institut, Berlin, 2015.
[31] DataCenter Insider, „Neue Stromversorgungsarchitekturen drücken die
Datacenter-Kosten“, 2017. [Online]. Verfügbar unter:
https://www.datacenter-insider.de/neue-stromversorgungsarchitekturen-
druecken-die-datacenter-kosten-a-590569/. [Zugegriffen: 25-Okt-2017].
[32] electronicspecifier.com, „On the road to software defined power
architecture“, 2017. [Online]. Verfügbar unter:
http://www.electronicspecifier.com/power/on-the-road-to-software-
defined-power-architecture. [Zugegriffen: 25-Okt-2017].
[33] W. Knight, „Intel wants to move beyond today’s architecture, with brain-
inspired and quantum chips.“, MIT Technology Review, 2018. [Online].
Verfügbar unter: https://www.technologyreview.com/s/609909/intels-
new-chips-are-more-brain-like-than-ever/. [Zugegriffen: 28-Mai-2018].
