ITIES - ERCIM - ERCIM News
←
→
Page content transcription
If your browser does not render page correctly, please read the page content below
ercim-news.ercim.eu Number 127 October 2021
ERCIM NEWS
Special theme:
I R
C CU
D
L AR CIT
A R T A N
IES M
S
Also in this issue
Digital Inclusion:
Using Localisation Technologies and Haptic Feedback
for a More Inclusive Society
JointCONTENTS
Editorial Information SPECIAL THEME
ERCIM News is the magazine of ERCIM. Published quarterly, it reports The special theme “Smart and Circular Cities” has been
on joint actions of the ERCIM partners, and aims to reflect the contribu- coordinated by the guest editors Erwin Schoitsch (AIT)
tion made by ERCIM to the European Community in Information and Georgios Mylonas (ISI, Athena Research and
Technology and Applied Mathematics. Through short articles and news Innovation Center)
items, it provides a forum for the exchange of information between the
institutes and also with the wider scientific community. This issue has a Introduction to the Special Theme
circulation of about 6,000 printed copies and is also available online ,at
4 Smart and Circular Cities
by Erwin Schoitsch (AIT) and Georgios Mylonas (ISI,
https://ercim-news@ercim.eu.
Athena Research and Innovation Center)
ERCIM News is published by ERCIM EEIG
6 Smart Governance for Cybersecurity
BP 93, F-06902 Sophia Antipolis Cedex, France
by Michael Kreutzer and Kirstin Scheel (Fraunhofer
+33 4 9238 5010, contact@ercim.eu
SIT)
Director: Philipp Hoschka, ISSN 0926-4981
7 Policies and Recommendations for IT Security in
Contributions Urban Environments from the Morgenstadt Urban
Contributions should be submitted to the local editor of your country Data Partnership Project
by Philipp Lämmel, Michell Boerger, Nikolay
Copyrightnotice Tcholtchev (Fraunhofer FOKUS) and Eva Ottendörfer
All authors, as identified in each article, retain copyright of their work. (Fraunhofer IAO)
ERCIM News is licensed under a Creative Commons Attribution 4.0
International License (CC-BY). 9 Data Privacy in Smart Cities – Federated Learning
to the Rescue?
Advertising by Anastasia Pustozerova and Rudolf Mayer (SBA
For current advertising rates and conditions, see Research)
https://ercim-news.ercim.eu/ or contact peter.kunz@ercim.eu
10 Considering Cybersecurity with Trustworthy IoT in
ERCIMNewsonlineedition:ercim-news.ercim.eu/
Smart Cities
by Christoph Klikovits (Forschung Burgenland),
Clemens Gnauer (Forschung Burgenland), Patrik
Nextissue:
Abraham (Fachhochschule Burgenland)
January 2022: Quantum Computing
12 Digital Twins for Cyber-Physical Threat Detection
Subscription
and Response
Subscribe to ERCIM News by sending an email to
by Matthias Eckhart, Andreas Ekelhart (SBA Research
en-subscriptions@ercim.eu and University of Vienna), and Roland Eisl (ENRAG)
EditorialBoard: 13 Circularity and Sustainability in Modern Smart
Central editor: Grids Through Innovative Energy Market
Peter Kunz, ERCIM office (peter.kunz@ercim.eu) Architectures
by Nikolaos Efthymiopoulos, Prodromos Makris,
Local Editors: Emmanouel Varvarigos (National Technical University
• Christine Azevedo Coste, Inria, France (christine.azevedo@inria.fr) of Athens)
• Andras Benczur, SZTAKI, Hungary (benczur@info.ilab.sztaki.hu)
• José Borbinha, Univ. of Technology Lisboa, Portugal (jlb@ist.utl.pt) 15 Enabling Smart Control and Fair Sharing of
• Are Magnus Bruaset, SIMULA, Norway (arem@simula.no) Renewable Resources in Energy Communities
• Monica Divitini, NTNU, Norway (divitini@ntnu.no) by Sonam Norbu, Benoit Couraud, Merlinda Andoni,
• Marie-Claire Forgue, ERCIM/W3C (mcf@w3.org) David Flynn (Heriot-Watt University, Edinburgh) and
• Lida Harami, FORTH-ICT , Greece (lida@ics.forth.gr) Valentin Robu (CWI and TU Delft)
• Athanasios Kalogeras, ISI, Greece (kalogeras@isi.gr)
17 Green Energy Planning of IoT Rule Automation
• Georgia Kapitsaki, Univ. of Cyprus, Cyprus (gkapi@cs.ucy.ac.cy)
Workflows in Smart Environments
• Annette Kik, CWI, The Netherlands (Annette.Kik@cwi.nl)
by Soteris Constantinou (University of Cyprus),
• Hung Son Nguyen, Unviv. of Warsaw, Poland (son@mimuw.edu.pl)
Andreas Konstantinidis (Frederick University) and
• Alexander Nouak, Fraunhofer-Gesellschaft, Germany
Demetrios Zeinalipour-Yazti (University of Cyprus)
(alexander.nouak@iuk.fraunhofer.de)
• Maria Rudenschöld, RISE, Sweden (maria.rudenschold@ri.se) 18 Using Telematics to Gather User Behaviour Data
• Harry Rudin, Switzerland (hrudin@smile.ch) from a Fleet of Electric Bicycles
• Erwin Schoitsch, AIT, Austria (erwin.schoitsch@ait.ac.at) by Sam Gunner, Eddie Wilson and Theo Tryfonas
• Thomas Tamisier,LIST, Luxembourg (thomas.tamisier@list.lu) (University of Bristol)
• Maurice ter Beek, ISTI-CNR, Italy (maurice.terbeek@isti.cnr.it)
2 ERCIM NEWS 127 October 2021
20 Tethys: An Edge Computing–Ready Water Metering REEARCH AND INNOVATION
System for Smart Cities
by Dimitrios Amaxilatis (SparkWorks Ltd.), Ioannis Digital Inclusion
Chatzigiannakis (Sapienza University of Rome) and 36 Using Localisation Technologies and Haptic
Simos Papadogeorgos (Power Made SA) Feedback for a More Inclusive Society
by Barbara Leporini (ISTI-CNR) and Maria Teresa
22 Circular Intelligence: Using a Smart Digital Paratore (ISTI-CNR)
Platform to Encourage the Collection of Used
Cooking Oil from Households Privacy-Preserving Computation
by Thanasis Gentimis, Theodore Chatzidimitriou and 38 Supporting Privacy Preservation by Distributed and
Antonis Kokossis (SymbioLabs) Federated Learning on the Edge
by Davide Bacciu (UNIPI), Patrizio Dazzi (CNR-ISTI)
23 Seamless Distributed Traffic Monitoring for Smart and Alberto Gotta (CNR-ISTI)
Cities Using Fibre Optic Acoustic Sensing
by Martin Litzenberger, Carmina Coronel and Christoph 39 CT Scans Reveal a Double Panel in an Oil Sketch
Wiesmeyr (AIT Austrian Institute of Technology after Rubens
GmbH) by Marta Domínguez-Delmás (UvA), Francien Bossema
(CWI), Erma Hermens (Rijksmuseum, UvA)
25 Forging the Future of Responsive Cities Through
Local Digital Twins
by Susie Ruston McAleer, Mark McAleer and Pavel
Kogut (21c Consultancy Ltd) ANNOUNCEMENTS
26 Recovering Non-Conventional Water Sources in the 40 26th Conference on Formal Methods for Industrial
Mediterranean: The HYDROUSA Project Critical Systems
Experience by Maurice ter Beek (ISTI-CNR)
by Simos Malamis, Stavroula Kappa, Eleni Nyktari and
Constantinos Noutsopoulos (National Technical 41 PerAwareCity & WSCC 2022 - Call for Papers
University of Athens)
42 W3C/SMPTE Joint Workshop on Professional
28 Vigilantes del Aire – Measuring Air Quality in Spain Media Production on the Web
by Daniel Bruno, Miguel Sevilla-Callejo, Enrique
Navarro (Pyrenean Institute of Ecology) and Francisco 42 ERCIM “Alain Bensoussan” Fellowship Programme
Sanz (Ibercivis Foundation)
43 Dagstuhl Seminars and Perspectives Workshops
30 A Holistic and Scalable Solution for Research,
Innovation and Education During the Energy
Transition
by Helen C. Leligou (University of West Attica) IN bRIEF
31 How Small Initiatives Create Smart City Dynamics: 43 Marta Kwiatkowska and Susan Murphy win CWI’s
The ICC Experience of Patras Van Wijngaarden Awards 2021
by Petros Ganos, Athanasios Kalogeras, Tanya Politi
and Lena Tsipouri
33 Digital Cities, Digital Tourism, Digital Arts – A
Research Agenda
by Christian Thomay and Markus Tauber (Research
Studios Austria FG), Christoph Schmittner (Austrian
Institute of Technology GmbH) and Beatriz Tadeo Fuica
(IRCAV, Sorbonne-Nouvelle University)
34 Circularity in Small Island Cities
by Sobah Abbas Petersen (SINTEF Digital) and Hanne
Cecilie Geirbo (Oslo Metropolitan University)
ERCIM NEWS 127 October 2021 3
Special Theme
Introduction to the Special Theme
Smart and Circular Cities
by Erwin Schoitsch (AIT) and Georgios Mylonas (ISI, Athena Research and Innovation Center)
Smart city technologies have been pro- section and a multidisciplinary communities, e.g., in rural areas. They
liferating at a rapid pace for some years approach. The fact that the implementa- discuss five preventive measures that
now, and at the same time, the divide tion and deployment of such systems might help to mitigate cyber incidents
between the natural and digital worlds involves the participation of a large towards smartification. When it comes
has lessened considerably. Multiple number of citizens has also opened a to IT security in urban environments,
sensing endpoints located in our envi- window for opportunities in fields such Lämmel et al. (page 7) discuss policies
ronment, offices, homes, devices, and as citizen science and co-creation. and recommendations towards the
even our body, produce continuous secure implementation and operation of
streams of sensor data. At the same time, This special theme of ERCIM News urban infrastructure. In this context, the
the concept of the circular economy has reports on academic and industry Urban Data Partnership (UDP) aims to
entered the mainstream, and we are now research that addresses technology, sys- accelerate the digital transformation of
seeing many communities and busi- tems, applications, and services in the cities and communities, while consid-
nesses adopt novel approaches based on Smart and Circular Cities domain. This ering data security. Moreover,
circularity (e.g., in the European Green section includes articles targeting Pustozerova and Mayer (page 9) discuss
Cities Network). European policy important elements such as security and aspects of using Federated Learning, an
(“Green Deal”) has reacted to the chal- privacy, as well as the integration of ML approach that is quickly gathering
lenges of climate change and environ- technologies like Digital Twins and the traction, for data security and privacy
mental footprint reduction, particularly Smart Grid, which are still evolving at a purposes in smart cities. They argue that
in the context of large urban agglomera- rapid pace. Experiences from deploy- although its adoption is growing, there
tion. Smartness (intelligence) must ments that use more conventional are still challenges to overcome, and
address not only the immediate goals of approaches from a technological stand- further investigation of defense mecha-
human wellbeing, assisted living and point offer us glimpses of the near nisms is needed. Furthermore, Klikovits
et al. (page 10) present an approach to
integrating the plethora of constantly
changing IoT devices and services
within a smart city setting, based on the
combination of an ID provider and the
Arrowhead framework.
Resilient Cities: Digital Twin and
Smart Grid
Moving on to novel technologies that
are quickly becoming part of the smart
city landscape, the Digital Twin is one
such technology. Digital Twins are cur-
rently being introduced to help us on
multiple fronts, including data security
and privacy. The SecurityTwin project,
Figure1:The17UNSustainableDevelopmentGoals(SDGs).Source:un.org. presented in this issue by Eckhart et al.
(page 12), aims to develop the funda-
mental methods for employing the dig-
comfort but, perhaps even more impor- future, while reports from smart city ital-twin concept to enhance the secu-
tantly, long-term sustainability, as projects using novel methodologies and rity of Cyber-Physical systems and pro-
defined by the 17 UN Sustainable technologies give us a better sense of vide the basis for implementing intru-
Development Goals (Figure 1). how smart city research will evolve to sion detection and response methods.
encompass new communities and tools. Meanwhile, the Smart Grid is another
In this environment, Smart Cities, AI field that has become a part of the
and the IoT, together with sustainability Security and Privacy Issues research landscape, with its importance
and circular economy, form a significant Kreutzer et al. (page 6) discuss a frame- accentuated by the need to accelerate
part of the current research landscape. work for facilitating the transition from towards more sustainable cities.
Several challenges have surfaced when existing smart city services to smart Efthymiopoulos et al. (page 13) present
designing and applying these systems in governance, arguing that thus far the FLEXGRID project, which investi-
fields like energy, sustainability, smart research has focused more on technical gates the constraints of current smart
transportation and digital twinning, aspects, leaving out of the equation grid architectures preventing integra-
especially when discussing their cross- more common ones, or even whole tion of large-scale distributed energy
4 ERCIM NEWS 127 October 2021
resources into distribution networks and anomalies. Such data can lead to indi- Sustainability projects in specific
aiming to mitigate circularity and sus- rect observations, e.g., indoor activity contexts
tainability in modern smart grids. The levels during the COVID-19 pandemic. Finally, there is the issue of focusing on
project is currently developing a digital Moreover, the circular economy is specific communities and co-creation to
platform to offer digital energy services entering our daily lives. Gentimis et al. deliver for smart and circular cities.
helping energy sector stakeholders to (page 22) present a digital platform that Ruston McAleer et al. (page 25) present
automate and optimise the planning, aims to encourage citizens to recycle the COMPAIR project, scheduled to
operation and management of systems used cooking oil in West Macedonia, begin in November 2021, which brings
and assets. Greece. The goal is for the platform to social and emotional intelligence into
be adopted by 10,000 households across the decision-making process within
At the same time, energy communities 13 municipalities in this area. cities. H. Lelligou (page 30) presents
have been forming throughout Europe the ASSET project, which focused on
to facilitate sustainable energy produc- Novel technologies and methods the educational community, an impor-
tion in a decentralised manner. Norbu et Going back to major issues at the heart tant community for the transition to
al. (page 15) discuss the Responsive of smart and circular cities, traffic mon- smart and circular cities. The project
Flexibility (ReFLEX) project, currently itoring, air quality and pollution moni- produced more than 25 training pro-
the UK’s largest smart energy demon- toring, as well as sustainable water grammes on energy transition, evalu-
strator, focusing on an energy commu- resource management immediately ated through pilot studies. Ganos et al.
nity at the Orkney Islands in Scotland spring to mind. Over recent years, the (page 31) present their experiences
comprising 200 households. They research community has attempted to from several smart city initiatives in the
developed algorithms for the smart con- tackle them using rather conventional city of Patras, Greece, and argue that
trol of energy assets and redistribution approaches. However, we are beginning even small communities with limited
mechanisms, achieving a fairer redistri- to see more innovative approaches resources can make a difference and
bution in the process. At a more adopted, utilising novel methodologies. facilitate change. Thomay et al. (page
localised scale, within home environ- In the context of real-time traffic moni- 33) tackle the issue of sustainable
ments, Constantinou et al. (page 17) toring, Litzenberger et al. (page 23) use tourism in smart cities, and how digital
propose the IMCF+ framework to facil- fiber optic acoustic sensing and the tourism can provide new ways to share
itate the smart consumption of energy at existing telecom fiber cable infrastruc- cities’ cultural heritage. Furthermore,
the time it is produced by e.g., photo- ture in a prototype study in Graz, Abbas Petersen and Geirbo (page 34)
voltaics on the roof of a building. Their Austria. Initial results suggest that it is discuss the Learning Flexibility project,
strategy relies on an AI-inspired algo- feasible to derive real-time traffic esti- focusing on waste management and cir-
rithm to schedule energy consumption mation using this approach. Ruston cularity in regions such as small islands.
of various devices, using a variety of McAleer et al. (page 25) discuss the The project takes a bottom-up
strategies while at the same time emerging concept of city-scale Digital approach, in contrast to the more con-
meeting indoor comfort level require- Twins and present the DUET project. ventional top-bottom approach adopted
ments. The pilots developed in the project by many related projects, aiming to
across three cities are expected to go identify innovative and sustainable
Optimising the use of established live in autumn 2021, tackling mobility solutions.
methods and tools and air quality, city planning and public
Several projects utilise more estab- decision-making. Bruno et al. (page 28) Overall, research on smart and circular
lished tools and technologies in the cir- discuss the use of plants, specifically cities in Europe appears to be moving
cular and smart city domain – for strawberries, as biosensors to monitor towards using innovative technologies
example, electric bikes and cars are rap- air quality. They report on a large-scale and approaches, reaching to novel
idly becoming part of the urban trans- deployment across Spain in the application domains, as well as encom-
port landscape. Gunner et al. (page 18) “Vigilantes del Aire” project, utilising passing communities in a more active
discuss results obtained by fitting moni- samples from 205 municipalities and 26 and engaging manner.
toring equipment to a fleet of electric Spanish provinces. Malamis et al. (page
bikes that were deployed as part of the 26) address sustainable water resource Please contact:
H2020 Lighthouse Project REPLI- management as investigated by the Erwin Schoitsch
CATE in Bristol, UK. This has pro- HYDROUSA project, which uses inno- Austrian Institute of Technology,
duced a dataset that could help us better vative nature-based solutions (NBS) to Austria
understand aspects such as route selec- manage a variety of water streams. The erwin.schoitsch@ait.ac.at
tion by e-bike users or identify city project promotes decentralised on-site
areas where cycling infrastructure water, materials and energy conserva- Georgios Mylonas
would be most beneficial. Smart water tion, treatment and reuse. It has estab- ISI, Athena Research and Innovation
metering is another quickly evolving lished large-scale demonstrators in Center, Greece
domain, following electric power three Greek islands, while its solutions mylonasg@athenarc.gr
metering. Amaxilatis et al. (page 20) are being evaluated in 25 early-adopter
describe Tethys, a large-scale water cases in other Mediterranean coastal
metering deployment in Thessaloniki, areas.
Greece, using water consumption data
to identify patterns, behaviours and
ERCIM NEWS 127 October 2021 5
Special Theme
Smart Governance for Cybersecurity
by Michael Kreutzer and Kirstin Scheel (Fraunhofer SIT)
The topic of smartification has become ubiquitous; urban planners and public organisations are
investing heavily in digitalisation projects. At the same time, cybersecurity often seems to still be a
sideshow. Is there a way to get from smart cities to smart governance?
The Fraunhofer Institute for Secure mendations of this governance frame- be aware of the necessity of security as
Information Technology SIT is one of work are transferable and help a cornerstone of digitalisation and smar-
the world’s leading research institutes strengthen the cybersecurity of all smart tification projects. If you want your
for cybersecurity and privacy protection. community projects. projects in this area to be sustainable,
It is part of the National Research Center forward-thinking, and accepted by citi-
for Applied Cybersecurity ATHENE and We have identified five preventive zens and public employees alike, cyber-
as such is an essential part of the cyber- measures that might help to mitigate security needs to be embedded in the
security strategy of the Federal cyber incidents in the process of smarti- organisational culture.
Government and the State of Hesse. fication. We intend these principles to
be applicable as a governance frame- Clear responsibilities need to be
As part of a project for the Hessian work on all areas of smart communities. assigned. This is particularly true in
Ministry of the Interior and Sports on The framework, represented in Figure hierarchically structured organisations.
cybersecurity processes within and 1, is based on the principles of: Lack of responsibility or diffusion
across Hessian municipalities, our (i) anchoring, thereof can counteract preventive meas-
research has led us to realise that existing (ii) responsibilities, ures. In addition, appropriate resources
work on smart city developments has (iii) unification, to act on said responsibilities are
tended to focus on technical challenges (iv) co-operation, and required. For example, reporting chan-
and/or theoretical attack scenarios. It (v) improvement. nels and response times need to be
emphasises specific aspects of different defined. It does not end there, though –
types on attack vectors, privacy impacts One thing to consider is that these are they also need to be continually prac-
and also considerations for transforma- not meant to be building on each other – ticed and lived in everyday life.
tive frameworks to developing smart each principle is equally important and Convenience, unsafe habits and
cities or protecting them from specific all need to be in effect to bring about the “workarounds” are the bane of not just
attacks. In addition, many papers still intended outcome. Considering that the cybersecurity. So, from an organisa-
only talk about smart cities when in fact most ingenious smart development is tional psychological perspective,
rural communities also benefit from digi- likely happening as part of an existing raising awareness and strengthening a
talisation – hence our preference for the legacy system, it becomes clear that the culture of responsibility is important.
term “smart communities”. governance needs to be embedded to
bring about a cohesive cybersecurity Another central idea is the unification
However, what seems to be missing in framework. across organisational units. In the field
the race to “smartification” is a frame- of IT this can mean a competently set up
work that helps formulate an encom- Cybersecurity needs to be anchored at infrastructure to prevent failures. Many
passing governance perspective for all the top level. Top management needs to cases of malware infestation can spread
projects, which ensures that cybersecu-
rity underpins all digital developments.
Our goal is an integrated multidiscipli-
nary security framework.
We started our project with a broad liter-
ature review, with a specific focus on
publications from/on Germany’s federal
structures [1], as the original research
ordinance is focused on the state of
Hesse [2]. In addition, we launched a
preliminary and continuing review of
publicly available sources on real cyber-
attacks on public infrastructures. We
supplemented this with structured inter-
views with public sector officials in the
field of cybersecurity. The observations
we are presenting here are a preliminary
culmination deduced from these and
need to be tested in practice.
Nonetheless, we believe that the recom- Figure1:Proposedstepstowardssmartgovernance.
6 ERCIM NEWS 127 October 2021
through systems that are not properly Dynamically changing environments References
segmented. IT experts are needed – and require continuous improvement. [1] G. R. Wollinger and A. Schulze
they need to continually update their Learning from internal and external mis- Eds.: “Handbuch Cybersecurity für
skills. takes is essential to keep up with these die öffentliche Verwaltung,
developments. Innovations as well as Wiesbaden: Kommunal- und
Operational co-operation and cross- paradigm shifts are the norm, especially Schul-Verlag, 2020”, [online]
divisional collaboration are also in the digital world. It is necessary to available: https://kwz.me/h8a
important. Especially in the public make learning an integral part of the [2] J. Remy and R. Stettner:
sector, resources need to be used effi- organisational culture. In the field of IT “Cybersicherheit als Aufgabe der
ciently and effectively. With regard to this is nothing new and is usually referred Länder,” Datenschutz Datensich,
cybersecurity, this can help achieve a to in the form of maturity models. vol. 45, no. 4, pp. 254–258, 2021,
higher level of protection overall. For However, we believe that this needs to be doi: 10.1007/s11623-021-1429-y.
complex IT systems, such as those in rooted at the heart of the entire organisa-
smart communities, to work securely tion and smart community. Please contact:
together, the different units need to be Kirstin Scheel
connected. Information needs to flow – We will continue our research in this Fraunhofer Institute for Secure
and to actually be used as well as field in the coming months as part of the Information Technology SIT, Germany
processed. Vertical and horizontal net- current project and hope to expand on it kirstin.scheel@sit.fraunhofer.de
working is needed. in the future. +49 6151 869 268
Policies and Recommendations for IT Security
in Urban Environments from the Morgenstadt
Urban Data Partnership Project
by Philipp Lämmel, Michell Boerger, Nikolay Tcholtchev (Fraunhofer FOKUS) and Eva Ottendörfer
(Fraunhofer IAO)
Urban ICT infrastructure is playing an increasingly decisive role as the technical backbone of smart
cities. To guarantee the protection of the public sector and citizens in this context, the security of this
infrastructure is of utmost importance and should be continuously monitored and improved. This article
presents measures and recommendations towards ensuring the security of urban ICT infrastructures.
The smart cities domain is becoming One aim of this initiative is to stimulate urban use cases. Therefore, all stake-
ever more relevant for our society. The the transfer of knowledge between holders, including the city government,
accelerating digitalisation of processes cities/communities by creating common should be encouraged from the begin-
in urban settings is expected to lead to knowledge as well as sharing experi- ning to create a culture of cybersecurity
long-term improvements, enhancing the ence and strategies regarding the effi- throughout all the involved public enti-
quality of life of inhabitants and creating cient and secure management of urban ties.
more liveable, sustainable, and inclusive data. In the long term, the UDP aims to
cities. Information and communications accelerate the digital transformation of Apply security frameworks and
technology (ICT) plays an essential role cities and communities, while consid- standards
as the backbone of digital transforma- ering (data) security in an urban envi- To secure a smart city/community, the
tion. New optimisation opportunities are ronment. Based on knowledge gleaned security of the ICT infrastructure must
arising due to the ICT-enabled emerging from the UDP, this article presents be addressed as early as the conception
capabilities for combining and evalu- measures and recommendations to phase. Security is important at every
ating new services and data sources. In ensure the security of urban ICT serv- step of the development lifecycle and
addition, digitalisation and the accompa- ices and systems in smart urban envi- vulnerabilities should be avoided at
nying transformation of the economy ronments. The fundamental policies and every level. To this end, the National
and our everyday lives offer the poten- recommendations are discussed below Institute of Standards and Technology
tial to optimise fundamental urban and summarised in Figure 1. (NIST) has published a cybersecurity
processes, e.g. in the domains of framework [1] covering many topics.
mobility, transportation and energy. Stakeholder engagement and This framework is a must-read for
governance anyone involved or interested in
To ensure that cities and communities do An open ecosystem of diverse stake- improving security in their city, com-
not have to face these diverse challenges holders who are aware of the impor- munity or organisation.
on their own, the Urban Data tance of cybersecurity in a smart city is
Partnership (UDP) was founded by the a fundamental driver for the sustainable In addition, in 2002, the OECD pub-
Fraunhofer Morgenstadt network [L1]. and secure implementation of smart lished revised guidelines for informa-
ERCIM NEWS 127 October 2021 7
Special Theme
tion systems and network security,
underpinned by nine principles [2]: (1)
awareness, (2) responsibility, (3) Stakeholder Engagement &
response, (4) ethics, (5) democracy, (6) Gover nance
Governance
risk assessment, (7) security design and
implementation, (8) security manage-
ment, and (9) reassessment. NIST
expanded upon these principles in their Apply Security
document Systems Security v
Avoid
Avoid kn
known Security
Frameworks &
Vulnerabilities
Vulnerabilities
Engineering: Considerations for a Standards
Standar
Multidisciplinary Approach in the
Engineering of Trustworthy Secure URBAN
Systems [3]. This document provides a ICT
SECURITY
taxonomy of security design principles
to be used as a basis for engineering
trustworthy, reliable, and secure sys-
tems.
Cover Train
Train
Avoid known security vulnerabilities Security Basics Security Personnel
and threats
Since software is becoming increas-
ingly complex and interconnected, the Figure1:Overviewofidentifiedmeasuresandrecommendationswhicharecrucialforensuring
difficulty of achieving application secu- thesecurityofurbanICTservicesandsystems.
rity is also increasing exponentially.
Therefore, the Open Web Application
Security Project (OWASP) published • Strong access controls: All systems Links:
the ten most critical security risks for that are not currently in use should be [L1] https://kwz.me/h79
web applications [L2]. These have disabled. Unused remote manage- [L2] https://kwz.me/h7f
become the de facto standard for appli- ment functions and ports should also
cation security. We recommend that all be disabled to prevent attackers from References:
actors involved in developing an urban accessing them. Furthermore, net- [1] NIST, “Framework for Improving
ICT infrastructure study the risks and work activities should be scanned Critical Infrastructure
resulting measures identified by the regularly, and suspicious internet Cybersecurity, Version 1.1,”
OWASP. traffic should be monitored with the National Institute of Standards and
help of security incident and event Technology, Gaithersburg, MD,
Cover security basics management tools to detect attacks at NIST CSWP 04162018, Apr. 2018.
The following security basics should be an early stage. DOI:
followed: 10.6028/NIST.CSWP.04162018.
• On-time software updates: All soft- Train security personnel [2] OECD Guidelines for the Security
ware used in an urban ICT environ- A further security-related challenge is of Information Systems and
ment should be kept up to date, so the training of staff to secure an urban Networks. OECD Publishing,
that no known security vulnerabilities infrastructure. Due to the rapid growth 2002. doi:
can be exploited. All firewalls and and expansion of smart cities, there is 10.1787/9789264059177-en-fr.
antivirus programs should be updated currently a shortage of security experts [3] R. Ross et al., “Systems Security
regularly. in the urban context. Therefore, the Engineering: Considerations for a
• Enforce secure passwords and poli- training and certification of profes- Multidisciplinary Approach in the
cies: Users should regularly update sionals for the development, construc- Engineering of Trustworthy Secure
their passwords to ensure that they tion, operation, and maintenance of Systems,” National Institute of
are unique and complex. Strict poli- urban ICT infrastructures should be Standards and Technology, NIST
cies should be enforced to ensure that urgently promoted. SP 800-160, Nov. 2016. DOI:
passwords are secure. Furthermore, 10.6028/NIST.SP.800-160.
establishing security operation cen- Summary
tres could be helpful to monitor secu- In summary, particular policies and rec- Please contact:
rity, mitigate vulnerabilities, and ommendations should be followed for Philipp Lämmel
respond to attacks. the secure implementation and opera- Fraunhofer Institute for Open
• Correct operating procedures: tion of urban ICT infrastructures, Communication Systems FOKUS,
Deploying firewalls is an important namely: stakeholder engagement and Germany
step in protecting a smart city/com- governance; application of security philipp.laemmel@fokus.fraunhofer.de
munity. Determining the type of traf- frameworks and standards; avoidance
fic allowed to pass through the fire- of known security vulnerabilities;
wall is one of the most central ways training of personnel; coverage of secu-
to protect a network from potential rity basics; and the establishment of
attacks. adequate security processes.
8 ERCIM NEWS 127 October 2021
Data Privacy in Smart Cities
– Federated Learning to the Rescue?
by Anastasia Pustozerova and Rudolf Mayer (SBA Research)
Within any smart system, data is vital for making the management of resources and assets more
efficient. At the same time, data is a potential vulnerability to data owners, and it could become a
threat in the hands of an adversary. Data security and privacy are therefore critical for building
sustainable smart systems like smart cities. In such systems, where data collection is distributed,
federated learning seems like a prime candidate to address the issue of data privacy. However, there
are still concerns that need to be addressed regarding privacy and security in federated learning.
Machine learning demands large and new attack vectors open up simply models on poisoned data or directly
amounts of data to build effective due to the distributed nature of feder- manipulate model updates [3]. An adver-
models that can help to improve serv- ated learning (see Figure 1). sary who is able to compromise the
ices. In many real-world scenarios, data aggregator can perform attacks on the
originates at the edge, e.g., smart meters Security risks (integrity and availability). global model. Another threat comes from
and sensors in smart power grids. In tra- Malicious participants of federated non-secure communication channels
ditional machine learning workflows, learning or adversaries leveraging trans- when an adversary is able to steal or mali-
data must be centralised from different ferred information can corrupt the ciously modify shared model updates.
sources before performing the model learning process to degrade the global
training. Concentrating all data in one model quality or to make it perform target Privacy risks (confidentiality). Model
place creates a single point of failure – misclassification. In smart cities, success- parameters exchanged during federated
an adversary that can potentially gain fully executed attacks can result in adver- learning represent an abstraction of the
access to this centralised data is a threat saries manipulating situations to favour training data. Adversaries might infer
to multiple entities. them – for example, by manipulating information about training data having
demand-driven pricing – or can even access to the model. In smart cities, data
Federated learning enhances data pri- result in the failure of critical services and generated by sensors and IoT devices
vacy in machine learning by suggesting infrastructure, and thus lead to major often involves personal privacy, and
a new perspective on applying machine safety issues. Security risks in federated this is thus a great concern. It is thus
learning for the analysis of distributed learning can originate through data or important to mitigate potential leaks of
data. The main idea is to train machine model poisoning (backdoor attacks), or this data through the machine learning
learning models closer to the place when an adversary alters the data at infer- process. Federated learning with the
where data originates – and just aggre- ence time (evasion attack). Backdoor increased exchange of models might,
gate these trained models instead of the attacks pose one of the biggest challenges however, increase the attack surface.
(sensitive or private) data. Federated in federated learning as they are espe- Adversaries can perform different
learning, therefore, eliminates the need cially hard to detect. The challenge is attacks on shared models in federated
to share and centralise sensitive data, increased by the secret nature of local learning, e.g., model inversion, trying to
allowing data owners to keep it private training data, which makes it hard to recreate the original samples from the
while at the same time offering compa- analyse the correctness of the contribu- model, or membership inference,
rable effectiveness of models. tion of clients. Malicious clients can train aiming to infer the membership of some
Federated learning architectures often
consist of data owners (clients), which
perform local training of the models on
their own data, and a central aggregator,
which collects the models from the
clients and averages them, producing a
global model. The global model can be
sent back to the clients for the next cycle
of training to improve its effectiveness,
and later utilised for predictions. Some
of the main challenges of federated
learning include communication costs,
data and systems heterogeneity. Many
works propose different optimisation
algorithms to tackle these issues, e.g.,
via client sampling or model and gra-
dient compression [1]. However, com- Figure1:Federatedlearningarchitectureandattackvectors.Anadversarywhoisableto
paratively little attention has been put on compromiseclients,aserverorcommunicationchannelscanthreatenthesecurityandprivacy
remaining privacy and security risks, ofthesystem.
ERCIM NEWS 127 October 2021 9
Special Theme
particular instance in the training set of Homomorphic Encryption (HE) allows in the technology. Further investigation
a target model [2]. Adversaries can be mathematical operations to be per- of defence mechanisms is therefore crit-
e.g., a compromised or malicious aggre- formed on encrypted data. Clients can ical for the successful application of
gator, or someone stealing models encrypt their model parameters, and the federated learning.
during client-server communication. coordinator could aggregate them but
not understand them. Like SMPC, HE This work has received funding from
Approaches for mitigating security and greatly increases computational costs. the European Union’s Horizon 2020
privacy risks in federated learning often research and innovation programme
still lag behind attacks, but are increas- Detecting attacks on the integrity and under grant agreement No 826078.
ingly in the focus of research activities. availability of the machine learning
process is even more difficult. Defences References:
Regarding privacy risks, several like anomaly detection and robust [1] P. Kairouz, H. Brendan McMahan,
approaches can be employed. aggregation aim to discover potentially et al.: “Advances and Open
Differential privacy (DP) aims to bring harmful models and eliminate their Problems in Federated Learning”,
uncertainty into the model outputs to malicious influence on the global Foundations and Trends in
hide personal contributions to the model. Yet they fail to detect targeted Machine Learning: Vol. 14: No.
model; clients can add noise to shared backdoor attacks, as poisoned models 1–2, pp 1-210, 2021.
model parameters or train a differen- look and behave similarly to models [2] A. Pustozerova and R. Mayer:
tially private machine learning algo- that were trained without backdoor [3]. “Information leaks in federated
rithm. The main downside of this learning”, in proc. of the Workshop
approach remains that noise degrades There has been a dramatic increase in on Decentralized IoT Systems and
models performance, thus there is a interest in federated learning in recent Security (DISS), 2020.
trade-off between privacy and utility. years. Many companies, including [3] N. Bouacida and P. Mohapatra:
Apple and Google, are already using “Vulnerabilities in Federated
Secure Multi-Party Computation federated learning for their services. Learning”, in IEEE Access, vol. 9,
(SMPC) provides a cryptographic pro- Interest in this technology is especially pp. 63229-63249, 2021.
tocol that allows joint computation of a high in medical applications and smart
function while keeping its inputs pri- cities, where personal data is processed, Please contact:
vate. In federated learning, this can and data privacy is a major concern. Anastasia Pustozerova
replace a central aggregator. However, However, there are still challenges to SBA Research, Austria
SMPC poses high computational costs, address in federated learning. apustozerova@sba-research.org
therefore limiting the scalability of fed- Mitigation of security and privacy risk
erated learning. is especially important for building trust
Considering Cybersecurity with Trustworthy IoT
in Smart Cities
by Christoph Klikovits (Forschung Burgenland), Clemens Gnauer (Forschung Burgenland), Patrik Abraham
(Fachhochschule Burgenland)
In today’s smart cities, the question remains how to securely integrate a multitude of different and
constantly changing Internet of Things (IoT) devices and services. This is where we propose the
combination of an identity provider (e.g.: ID-Austria [L3]) and the Arrowhead framework [L2] to verify
sensors by matching them with a known legal identity. By providing an application that assures a
secure authentication and trustworthy communication for people, sensors, and services in a smart city.
A variety of technological innovations this topic and researches an approach to cities can rely on secure and trustworthy
have changed the characters of cities in increase the trustworthiness and secu- IoT data. Furthermore, the Arrowhead
recent years. There are millions of rity of IoT devices. The aim of this project is implementing more and more
devices with sensors and actuators dis- approach is to link existing IoT devices support services like the on-boarding
patched with an upward trend in cities or services with a known legal identity. procedure [2] which is used for the
(weather, water and gas metering, traffic An identity provider (ID-Austria) and autonomous integration of devices into
lights and controls, waste management, an additional support service of the the service-oriented arrowhead
etc.). Applying and using evermore of Arrowhead framework are combined ecosystem. This procedure strengthens
these in the context of interconnected through an integrated approach. the secure and trustworthy integration
IoT systems raises the challenge of pro- Arrowhead was created for the orches- of devices or services. A proof of con-
viding trust and security in this context. tration of large scale IoT data. It offers cept could be developed in the EFRE
The Center for Cyber Security of strong security mechanisms and in com- project (FE07) “Civis 4.0 Patria” and
Forschung Burgenland [L1] deals with bination with the admission ticket, presented in the FIP / IEEE
10 ERCIM NEWS 127 October 2021Figure1:Architectureoftrustworthy
IoTinsmartcities.
International Symposium on Integrated person or organization is verified and sion ticket, the backend is informed by
Network Management (IM) workshop clearly identified. The identified person the ATP and allowed to publish the pay-
[1]. or organization can use their identifica- load into the frontend. If an admission
tion (bPK) to register various devices or ticket is invalid or does not exist, the
As shown in Figure 1, various tech- services in the backend layer. When frontend release of the payload is not
niques and tools are used to create a dig- registering, various additional parame- permitted and is discarded. As applied
ital admission ticket, which is created ters (IP-Address, Mac-Address, device in a proof of concept, it shows trust-
based on several parameters (IP, MAC ID, etc.), related to the device or service worthy and secure communication in
address, etc.). In addition to the parame- can be specified. These parameters, smart cities, where smart devices or
ters listed, the electronic proof of iden- including a legal identity (bPK), are for- services are increasingly used.
tity from an identity provider is inte- warded to the Arrowhead framework. Matching a unique personal identifier
grated into the creation of the digital Additionally, to the core services of the with devices or services enables to link
admission tickets. Three instances and Arrowhead framework, a further a person responsible with e.g., IoT sen-
steps linked in an integrated approach service called the Admission Ticket sors, whereby trustworthiness, accept-
are required for the creation of an Provider (ATP) was developed. The ance and security of devices and serv-
admission ticket. Step 1: Firstly, a ATP is responsible for generating a hash ices in a smart city can be strengthened.
person or organization is involved in the (admission ticket) by combining the
proposed proof of concept shown in received device or service parameters Links:
Figure 1. This natural person or corpo- (e.g., IP address or Mac-address), the [L1] https://www.forschung-
rate body requires an electronic proof of unique personal identifier (bPK) and burgenland.at/cybershysecurity/
identification from an identification using the SHA-256 function. This [L2] https://www.arrowhead.eu
provider (e.g., ID-Austria [L3]). While delivers a string with 64 characters [L3] https://kwz.me/h7J
logging into a backend layer (e.g., Civis which will be stored in the ATP-data- [L4] https://kwz.me/h7M
4.0 Patria backend), the login data is base and transmitted to the backend for
verified by the identification provider further use. Afterwards, the person or References:
using an interface. By using the inter- organization who registered the device [1] C. Klikovits, P. Abraham and R.
face to the identification provider, the or service will receive the admission Rambacher: “A Framework to
person or organisation does not need ticket and must store it on the device identify People, Devices and
any additional login data for the (e.g., sensor, Raspberry Pi or smart- Services in Cyber-physical system
backend layer. Step 2: In the verifica- phone) or service-platform. of systems,” 2021 IFIP/IEEE
tion process the identification provider International Symposium on
determines an unique personal identi- The certified hash (admission ticket) Integrated Network Management
fier, called bPK [L4], in two steps: created by the ATP is stored on both the (IM), 2021, pp. 914-919.
Firstly, a character string is formed from respective device or service and in the [2] A. Bicaku, et al.: “Interacting with
a master number (central register or ATP database. An implemented the arrowhead local cloud: On-
association register, commercial reg- backend process automatically for- boarding procedure”, 743-748.
ister entry) and the procedural area. wards the respective payload including 10.1109/ICPHYS.2018.8390800,
Secondly, a specific hash algorithm cal- the supplied admission ticket to the ATP 2018.
culates a secure one-way cryptographic in the Arrowhead framework every time
derivation from this character string and a device or service transmits data. Please contact:
encodes the bPK with the Base64 stan- Furthermore, the admission ticket of the Christoph Klikovits
dard. Step 3: After the bPK has been device or service is compared with the Forschung Burgenland, Austria
transmitted from the identification stored admission ticket in the ATP. After christoph.klikovits@forschung-
provider to the backend layer, the a successful verification of the admis- burgenland
ERCIM NEWS 127 October 2021 11Special Theme
Digital Twins for Cyber-Physical Threat Detection
and Response
by Matthias Eckhart, Andreas Ekelhart (SBA Research and University of Vienna), and Roland Eisl (ENRAG)
Since cyber-physical systems are the backbone of smart cities and innovative industrial applications,
their safe and secure operation is paramount. However, due to the steadily increasing aggressiveness,
sophistication, and stealth of cyberattacks, new methods for threat detection and response are needed.
The concept of digital twins opens up new avenues of research to address these gaps.
Recent security incidents involving phases of its lifecycle, consumes real- parallel to their physical counterparts,
cyber-physical systems (CPSs), such as time and historical data if required, and closely following their states, provide
the 2021 Colonial Pipeline cyberattack, has sufficient fidelity to allow the the means to inspect the behaviour of
have again demonstrated the vulnera- implementation of the desired security the CPS without the risk of interference.
bility of critical infrastructure. While measure.” [3] Since digital twins are not This unique feature allows rigorous
the current state of CPS security is used for redundancy purposes when monitoring of multiple CPS layers (e.g.,
already strained, smart technology applied within the context of security, physics, network, logic) and can be
trends proceed to evolve, pushing tradi- the CPS is virtually replicated by means exploited for detecting intrusions.
tional protection mechanisms to their of emulation, simulation, and modelling However, such a security-focused use
limits. As a result, new methods to sup- techniques to an extent that enables the case necessitates a state replication
port the implementation of a holistic implementation of security-enhancing mechanism to keep the digital twins in
security approach are needed. features and activities. For example, sync with their physical counterparts,
Considering the interdependency of the digital twins that possess a sufficient and further assumes that the virtual
cyber and physical domains in which degree of fidelity allow thorough secu- replicas exhibit benign behaviour [2]. If
these systems function, adequately pro- rity testing during both the engineering an alarm is raised, the digital twins can
tecting CPSs represents a pressing chal- and the operation phase [1]. This use then be used to identify possible coun-
lenge. A few years ago, researchers case of the digital-twin concept spares termeasures and to assess their effec-
started to explore how the concept of systems integrators and operators of tiveness as well as their effects on the
digital twins can be utilised to tackle CPSs the need to build custom testbeds physical process from a simulation
this challenge [3]. or conduct security tests with the real point of view. As initial efforts were
infrastructure, thereby providing cost directed toward developing the basic
Within the context of security, the term savings and preventing uncontrolled principles of this concept [3], more
“digital twin” can be defined as “... a vir- interactions with live systems that may research is required to efficiently create,
tual replica of a system that accompa- lead to extensive (physical) damages. operate, and maintain these security-
nies its physical counterpart during Furthermore, digital twins that run in focused digital twins.
Lifecycle
Engineering Operation
Physical Environment
Basic & Detaile d Engineering On-Site Eng. Cyber-Physical Sys tem R obot 1
MC AD
Pr oduct E CAD S witch 1
HM I 1
Virtual
S en so r 1
Pr oce ss Re sour ce
CSE ng. Commiss ioning
.. . PLC 1 Mo tor 1 Co nvey or B elt 1
Backp ro pagatio n & Feed back
Passive Dat a Collectio n
Engine ering Data Sources
er
Sim ulation &
Ingest io n
Data Laye
Data ͙
... Emulation ͙
...
Knowledge S ystem Log s Netwo rk S en so r
LOG
Knowledge Base Traffic Measurem ents
Parsin g & Semant ic Lifting
Framework Sp ec.
Intrusion Dete ction Me thods
Digital Twins
Virtual Environment
Ph ysical Pro cess
Respons e Mechanisms
Generation HM I 1
S witch 1
Securityy Tes ting
Co nvey or B elt 1, S en so r 1,
e s in g
PLC 1
St imu li
St imu li
Mo tor 1, R obot 1
State
Proces
Replication Ph ysical Pro
o cess & I/O S imulation
Sy stem Emu lation & Simulation
Net wo rk St ack Emulat io n
Figure1:High-levelarchitectureofthedigital-twinframework.
12 ERCIM NEWS 127 October 2021The SecurityTwin project [L1] aims to manner but also of recovering the dig- References: develop the fundamental methods for ital twins from state mismatches. [1] M. Eckhart, A. Ekelhart: “Towards employing the digital-twin concept to Security-Aware Virtual enhance the security of CPSs. As part of Using the architecture we described, Environments for Digital Twins”, this project, researchers at SBA our framework will provide the basis Proc. of the 4th ACM Workshop on Research and the University of Vienna, for implementing intrusion detection Cyber-Physical System Security. together with industry professionals at and response methods. Owing to the ACM, 2018. ENRAG and condignum, will create a physical models and simulations inte- [2] M. Eckhart, A. Ekelhart: “A framework to efficiently build digital grated into the digital twins, the Specification-based State replicas of CPSs based on engineering designed intrusion detection system Replication Approach for Digital data, emulating components as well as incorporates knowledge about the phys- Twins”, Proc. of the 2018 networks, and simulating physical ical process under control and thereby Workshop on Cyber-Physical processes. Figure 1 illustrates the archi- will yield alerts if the process is steering Systems Security and Privacy. tecture of the digital-twin framework on toward an unintended state. Upon ACM, 2018. a high level. We aim to develop a detection of adverse events, response [3] M. Eckhart, A. Ekelhart: “Digital knowledge base that incorporates measures can be identified and their Twins for Cyber-Physical Systems know-how from numerous heteroge- applicability, as well as consequences, Security: State of the Art and neous data sources (e.g., engineering assessed by observing the behaviour of Outlook”, Security and Quality in data repositories, domain knowledge) the virtual replicas. Cyber-Physical Systems and provides the semantic foundation Engineering. Springer, 2019. for generating the digital twins. This Building upon our earlier work [1, 2, 3], knowledge base comprises: (i) informa- we are currently in the process of devel- Please contact: tion about the CPS itself (sourced from oping the framework as described Matthias Eckhart, Andreas Ekelhart engineering artifacts); (ii) information above. In addition to our contribution as SBA Research and University of concerning the simulation and emula- part of the SecurityTwin project [L1], Vienna, Austria tion used as part of the digital twins; and we want to actively stimulate scientific meckhart@sba-research.org, (iii) operational data from the real CPS exchange in this emerging research aekelhart@sba-research.org for state replication and intrusion detec- area. We are therefore organising the https://www.sba-research.org/ tion. The digital twins can then be auto- Dagstuhl seminar 22171 [L2], which is https://www.sqi.at/ matically generated by instructing the dedicated to this topic and are encour- integrated emulation solutions (e.g., aging other researchers to share their Roland Eisl QEMU) and initialising the embedded perspectives. ENRAG, Austria simulation models. Moreover, a syn- roland.eisl@enrag.at chronisation mechanism will be devel- Links: https://www.enrag.at/ oped, which is not only capable of auto- [L1] https://kwz.me/h7j matically replicating states in a timely [L2] https://kwz.me/h7q Circularity and Sustainability in Modern Smart Grids Through Innovative Energy Market Architectures by Nikolaos Efthymiopoulos, Prodromos Makris, Emmanouel Varvarigos (National Technical University of Athens) Circularity and sustainability in modern smart grids require open data models that can support dynamic and efficient distribution-network-aware energy management. In this context, the FLEXGRID [L1] project is developing a digital platform that will offer digital energy services (DESs) that help energy sector stakeholders (i.e., Distribution System Operators (DSOs), Transmission System Operators (TSOs), market operators, Renewable Energy Sources (RES) producers, retailers, flexibility aggregators) to: (i) automate and optimise the planning, operation and management of their systems and assets, and (ii) interact in a dynamic and efficient way with the electricity system and other stakeholders. The large-scale integration of stated in the EU Clean Energy Package The first reason is that DSOs use con- Distributed Energy Resources (DERs), [1]. In this context, the FLEXGRID servative constraints in distributed DER such as PV/wind generation (RES), project is investigating the constraints installation to ensure reliable and secure electric vehicles (EVs), energy storage of the current smart grid architecture operation of their network. The root systems (ESS) and demand side man- that prevent large scale DER integration cause of this conservatism is the agement (DSM) equipment in distribu- in distribution networks and conse- inability of DSOs to dynamically and tion networks poses new challenges and quently mitigates circularity and sus- accurately monitor and manage their opportunities for the power sector, as tainability in modern smart grids. networks. The development of a ERCIM NEWS 127 October 2021 13
You can also read
