Industry 4.0 and Engineering for a Sustainable Future [1st ed.] 978-3-030-12952-1;978-3-030-12953-8

Table of contents :
Front Matter ....Pages i-xviii
Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead (Mohammad Dastbaz)....Pages 1-11
Why Industry 4.0? (Peter Cochrane)....Pages 13-21
Connectivity for Industry 4.0 (Kristina Gold, Kenneth Wallstedt, Jari Vikberg, Joachim Sachs)....Pages 23-47
Wireless Comms. Beyond 2020 (William Webb)....Pages 49-64
Digital Transformation (Paul Graham)....Pages 65-76
Big Data, Small Data, and Getting Products Right First Time (Human Ramezani, Andre Luckow)....Pages 77-90
The Internet of Things and Sustainable Manufacturing (David Heatley, Mohamed Abdel-Maguid)....Pages 91-115
Security Challenges in the Industry 4.0 Era (Mohammed M. Alani, Mohamed Alloghani)....Pages 117-136
Meeting the Future Challenges in Cyber Security (Ben Azvine, Andy Jones)....Pages 137-152
Ready for Industry 4.0? The Case of Central and Eastern Europe (Wim Naudé, Aleksander Surdej, Martin Cameron)....Pages 153-175
From Big to Small Data (Peter Cochrane, Ahmed Elmagarmid)....Pages 177-190
The Role of Blockchain in Underpinning Mission Critical Infrastructure (Hamid Jahankhani, Stefan Kendzierskyj)....Pages 191-210
The Planning and Design of Buildings: Urban Heat Islands—Mitigation (Christopher Gorse, James Parker, Felix Thomas, Martin Fletcher, Graham Ferrier, Neill Ryan)....Pages 211-225
Back Matter ....Pages 227-235

Citation preview

Mohammad Dastbaz Peter Cochrane Editors

Industry 4.0 and Engineering for a Sustainable Future

Industry 4.0 and Engineering for a Sustainable Future

Mohammad Dastbaz  •  Peter Cochrane Editors

Industry 4.0 and Engineering for a Sustainable Future

Editors Mohammad Dastbaz University of Suffolk Ipswich, Suffolk, UK

Peter Cochrane Cochrane Associates Limited University of Suffolk Ipswich, Suffolk, UK

ISBN 978-3-030-12952-1    ISBN 978-3-030-12953-8 (eBook) https://doi.org/10.1007/978-3-030-12953-8 Library of Congress Control Number: 2019936003 © Springer Nature Switzerland AG 2019 This work is subject to copyright. All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

Acknowledgements

We would like to thank all the contributing authors for their tireless work and for providing us with their valuable research work which has made this edited volume possible. We would also like to thank the Springer editorial and production team, Amanda Quinn, Brian Halm and Brinda Megasyamalan, for their patience and valuable advice and support. Special thanks go to Atlanta Blair who worked tirelessly in organising our almost impossible schedules, getting all the necessary forms done and sending numerous e-mails and gentle reminders when necessary, and also to Jane Cochrane, for helping Peter and I to get together for some key discussions around the book and its key concepts. Finally, our thanks go to all our colleagues at the University of Suffolk; BT Research Centre at Adastral Park; Matrixx; Ericsson; BMW; Maastricht University; Saïd Business School, University of Oxford; Cracow University of Economics; TRADE Research Advisory (Pty) Ltd; the University of East Anglia; the University of Northumbria; Leeds Beckett University; and Qatar Computing Research Institute that have made significant contribution to our work and have informed the ideas and core discussions, which are presented in this edited volume. Mohammad Dastbaz Peter Cochrane

v

Contents

1 Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead ����������������������������������������������������������������������    1 Mohammad Dastbaz 2 Why Industry 4.0? ����������������������������������������������������������������������������������   13 Peter Cochrane 3 Connectivity for Industry 4.0������������������������������������������������������������������   23 Kristina Gold, Kenneth Wallstedt, Jari Vikberg, and Joachim Sachs 4 Wireless Comms. Beyond 2020 ��������������������������������������������������������������   49 William Webb 5 Digital Transformation����������������������������������������������������������������������������   65 Paul Graham 6 Big Data, Small Data, and Getting Products Right First Time ����������   77 Human Ramezani and Andre Luckow 7 The Internet of Things and Sustainable Manufacturing����������������������   91 David Heatley and Mohamed Abdel-Maguid 8 Security Challenges in the Industry 4.0 Era������������������������������������������  117 Mohammed M. Alani and Mohamed Alloghani 9 Meeting the Future Challenges in Cyber Security��������������������������������  137 Ben Azvine and Andy Jones 10 Ready for Industry 4.0? The Case of Central and Eastern Europe ��������������������������������������������������������������������������������  153 Wim Naudé, Aleksander Surdej, and Martin Cameron 11 From Big to Small Data ��������������������������������������������������������������������������  177 Peter Cochrane and Ahmed Elmagarmid

vii

viii

Contents

12 The Role of Blockchain in Underpinning Mission Critical Infrastructure ������������������������������������������������������������������������������������������  191 Hamid Jahankhani and Stefan Kendzierskyj 13 The Planning and Design of Buildings: Urban Heat Islands—Mitigation������������������������������������������������������������  211 Christopher Gorse, James Parker, Felix Thomas, Martin Fletcher, Graham Ferrier, and Neill Ryan Index������������������������������������������������������������������������������������������������������������������  227

Contributors

Mohamed  Abdel-Maguid  is a professional engineer, entrepreneur and research and enterprise-focused academic, with 25 years of experience in both business and academia. He joined the University in 2012 as the founding professor of the School of Science, Technology and Engineering and professor of Smart Systems. Professor Abdel-Maguid’s research is focused on two strands. The first strand focuses on investigating new algorithms and techniques to underpin the development of multimodal signal processing and communication systems (unified, ubiquitous, cognitive, cooperative). The second strand is concerned with the application of digital technologies to the development of smart systems as the main building blocks of future cities and smart connected rural communities. Mohammed M. Alani  is an associate professor in Computer Engineering and currently is the provost at Al Khwarizmi International College, Abu Dhabi, UAE. He received his PhD in Computer Engineering in 2007 and since then has worked in various institutions in the Middle East. He has published five books in different areas of networking, security and big data, in addition to many papers in international peer-reviewed journals and conference. Mohamed Alloghani  is a PhD candidate and has been actively involved as a member of international conferences. He has 15 years of work experience in the computing domain and is currently working for the Abu Dhabi Health Services Company (SEHA) as an ICT corporate projects manager who manages all technical ICT projects, including but not limited to artificial intelligence, robotics, mobile computing, Web systems development and enterprise medical systems. He has published numerous refereed research papers and undergoing book chapters in multidisciplinary areas including e-governance, technology-enhanced learning, intelligent computing, artificial intelligence, big data, Internet of Things (IoT), mobile app development and mHealth application development. Ben Azvine  is responsible for setting direction and strategy for security research, identifying innovation opportunities and leading a strong international team of ix

x

Contributors

researcher to develop new capabilities in collaboration with industrial and academic partners. He has 30 years’ experience in both academia and industry. His previous roles included leading the IT research centre and head of business intelligence and customer analytics research at BT Group Chief Technology Office. He holds a PhD in Intelligent Control Systems from Manchester University and an Executive MBA from Imperial College, London. He is an inventor of 50 patent applications and holds visiting professorship positions at the Universities of Bristol, Cranfield, Suffolk and Bournemouth in the UK. Ben is a current fellow of the Institution of Engineering and Technology (IET) and the Institute of Telecommunications Professionals (ITP) and has acted as the chairman of the European Network of Excellence for Uncertainty Management Techniques. His current research includes visual analytics for cyber defence, cloud and Internet security, AI and machine learning for network anomaly detection and future identity and access management. Martin Cameron  is a quantitative economist specialising in quantitative executive decision support modelling, economic impact analysis and industrial and trade policy as well as engineering management decision support for large infrastructure projects. He is managing director of TRADE Research Advisory (Pty) Ltd, a spin-­ out company of the North-West University, South Africa. TRADE Advisory provides strategic and practical assistance to government and business on optimization of their regional or global presence, from the perspective of economic development and export expansion. Martin also works in the field of advanced (renewable) energy as independent economic advisor to the South African Independent Power Producer’s Programme. He is author of various papers on the topic of international trade, including recently for the International Growth Centre of the London School of Economics on policies for improving export promotion for Rwanda. He has also worked for the South African Indian Ocean Rim Association’s  Academic Group (IORA-AG) expanding intra-Indian Ocean Rim Association (IORA) trade. Graham Ferrier  is a reader in Earth Observation Science and head of Geology at the University of Hull with 20 years of experience in the spatial analysis and modelling of a wide range of environmental processes integrating a wide range of novel spatial datasets and analysis algorithms to provide new insights into the spatial and temporal dimensions. He is a member of the NERC Peer Review College and the NERC Science & Facilities Earth Observation Commissioning Panel and has been an external evaluator on a number of EU funding programmes. He is a fellow of the Geological Society and is currently principal investigator (PI) on NERC- and STFC-­ funded projects. He is highly multidisciplinary and is currently participating in a number of externally funded projects which are parameterising innovative modelling methods with a range of novel observation technologies to increase understanding of the influence of green infrastructure on local climate processes. Martin  Fletcher  is a research fellow with the Leeds Sustainability Institute at Leeds Beckett University. Martin’s academic background is in renewable energy, completing his masters at Newcastle University before working in the Centre for the

Contributors

xi

Green Knowledge Economy at Bournemouth University. Since joining Leeds Beckett in 2012, Martin has specialised in the performance evaluation of occupied buildings. Specifically, Martin’s research focuses on the energy use, thermal comfort and internal conditions in buildings. Martin is an established author, with publications in high-ranking journals. Kristina  Gold  has worked in the mobile industry for 25 years, with a focus on mobile communication ranging from development of smartphones to radio base station. She is currently working at Ericsson as director of Technology Foresight. Kristina is also engaged in academia as a board member within the Swedish National Infrastructure for Computing. Kristina holds an MSc in Engineering Physics from Uppsala University. Christopher Gorse,  BSc (Hons), MSc, PhD, MCIOB, MAPM, FHEA is a professor of Construction and Project Management, director of the Leeds Sustainability Institute and head of the Centre of the Built Environment at Leeds Beckett University. He is the current chair of the Association of Researchers in Construction Management (ARCOM) and chair of the International Conference for Sustainable Ecological Engineering Design for Society (SEEDS). Chris started out as an engineer and project manager, with a background in contracting and consultancy. He is a Chartered Building, Environmentalist and Engineering Professors’ council member, holding principal investigator positions for major construction, environment and energy research projects. He is also task leader for International Energy Agency on Energy and Environmental projects. The teams that Chris leads are multidisciplined and skilled with expertise in building, building forensics, climate, materials, behaviour, energy, data analytics and simulation. Chris is an established author, with leading text on sustainability, technology and management. Paul  Graham  joined Matrixx Software, a Silicon Valley company supplying a leading-edge digital commerce platform, in 2014 as directory of UK Engineering to establish an engineering centre to utilise local talent to expand Matrixx’s global footprint. He has 25 years’ experience in telecommunications, having been a member of another successful start-up, eServGlobal, part of which was acquired by Oracle in 2010. Paul’s areas of specialty originated in mobility, primarily concerned with call control and value-added services. He then progressed into real-time charging and the necessary support systems in OSS/BSS. His more recent area of expertise is digital commerce, looking at it with an end-to-end perspective, so as the charging and payment aspects as well as the mobile app self-care. He is a co-author of the book MMS: Technologies, Usage and Business Models. David Heatley  During his 30+-year career in research and development, David has held senior positions across a broad front of technology developments and application areas including wireless networks (radio and infrared), optical fibre systems, sensor networks, IoT applications, telemedicine and more. He is an acknowledged authority in these fields and is regularly called upon to deliver consultancy.

xii

Contributors

Publications of his work appear in many journals and conference digests, he is co-­editor of a book on telecom networks, and he has contributed chapters in several other books. He is the recipient of awards in recognition of his innovative work and is accredited with more than 20 filed patents. He is a chartered engineer and member of the Institution of Engineering and Technology and holds a PhD in Optical Fibre Systems and MSc and BSc degrees in Telecommunications and Electronic Engineering. He is also a visiting professor of e-Health Innovation. Hamid Jahankhani  gained his PhD from the Queen Mary College, University of London. In 1999, he moved to the University of East London (UEL) to become the first professor of Information Security and Cyber Criminology at the university in 2010. Over the last 15 years, Hamid has also been involved in developing new and innovative programmes and introducing “block mode” delivery approach at UEL, including MSc in Information Security and Computer Forensics and professional doctorate in Information Security. Hamid’s principal research area for a number of years has been in the field of cyber security, information security and digital forensics. In partnership with the key industrial sectors, he has examined and established several innovative research projects that are of direct relevance to the needs of UK and European information security, digital forensics industries, critical national infrastructure and law enforcement agencies. Professor Jahankhani is the editor-in-chief of the International Journal of Electronic Security and Digital Forensics (www.inderscience.com/ijesdf) published by Inderscience and general chair of the annual International Conference on Global Security, Safety and Sustainability (ICGS3). Hamid has edited and contributed to over 15 books and has over 150 conference and journal publications together with various BBC Radio interviews. Hamid has supervised to completion 13 PhD and professional doctorate degree students and overseen 67 PhD students progressing. In summer 2017, Hamid was trained as the Government Communications Headquarters (GCHQ) “cyberist” to train the next generation of cyber security experts through GCHQ CyberFirst initiative. Professor Jahankhani has a number of books, Cyber Criminology, Blockchain and Clinical Trial: Securing Patient’s Data, Cyber Security Practitioner’s Guide and Digital Twins Technologies and Smart Cities, to be published by Summer of 2019. September 2018. Andy Jones  has a defence and defence research background. He was a principal lecturer at the University of Glamorgan (now the University of South Wales) in the subjects of Network Security and Computer Crime and a researcher on the threats to information systems and computer forensics. He developed and managed a well-­ equipped Computer Forensics Laboratory and took the lead on a large number of computer investigations and data recovery tasks. After this, he joined the Security Research Centre at BT where he became the head of information security research. He managed a number of research projects and led a series of projects into residual data on second-hand media. He is currently the director of the Cyber Security Centre at the University of Hertfordshire. He holds a PhD in the area of Threats to

Contributors

xiii

Information Systems. He has written seven books on topics including information warfare, risk management and digital forensics and cybercrime and has also had more than 100 peer-reviewed papers on the same subjects published. Stefan Kendzierskyj  is an experienced commercial director with over 25 years’ experience gained in a number of technology companies covering industries such as telecommunications, software development, outsourcing, publishing, eLearning, and content transformation. Stefan has a strategic approach when presenting solutions at Board level and finding innovative ways for businesses to transform their models and reach new markets. This has been in the form of collaborating with some high-profile publishers and large associations that need to reach a global audience with their vision and messaging, particularly in the healthcare and commercial sector. Stefan also provides publisher services for the UK Disaster Victim Identification team and Interpol in the form of process tools to aid in investigation work for mass fatalities where the process of identifying victims of mass disaster may not be possible by visual recognition. This consultative approach over the years has led to improvements to the methodology, and constant attention applied to ensure an agile and adaptive approach gives those in the field efficient practices. Stefan holds an M.Sc. in Cyber Security and has attained a level of Distinction. He has also published a number of articles and books with large publishers concerning blockchain, security/privacy of data (particularly in the healthcare sector), and securing critical national infrastructure and in the process of other publishing work. Stefan is also a speaker at conferences or any thought leadership events. Andre Luckow  is researcher and manager in the automotive industry. His work lies at the intersection of emerging technologies, such as blockchain and artificial intelligence, and automotive and mobility applications. He holds a PhD in the field of Distributed Computing from the University of Potsdam, Germany. Wim  Naudé  is a professor in Business and Entrepreneurship at Maastricht University, professor in Development Economics and Entrepreneurship at the Maastricht School of Management (MSM), visiting professor in Technology, Innovation, Marketing and Entrepreneurship at RWTH Aachen University and academic visitor at Saïd Business School, University of Oxford. He is also a research fellow at the IZA Institute of Labor Economics, the Global Labor Organization and the African Studies Centre, University of Leiden. Wim is one of the world’s foremost scholars on entrepreneurship economics and development. He has been a policy advisor and contributor to virtually all global development organisations including the WTO, OECD, UNIDO, UNCTAD, UNU, EU, World Bank, Global Development Network and more. His is a regular keynote speaker and corporate lecturer and has lectured in the Brown International Advanced Research Institutes of Brown University in the USA. His most cited works include a chapter on entrepreneurship and development that is included in the 2014 Oxford University Press book International Development: Ideas, Experience, and Prospects, which contains

xiv

Contributors

a foreword by Nobel Laureate Amartya Sen and the 2011 paper “Entrepreneurship and Human Development: A Capability Approach” in a special edition of the Journal of Public Economics edited by Thomas Piketty. James  Parker  is a reader in the Leeds Sustainability Institute at Leeds Beckett University. James specialises in building performance simulation modelling with a particular interests in retrofit, natural ventilation and summer overheating. He has been involved in a wide range of externally funded research projects, including a lead role in an international project designed to evaluate and visualise the impact of green infrastructure in urban heat island areas. James also runs an MSc module on sustainable refurbishment as well as supervises numerous PhD and BSc students. Human  Ramezani  studied Computer Science at the Technical University of Kaiserslautern, Germany, and joined the BMW Group in 1998. He introduced virtual and augmented reality methods effectively into company processes before joining the IT department in 2004. There, he successfully defined and implemented an IT innovation management and scouting process to show the value proposition of innovative IT technologies for many fields of application. Human was involved in several research projects and active member of innovation networks. He is BMW Netherlands head of IT since 2015. Human is interested in technology in general and enjoys movies and photography. Currently, he lives with his wife in Den Haag, Netherlands. Neill  Ryan  As founder and CEO of VRM Tech, Neill is a technology industry veteran with over 20 years’ experience working at a C level. Prior to VRM, Neill was a co-founder of PSI Mobile, a SaaS provider to energy and telco companies for mobilising sales contracts. He helped establish and run Oneview, a healthcare technology company that is now listed on the Australian stock market. VRM specialises in creating smart data for the built environment to assure the quality of works and improve the performance of energy systems. Joachim Sachs  is a principal researcher at Ericsson Research. He joined Ericsson in 1997 and is currently coordinating research activities on wireless connectivity for the Industrial Internet of Things. Over the years, he has worked with radio network design and analysis for various communication standards. After studies in Electrical Engineering in Germany, France, Norway and Scotland, he received diploma and doctorate degrees from RWTH Aachen University and Technical University of Berlin, respectively. In 2009, he spent the year as visiting scholar at Stanford University. Aleksander  Surdej  is a professor of International Economics at the Cracow University of Economics. He holds PhD degree from Jagiellonian University and completed an International Studies programme at SAIS at the Johns Hopkins University and a PhD programme at the European University Institute in Florence. He conducted research at the Robert Schuman Centre, the UNU/WIDER in Helsinki

Contributors

xv

and Netherlands Institute for Advanced Study (NIAS) in the Netherlands. Aleksander Surdej is an expert in international economics, international regulatory policies, development economics and global public policies. He is an author and editor of 14 books and more than 70 articles related to international economics published in Polish and international scientific journals. Recently, Aleksander has been working on the design of the mechanisms of co-operation in the international economy stressing the role of international organisations in the provision of global public goods and in regulating business environment. He has furthermore coordinated and participated in more than a dozen international research projects related to informal work, higher education, vocational training, family firms and small businesses. He lectured in the USA, Italy and Germany. From 2016, Aleksander Surdej is a permanent representative of Poland to the Organisation for Economic Co-operation and Development (OECD). Felix Thomas  is a research assistant in the Leeds Sustainability Institute, and his work is focused on the thermal performance of dwellings undergoing thermal upgrade and major retrofit. He specialises in elemental thermal modelling and in situ performance testing of dwellings, particularly issues around thermal bridging and moisture. He has collaborated on a wide range of projects within the group and has led the production of a number of publications. He is currently working on his PhD, which is investigating thermal bridging caused by gaps in retrofitted internal wall insulation and the resulting risks of condensation and mould formation. Jari  Vikberg  is a senior expert in network architecture and the chief network architect at CTO office. He joined Ericsson in 1993 and has both wide and deep technology competence covering network architectures for all generations of RANs and CNs. He is also skilled in the application layer and other domains and the impact and relation these have to mobile networks. He holds an MSc in Computer Science from the University of Helsinki, Finland. Kenneth  Wallstedt  is director, Technology Strategies, in Ericsson’s CTO office where he focuses on the company’s radio and spectrum management strategy. He joined Ericsson in 1990 and has since then worked in various leading positions in Ericsson’s research, development and market units in Canada, Sweden and the USA.  He holds an MSc in Electrical Engineering from KTH Royal Institute of Technology in Stockholm, Sweden. William  Webb  is an independent consultant at Webb Search and CEO of the Weightless SIG, a body standardising a new M2M technology. He was one of the founding directors of Neul, a company developing machine-to-machine (M2M) technologies and networks, which was formed at the start of 2011 and subsequently sold to Huawei. Prior to this, William was a director at Ofcom where he managed a team providing technical advice and performing research. He has worked for a range of communications consultancies and spent 3 years providing strategic

xvi

Contributors

management across Motorola’s entire communications portfolio, based in Chicago. He was IET president from 2014 to 2015. William has published 16 books, 100 papers and 18 patents. He is a visiting professor at Southampton University and a fellow of the Royal Academy of Engineering, the Institute of Electrical and Electronics Engineers (IEEE) and the IET. He has been awarded multiple honorary doctorates by the UK’s leading universities.

About the Editors

Mohammad  Dastbaz  has published over 60 refereed journal and conference papers, books and book chapters. He is on a number of editorial boards of international journals, has been chair of a number of international conferences (including IEEE’s Information Visualisation) and remains a member on a number of international conference committees. His latest publications include an edited volume published by Elsevier publishers, on Green Information Technologies (2015); Building Sustainable Futures: Design and the Built Environment, published by Springer (2016), a collaboration with the University of California, Berkeley; and a series of four edited volumes on Technology and Sustainable Futures being published by Springer. He is also the editor of the first three books in the series Building Information Modelling, Building Performance, Design and Smart Construction, a collaboration with the University of Cambridge (2017); Technology for Smart Futures (2017), a collaboration with the University of Georgia and Sheffield Hallam University; and Smart Futures, Challenges of Urbanisation, and Social Sustainability (2018) a collaboration with the Maastricht School of Management and Leicester School of Architecture. Professor Mohammad Dastbaz is a fellow of the British Computer Society (BCS, The Chartered Institute for IT), the Higher Education Academy and the Royal Society of Arts (RSA). Peter Cochrane, OBE  As a seasoned professional with decades of hands-on management, technology and operational experience, Peter was head of research and chief technology officer (CTO) at BT (2000) with 1000 strong teams spanning optical fibre, fixed and mobile networks, complex systems, AI (artificial intelligence), AL (artificial life), futures, human behaviour and interfaces. Since leaving BT, he has been employed by the defence, logistics, travel, retail, energy, healthcare, transport and pharma sectors and as an advisor to governments and numerous companies including Rolls-Royce and Facebook.

xvii

xviii

About the Editors

As an entrepreneur, he engaged in start-ups and global investments. The most visible are Ebookers and Shazam Entertainment. Peter was the UK’s first Professor for the Public Understanding of Science and Technology at Bristol in 1998, received the Queen’s Award for Innovation and Export in 1990, has numerous honorary doctorates and is an OBE (1999). He recently became a professor of Sentient Systems at the University of Suffolk (UoS), Ipswich.

Chapter 1

Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead Mohammad Dastbaz

Introduction In the aftermath of the 2008 global financial crises, a debate around how to recover from the crises and ensure future growth, as well as the role of technology in the future of this growth, is pursued. In a report titled “Industrie 4.0” (Industry 4.0), the German government through its “Ministry of Education and Research” and the “Ministry for Economic Affairs and Energy” proposed a national strategic initiative focused on building a digital society and pushing digital manufacturing into an ever-­expanding interconnection of products, value chains and business models (European Commission Report 2017). While the report generated a lot of interest and it was followed by the German government initiating ten “Future Projects”, the reality is that 7  years later and despite numerous debate and position papers both from industry and academia, the full concept and potential of Industry 4.0 remain largely poorly understood and not widely implemented or exploited. In a report published by Deloitte on “Industry 4: Are you ready”, in January 2018, Punit Renjen, its global CEO, writes: “… Wristwatches monitoring vital signs to warn of impending heart attacks. Factories running at optimal capacity, with every process monitored and adjusted in real time. With the emergence of big data, cloud computing, the Internet of Things, 3D printing, and more, this is the world being ushered in by the fourth industrial revolution (Industry 4.0)” (Deloitte Review 2018). The report highlights the fact that only 14% of CXOs are confident that their organisations are ready to fully harness and benefit from what i4.0 has to offer. Before attempting to provide a framework of what i4.0 in its current stage of development is, what are the key areas of technology that define i4.0 and how we can move beyond the jargons and the hype, it will be useful to provide a historical context to i.4.0 and its predecessors and how they impacted our development over the past 250 years. M. Dastbaz (*) University of Suffolk, Ipswich, Suffolk, UK e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_1

1

2

M. Dastbaz

Historical Context The extant literature indicates that the Industrial Revolution started in the 1760s in Britain. With the emergence of steam engine and steam power at the dawn of the first Industrial Revolution (and from now on i.1.0), a significant shift in the mode of production from cottage industry production to large-scale mechanisation took place. The dawn of the Industrial Revolution and the shift from framing/cottage-­ based production to large-scale factories also signalled the beginning of an era of significant scientific discoveries and innovation (Dastbaz et al. 2016). From “the spinning jenny” that increased wool mills productivity in 1764 to James Watt’s first reliable steam engine in 1775, the “telegraph communications” in the early 1800, and finally Joseph Aspdin who, in 1824, devised and patented a chemical process for making “Portland cement”, the world rapidly changed, in both producing manufactured goods to how we rapidly developed population centres (town and cities). According to Eric McLamb (2011), following the Industrial Revolution in the late 1700s, the world’s population grew by 57% to 700 million and then quickly reached the billion mark by 1800, and within the first 100 years of the Industrial Revolution, it grew by a further 600 million to reach 1.6 billion (MacLam 2011) (Fig. 1.1).

Fig. 1.1  Time line for Industrial Revolution 4.0

1  Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead

3

 ey Technical Advances: High-Frequency Transistors, IC, K and ARPANET Looking at the how our current technological advances have been affected, one could identify a number of key technical advances that all happened within a span of two decades following the end of the World War II. According to historic documents, the transistor was invented by three American physicists, John Bardeen, Walter H.  Brattain and William B.  Shockley, at the American Telephone and Telegraph Company’s Bell Laboratories in 1947–1948 (Development of Transistors n.d). The transistor’s high reliability and low consumption as compared to the electron tube, as well as its capacity to compress complex circuitry into a small device, made significant changes to the development of modern electronics (Fig. 1.2). Following the invention of the transistor, it was the emergence of integrated circuit (IC), first successfully tested in September 1958, by Jack Kilby, that paved the way for the emergence of the modern computing. In his patent application on 6 February 1959, Kilby described his new device as “a body of semiconductor material … wherein all the components of the electronic circuit are completely integrated”(Winston 1998).

Fig. 1.2  An old Motorola television from 1948 with electron tubes. (Image sources: https://www. antiqueradio.org/motvt73.htm)

4

M. Dastbaz

The capability provided by the IC to produce microchips with large number of transistors, diodes and resistors made possible the production of both powerful mainframe (large-scale) computers and microcomputers with higher operating speeds. At the same period that rapid changes were taking place in the hardware industry and companies like Intel (1968) and Apple (1976) were developing new products not aimed at large military or scientific labs but for the mass market, there was significant developments made in creating systems solutions (software) to work with the emerging new power of these computers. The emergence of companies such as Microsoft (1975) and the development of new “operating systems” [such as DOS (disk operating system)] which provided a more manageable, user-friendly interface with the new microcomputers can be viewed as a revolution and the dawn of the digital age. Gordon Moore, the co-founder of Fairchild Semiconductor and CEO of Intel Corporation, noted in 1965 that there was the strong possibility of doubling the number of components per integrated circuit every 2 years (this is now referred to as “Moore’s law” although this was only an observation). In his paper “Cramming More Components onto Integrated Circuits”, Moore wrote: “The future of integrated electronics is the future of electronics itself. The advantages of integration will bring about a proliferation of electronics, pushing this science into many new areas”(Moore 1965) (Fig. 1.3).

Fig. 1.3  Increasing number of transistors per microprocessor over the last four and a half decades. (Source: https://ourworldindata.org/technological-progress)

1  Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead

5

An important part of the current information revolution relates to the concept of creating “networks” of machines that exchange data across long distances. Perhaps the most notable of network projects that prepared the ground for the emergence of the Internet was Advanced Research Projects Agency Network (ARPANET). The first network structure that used an innovative packet switching data exchange protocol was called “Transmission Control Protocol (TCP)”. On 29 October, 1969, the first message was sent over the ARPANET link from Leonard Kleinrock in UCLA to a second node at Stanford Research Institute in Menlo Park, California. The message was simply “Lo” instead of the intended word “login” (“Charley Kline Sends the First Message Over the ARPANET” n.d.) It was this combination of advances in hardware and software design as well as the emergence of network technologies that provided the building blocks of the new digital universe with its billions of connected users and yottabytes (YB) of data being exchanged (1 YB  =  1024bytes  =  1 000 000 000 000 000 000 000 000 bytes = 1000 zettabytes (ZB) = 1 trillion terabytes (TB)).

Industry 4.0: Hype or Reality? The rapid changes in the power of computing and the wide reach of the technology, which almost covers the entire planet and every aspect of our lives, fundamentally changed the way the industries worked, businesses operated and how we live and communicate. It’s also worth noting that while there has been some justified criticism about hyping the concept of i4.0 to sell products and solutions that are hardly anything significantly different to what has been around for a while, there are significant new products and solutions following the technological advances that we have seen over the last decade that will bring significant changes. A report by KPMG titled “Beyond the hype: Separating ambition from reality in Industry 4.0” warns against unrealistic expectations and warns: “Everyone wants to talk about i4.0. From industry conferences and magazines through to boardroom tables and shareholder meetings, i4.0 is at the top of the agenda. The pressure on executives to adapt and compete is tremendous. But there is also a lot of hype. Projections for the i4.0 market run into the trillions. Forecasts for potential value creation are eyewatering. Revenue expectations at manufacturers and at service providers — are flying high. Depending on who you talk to, the disruption for value chains, employees and business models may be fundamental…” (KPMG Report 2017). In response to the sceptics, it is fair to point out that the modern computing technology, as already stated, has been around for over six decades, but the pace of change and the hardware and software power to enable this change are not comparable to anything we have seen before. Therefore, it is safe to assume that we have entered a new era of an Industrial Revolution that is as significant as the first one and will no doubt change the future of humanity for good. While it is important to note

6

M. Dastbaz

that we are at the beginning of this road, and there are still significant challenges ahead, we also acknowledge the opportunity and look to see how i4.0 will impact our future.

Understanding Industry 4.0 The extant literature provides a range of definitions and identifies several key components and challenges where i4.0 is concerned. For the purpose of this book, the key features and technological enablers for Industry 4.0 are summarised as below: 1. Internet of Things (IoT) – The backbone of the ever more connected world where various devices be it personal or industrial are connected through the Internet, thus creating a new digitally connected world with billions of users and providing new possibilities for knowledge, data and process sharing and exchange. 2. Cloud Computing – According to Microsoft, “cloud computing is the delivery of computing services – servers, storage, databases, networking, software, analytics, intelligence and more – over the Internet (“the cloud”) to offer faster innovation, flexible resources and economies of scale” (Microsoft n.d.). Everything from our “Gmail” to our Internet search and financial exchanges online uses cloud-based services. According to Statistica: 2018, approximately 3.6 billion Internet users are projected to access cloud computing services, up from 2.4 billion users in 2013. 3. Big Data – Over the past decade, there has been a significant amount of data produced, stored and shared across the Internet. The concept of “big data” deals with volumes of data storage and traffic unparalleled before the age of Industry 4.0. While it is difficult to provide an accurate estimate to the amount of data stored and trafficked across our digital universe (as this is a constantly changing volume), one way to provide an estimate for data currently stored on the Internet or on various cloud services is to look at data held by all the big online storage and service companies such as Google, Amazon, Microsoft and Facebook. Estimates are that the big four hold at least 1,200 petabytes between them, that is, 1.2 million terabytes of data. 4. Digital Manufacturing/Production (Factory 2.0) – From the emergence of 3D printers to laser machinery, and robots replacing humans in factories, it is not difficult to see how manufacturing has changed in the twenty-first century. There are several interesting case studies where it can be demonstrated how technological advances have changed the design and production process of goods. The car industry is one such case study where the old production lines, with hundreds of thousands industrial workers, have been replaced with robots (Fig. 1.4). 5. Industry 4.0 Logistics – One of the key discussion areas around Industry 4.0, besides its usage in manufacturing and production, has been around logistics  and how traditional warehouse operations have been revolutionised.

1  Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead

7

Fig. 1.4 BMW plant in Spartanburg, South Carolina. (Source: https://www.bmwblog. com/2015/10/30/behind-the-scenes-in-a-rare-factory-tour-of-the-bmw-spartanburg/)

Logistics  4.0 and Supply Chain Management 4.0 deals with the use of a ­combination of i4.0 technological enablers to provide for various aspects of end-to-end logistics and supply chain management. A good example of i4.0 in logistics and supply chain management is the ever-growing operations of Amazon, one of the world’s largest online retailers that has replaced its traditional warehouse and managing delivery operations with advanced computer systems and robots (Fig. 1.5). 6. Cyber-physical Systems (CPS)  – It is the concept of integrating physical processes with computation and networking power, where intelligent software process, monitor and control physical processes. An interesting example of CPS can be found in research around “Smart Living” where the movement of elderly people living alone can be monitored for potential problems and then provide support by raising an alarm automatically. 7 . Re-emergence of Artificial Intelligence (AI) and Autonomous Systems – One of the most striking technological developments over the past decades has been around AI and the potential new areas of application and development. While AI systems have been around for decades and as a young graduate doing ­information systems engineering in the 1980s and programming with languages such as PROLOG (programming in logic) and LISP (list processor), the reality was that mainframe systems such as DEC10 could not provide enough processing power to develop the complex systems that we are able to develop today. Using very powerful parallel processing systems, we are now able to develop ever more complex systems, tackling difficult problems in medicine, engineering, etc. and

8

M. Dastbaz

Fig. 1.5 Amazon bots in action in their warehouse. (Source: https://qz.com/709541/ amazon-is-just-beginning-to-use-robots-in-its-warehouses-and-theyre-already-making-a-hugedifference/)

Fig. 1.6  Erica Aoi the first android employee on Nippon TV. (Source: https://www.broadcastingcable.com/post-type-the-wire/nippontvericaaoi)

finding new solutions. The use of AI as a research and development tool has opened a whole new era of innovation, and with it several difficult philosophical questions and challenges were solved. It is worth noting that in April 2018, it was announced that “Nippon Television Network Corporation” (Nippon TV) will be welcoming its new “anchor Erica Aoi” as its first android employee (Fig. 1.6). 8 . Augmented Reality – The extant literature points to the ever-growing use of augmented reality in smart manufacturing, medical research and engineering. Volker Paelke (2014) presented “an augmented reality system” that supported human workers in a rapidly changing production environment. The system provided

1  Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead

9

spatially registered information on the task directly in the user’s field of view that could guide users through unfamiliar tasks (e.g. assembly of new products) and visualise information directly in the spatial context where it is relevant (Paelje 2014).

Some of the Key Challenges While, as discussed, Industry 4.0 brings significant potentials for innovation and opens new possibilities to revolutionise business and industry processes as well as our lives, it also comes with a number of key challenges and issues that require careful consideration and addressing. These can be summarised as follows: 1. Cybersecurity – As recent news headlines and horror stories of large-scale hacking and stealing of personal data suggests our new digital universe, our connected world where massive amount of data is shared across networks is a very vulnerable environment. Cybersecurity and threats posed by online data exposure is a critical issue that requires urgent new solutions. A report published by USA Today on 28 December 2018 noted that: “Billions of people were affected by data breaches and cyberattacks in 2018 – 765 million in the months of April, May and June alone – with losses surpassing tens of millions of dollars, according to global digital security firm Positive Technologies. Cyberattacks increased 32% in the first 3 months of the year and 47% during the April-June period, compared to the same periods in 2017”(USA Today 2018). Clearly there is an urgent need to develop a more secure network infrastructure providing secure communications and data exchange between companies and users. Without a robust cybersecurity protocol, the full industrial implementation of i4.0 will be quite challenging. The emergence of “Blockchain” (discussed later in this volume), which uses the bitcoin technology to create an open distributed ledger that can record transactions between two parties efficiently and in a verifiable way, could be an important development in addressing some of the cybersecurity challenges facing Industry 4.0. 2. I4.0 Energy Cost and Requirement – The new digital era brings with it its significant energy costs and requirements. With more than 3.5 billion smart phones and 3 billion PCs, laptops and tablets that require constant energy to operate, and that is before measuring the cost of billions of websites, tens of millions of servers and yottabyte of data that is exchanged in this universe 24/7 365 days a year, it is not difficult to ascertain that we will sooner than later will be facing serious energy shortages to keep up with the growth and increasing demand of our new universe. While it is almost imposable to measure the energy cost and requirements of our current digital universe, a report by the “Department of Energy’s Lawrence Berkeley National Laboratory” calculated that to run the Internet we require some 70 billion kilowatt hours per year. This amounts to 1.8% of the total American electricity consumption. At an average cost of 10 cents per kwh, the annual cost of runs into $7 billion (Helman 2016). With 4.1 billion Internet users,

10

M. Dastbaz

1.9 billion websites and 1.5 billion e-mails sent in a day (Data Source n.d.) (data relates to 30 December 2018), it is not hard to imagine that we are rapidly moving towards a situation where the energy requirement of our new digital universe will soon outstrip the electricity we are producing. Research on developing alternative sources of energy, more efficient intelligent network infrastructures and better battery technologies are now a critical area of development if we are to stay on course to fulfil the potential of i4.0. 3 . The Social Cost – Literature and research relating to i4.0 have already noted the significant potential challenge to the fabric of our societies including jobs, the nature of employment, people’s rights and the relationship between humans and the “humanoids”. Perhaps the more worrying aspect is that we are already seeing that more and more businesses see robots as means of maximising output, efficiency and profit and potentially displacing human workers altogether. A report by “McKinsey Global Institute” titled “Jobs Lost, Jobs Gained: Workforce Transitions In A Time Of Automation” published in December 2017 predicts that it is possible that by 2030, as many as 800 million jobs could be lost worldwide to automation. The report states that: “Our scenarios across 46 countries suggest that between almost zero and one third of work activities could be displaced by 2030, with a midpoint of 15%. The proportion varies widely across countries, with advanced economies more affected by automation than developing ones… Our scenarios suggest that by 2030, 75 million to 375 million workers (3–14% of the global workforce) will need to switch occupational categories. Moreover, all workers will need to adapt, as their occupations evolve alongside increasingly capable machines. Some of that adaptation will require higher educational attainment…”. Industry 4.0 is clearly not a fad and is here to stay with us for the unforeseeable future changing the way we work and we live. Industry 4.0 is no longer hype or merely a wild claim to sell old products under a new badge. As the “Industrie 4.0” German government report proposed, we are now living in a “digital society” with its benefits and pitfalls. Access to a universe of data and information, the availability of services in a global market place which are no longer bound by local shortcomings and our ability to use AI to solve very complex medical, engineering and environmental problems clearly indicate that we live in times of significant change. While we will be wise to acknowledge that this is only the beginning of our Industry 4.0 journey and we still have some key questions and challenges to address, and if one looks at the progress made over the last 25 years and then projects this forward to the next 25 years, it is not difficult to predict that we will all be living, working and communicating in a totally different world. Perhaps we could call this the “Digital Universe 2.0”, but what is important to be mindful of is that this transformation could come with significant cost and challenges that we need to carefully plan for and manage. This edited volume sets out to critically discuss and evaluate some of the key issues and challenges facing i4.0 and its future.

1  Industry 4.0 (i4.0): The Hype, the Reality, and the Challenges Ahead

11

References “Charley Kline Sends the First Message Over the ARPANET”, from Jeremy Norman’s “HistoryofInformation.com”; http://www.historyofinformation.com/detail.php?entryid=1108 Dastbaz, M., et  al. (2016). Building sustainable futures. Springer International Publisher Switzerland. Data source is from “Internet Live stats”. http://www.internetlivestats.com/ Deloitte Review “Industry 4: Are you ready”; Issue 22, January 2018. https://documents.deloitte. com/insights/DeloitteReview22 “Development of Transistors”, Encyclopaedia Brirtanica: https://www.britannica.com/technology/ transistor European Commission Report: “Digital Transformation Monitor Germany: Industrie 4.0”; Jan 2017. https://ec.europa.eu/growth/tools-databases/dem/monitor/sites/default/files/ DTM_Industrie%204.0.pdf Helman, C. (2016). “Berkeley Lab: It takes 70 billion kilowatt hours a year to run the Internet”, https://www.forbes.com/sites/christopherhelman/2016/06/28/how-much-electricity-does-ittake-to-run-the-internet/#46d85c0c1fff https://www.mckinsey.com/featured-insights/future-of-work/jobs-lost-jobs-gained-what-the-future-ofwork-will-mean-for-jobs-skills-and-wages https://www.sciencefocus.com/future-technology/how-much-data-is-on-the-internet/ KPMG Report: “Beyond the hype: Separating ambition from reality in i4.0”, 2017; source: https:// home.kpmg.com/xx/en/home/insights/2017/05/beyond-the-hype-separating-ambition-fromreality.html MacLam, E. (2011). The Ecological Impact of the Industrial Revolution; http://www.ecology. com/2011/09/18/ecological-impact-industrial-revolution/ Microsoft: “What is cloud computing?”; source: https://azure.microsoft.com/en-gb/overview/ what-is-cloud-computing/ Moore, G. E. (1965). “Cramming more components onto integrated circuits”; Electronics, pp. 114– 117. Publisher Item Identifier S 0018-9219(98)00753 http://www.cs.utexas.edu/~fussell/ courses/cs352h/papers/moore.pdf Paelje, V. (2014). Augmented reality in the smart factory: Supporting workers in an industry 4.0. environment. Proceedings of the 2014 IEEE Emerging Technology and Factory Automation (ETFA); https://ieeexplore.ieee.org/xpl/mostRecentIssue.jsp?punumber=6994138 USA Today: “Your data was probably stolen in cyberattack in 2018  – and you should care”, 28th December 2018. https://eu.usatoday.com/story/money/2018/12/28/data-breaches-2018billions-hit-growing-number-cyberattacks/2413411002/ US Patent 3,138,743, filed 6 February 1959, issued 23 June 1964. Winston, Brian (1998). Media Technology and Society: A History: From the Telegraph to the Internet. Routledge. p.  221. ISBN 978-0415142304.

Chapter 2

Why Industry 4.0? Peter Cochrane

Preamble The planet on which we all live and work presents a closed, finite, complex and sensitive environment that is being stressed by the presence of 7.5Bn humans and their activities. Since 1972 (Club of Rome Reports), we have been fully cognisant that the mineral and biological resources of the planet were being denuded and stressed along with the creation of a rising tide of pollution. Today we are suffering the consequences of our inaction in the form of global warming, water and food shortages, sea level rises and more frequent and extreme weather events, along with pollutants such as microplastics entering the food chain. No life-form can survive in its own waste material

What we are doing is not sustainable, and this problem is much bigger than changing all the light bulbs! Continuing to polish and refine our current industries and logistic chains to make them ever more efficient is not enough. It only puts off the day of reckoning and the ultimate collapse of societies. We really do have to rethink our use of energy, materials manufacturing (https://internetofthingsagenda.techtarget.com/ feature/4D-printing-is-the-catchphrase-programmable-materials-the-newsmakers), including fabrication and supply chains plus the whole arena of reuse, repurpose and recycling (3R). We have no choice! (Fig. 2.1). It is also very evident that our efforts at 3R to date have been lamentably poor with schemes seldom giving positive returns and often costing more energy, time and materials than they save (Recycling Failure). For example, most recycled plastics finish up in landfill due to contamination and their subsequent rejection by industry, whilst the hard plastic caps resist recovery and often finish up as hardcore for roads and some concretes. Probably the biggest failing has been political P. Cochrane (*) Cochrane Associates Limited, University of Suffolk, Ipswich, Suffolk, UK e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_2

13

14

P. Cochrane

Fig. 2.1  The impact of human activity and carbon pollution

decision making driven by perception, instinct and the heart. What we actually need is designed and engineered 3R systems based on hard science, facts and good engineering (Design for 3R). Perhaps an even bigger problem-set suffering from misguided policies and political instinct is energy. We have a multitude of ways we can create clean energy, and that is not the big problem, energy storage is. The wind does not blow every day, and the sun is not always shining. Moreover, solar panels and wind turbines do not come for free – they cost energy and pollution too, as do electric vehicles (Cost of Green). All forms of network including power suffer from ‘peaks and trough’ in demand and supply. For energy it can be extreme, and so we still see a need for nuclear, coal, oil and gas power stations that are not going away soon. What is needed is greater and far more efficient electricity storage facilities at a national, regional, local, campus and individual building level. This is an urgent need that has yet to be fully appreciated politically and one that is currently lacking necessary resources and R&D funding (Energy Storage). Exacerbating all of this are the capitalist and simple-minded economic systems that are patently unfit for purpose. It is nonsense to suppose that we can manage the complexities of a planet on the basis of a single parameter – money! At the very least, all our decisions should be subject to two further considerations: the impact on society and the impact on ecologies. However,  there is most likely a fourth dimension in the form of a global aggregation of the impact of the totality of our decisions and actions on raw material stocks and the likely impact of any associated pollution. In short, a far more comprehensive economic model is required that takes into account long-term outcomes (Beyond Capitalism). These then are the major industrial, social and economic problems facing our species, which for the most part are not intractable, and we can already see, or have solutions. They also demand that we change almost everything (mind-sets, economics, politics, companies, industries and the way they operate, create, deliver and support products, plus their end of life recovery and disposal via 3R chains that work at very high efficiency)! This is unlikely to be an easy transition from where we are to where we need to be!

2  Why Industry 4.0?

15

We have to extract materials, transform them into components and products , delivered and supported them for life, and then recover all materials at efficiencies of >> 90% at the lowest possible energy cost

Axioms • We cannot succeed in our mission by being entirely reliant upon the raw materials provided by Mother Nature. • We need new (artificial) materials with new and extraordinary properties and capabilities that are currently unavailable. • Our fabrication and manipulation methods have to be far less wasteful and inefficient. • Logistics chains are a dominant cost and a move towards production at point of need is a necessity. • Reuse, repurpose and recycle (3R) with the very high-efficiency recovery of plastics, metals and ceramics at low energy cost is a necessity. • Energy efficiency has to be improved drastically. • Burning oils and the destruction of limited valuable resources have to be minimised.

Industrial Revolutions We have witnessed three concurrent industrial revolutions during the past (near) 300 years that projected societies into the unique positions they now occupy. From the handmade products of the cottage industries to water-powered factories of I1.0 (circa 1750), to a steam- and oil-powered I2.0 (circa 1900) and mass production, and on the digital revolution (circa 1980), we are now at the starting grid of a materials-, AI- and robotics-driven I4.0 (circa 2015) destined to eclipse all that has gone before (Industrial Revolutions). It is worthy of note that the progression time from one industrial mode to the next has occurred at an increasing rate with a shortening of intervals of 150, to 80, to 35 years, with each realising progressively greater exponential innovation, production and population change. In the golden era of the digital revolution, we experienced an ability to produce IT that is 1000-fold more capable at 1000th of the price every decade since 1960 (Moore’s Law). Coined as ‘Moore’s law’ in the contest of transistor density per unit area, variants also apply to data storage, processing, communication, networking, media, data traffic and AI (Fig. 2.2). So successful have we been in the exploitation of the I3.0 technologies, we now know and understand, and can do, more than ever before in our history. “The discoveries, understandings and knowledge seen in the past 20 years by far eclipse (exponentially) those of the previous 2000”

16

P. Cochrane

In  30% weight reduction (Nano Steels). Nono Fabrics: Textiles modified by the inclusion of nanoparticles, structures, and/ or fibres to realise stronger, longer wearing, greater elasticity, self-cleaning, hydrophobic materials. These techniques and materials also extend into tissue and bone engineering as well as some plastics (Nano Fabrics). Coatings: High resilience, long life, low friction and hydrophobic properties’ look set to transform the shipping, vehicle, construction, soft furnishing and clothing sectors with potentially self-repairing and cleaning properties (Hydrophobic Coatings). “We have to stop supplying more and more to the few, and start providing sufficient for the many”

18

P. Cochrane

The Fabrication Space So far our history has been dominated by the top-down manipulation of bulk materials to create the tools and products that we need. The hammer, chisel, file, forge, drill, lathe, miller and shaper have been the prime instruments of industry along with forging, casting and extrusion. However, the relatively recent arrival of additive processes such as 3D and 4D printing is invoking major changes from ‘just-in-­ time’ production to teleported design/instruction sets feeding local printers. Three primary examples here are the following: 1. The aerospace industry has exceeded 50% of composite and plastic parts in aircraft construction. Huge robots are employed to lay down carbon fibres in contiguous high precision layers to achieve greater strength and flexibility than any metal. Notably, the R&D and design shops are often distanced by continents (Aerospace Composites). 2. The US military has deployed 3D printers on the front line for the production of replacement parts for aircraft, vehicles and weapon systems. This is often augmented by remote augmented reality back to the designers and producers for complex problems, updates and upgrades (Military 3D Printers). 3. Automobile producers are also employing more and more components produced by additive processes. Everything from manifolds to interior trim are being supplied ‘good to go’, whilst the tyre manufacturers are investigating printed wheels with solid tyres, flexible spokes, integrated motors and even the suspension elements (Auto-Industry 3D Printing). Looking to the future, a deal of energy and focus is being devoted to programmable materials that can give shape or form on command (https://selfassemblylab. mit.edu/programmable-materials). This parallels Mother Nature with the genome instructing proteins to become a particular organ in some animal. Such an ambition may seem fanciful or even science fiction, but significant progress is being made. The big question is will this be ready in time for I4.0 or will it be the driving force behind a future I5.0? In I4.0 we think of biology and the composites of metals, plastics and ceramics as separate and distinct entities, but in the future we will most likely see biological elements brought into this composite fold. But it will most likely happen in I5.0/6.0 timeframe and may even be the trigger for the next big industrial transformation.

The IoT Space In the literature there is scant recognition of that fact that I4.0 and any sustainable future is critically dependent on the ability to track every item we make, along with its complete production record, deployment and use in order to achieve our essential 3R objectives. But this happens to be the inherent role and responsibility

2  Why Industry 4.0?

19

afforded by the IoT. So by accident or by design, every element necessary for I4.0 to become the essential game changer in the move to a green future is already in place and well ahead in its development. To see a fuller perspective, we only have to look at the progressive development of production lines of the twentieth century where it soon became evident that there was a lot of value in knowing where materials and components came from and when, and who did what to them as the progressed along the assembly process. Traditionally, these were paper records called ‘travellers’ that eventually ran from start to finish: raw material-to-component, component-to-item and item-to-final product. Over time this included all test and measure at every stage as a part of an integrated quality assurance chain that generally ended at the post of shipping (Smart production Lines). Fast-Forward to this century, those paper travellers have become a few lines of code embedded in a chip that is easily accessible at any time through the product life. And recently this has been extended beyond the point of shipping to include point of sale, delivery, servicing and repair, and more controversially user behaviour. For example, top-end domestic appliances record details of when and how much coffee they dispense, how often and how much washing and drying are completed and so on (Fig. 2.3). Gathering details of customer behaviour is controversial, but it is something that the IT industry has been doing for decades (IoT in Cars, Office and Home) (Fig. 2.4). What is the unseen advantage to users in all this? For the first time, we, the users, become a part of the design loop! A washing machine with 83 selectable programmes might be shown to require only 8. A coffee machine that can count might offer some health advice to someone about to drink their 6th espresso of the m ­ orning. A car that has a fault or needs a service could go beyond flagging this fact to you and could direct you to the nearest dealer/garage and/or book an appointment with all the spare parts being delivered, or printed on site, just in time. But when we need a spare part, a prime option might be one recovered from a recently scrapped model that had only seen a small amount of use. This then is almost the complete picture, but there is one more important feature induced by adding more intelligence into everything we make, and in short: “Things that think want to link”

Fig. 2.4  When things are smart and can communicate, they will

20

P. Cochrane

Fig. 2.5  The cyber security of things has to be automated and autonomous

Our vehicles, domestic and office appliances and machines will want to connect, to share information and to protect each other from cyberattack as we roll out autoimmunity into the IoT (Fig. 2.5). In a sense, I4.0 products on the IoT will be alive, smart, adaptable and learning from us, the users. And to the end of their life, we will know how they have been used, how much, where they are and where they have been and their service record, along with the treasure trove of raw materials they contain. At this point we will be able to realise very efficiency 3R, and we will be far closer to achieving a sustainable future.

The Far Future Clustering people and resources in ever larger megacities has been perceived to be the most efficient way of organising societies, their provisions and supply. But that model does not look quite so certain on the back of the changes to be powered by and made viable by I4.0. For sure, we can’t all be farmers or even live in the country, but there is a new vision abroad that sees the potential for a migration to much smaller and distributed conurbations linked by high-speed transport and information networks with local energy, water supplies and some food production in vertical farms. And with the aid of I5.0 manufacturing that has embraced programmable materials, the differences between biology and technology will be even more blurred (https://www.fraunhofer.de/en/research/key-strategic-initiatives/programmablematerials.html). Only time will tell! “There is only one way to predict the future, and that is to build it”

2  Why Industry 4.0?

21

Further Reading Aerospace Composites: http://compositesmanufacturingmagazine.com/category/aerospace/ AI and Robotics: https://policyatmanchester.shorthandstories.com/on_ai_and_robotics/index.html Auto-Industry 3D Printing: www.stratasys.com/-/media/files/white-papers-new/wp_du_fivewaysauto.pdf Beyond Capitalism: https://www.theguardian.com/sustainable-business/2015/apr/21/regenerativeeconomy-holism-economy-climate-change-inequality. Club of Rome Reports: https://www.clubofrome.org/activities/reports/. Connected Everything: https://www.technologyreview.com/s/601013/the-internet-of-thingsroadmap-to-a-connected-world/ Cost of Green: https://insights.som.yale.edu/insights/can-we-afford-sustainability Design for 3R: https://www.sciencedirect.com/topics/materials-science/design-for-recycling Energy Storage: https://www.nytimes.com/2010/09/30/business/energy-environment/30iht-renstore.html Hydrophobic Coatings: https://onlinelibrary.wiley.com/doi/abs/10.1002/adfm.200801916 Industrial Revolutions: https://www.britannica.com/event/Industrial-Revolution IoT in Cars, Office and Home: http://www.itpro.co.uk/strategy/28543/how-the-internet-of-thingscan-transform-customer-experiences IoT in ID4.0: https://ieeexplore.ieee.org/abstract/document/7467436/ Man-Machine Partnership: https://www.cas.org/blog/evolving-partnership-future-man-and-machine Micro-plastics: https://phys.org/news/2018-05-microplastic-lowest-food-web-analysis.html Military 3D Printers: https://www.army-technology.com/features/made-measure-next-generationmilitary-3d-printing/ Moore’s Law: https://cacm.acm.org/magazines/2017/1/211094-exponential-laws-of-computinggrowth/abstract Nano Fabrics: https://www.tandfonline.com/doi/abs/10.1080/00405167.2011.570027 Nano Steels: https://www.sciencedirect.com/topics/materials-science/nanostructured-steel Programmable Materials: https://www.sciencedirect.com/science/article/pii/S0020768316302712 ?via%3Dihub Recycling Failure: https://www.economist.com/united-states/2002/07/04/waste-of-time Self Repairing/Healing Concrete: https://link.springer.com/article/10.1007/s10295-011-1037-1 Smart production Lines: https://www2.deloitte.com/insights/us/en/focus/industry-4-0/smart-factory-connected-manufacturing.html

Supporting Slide Sets https://www.slideshare.net/PeterCochrane/ai-fables-facts-and-futures-threat-promise-or-saviour-103261394 https://www.slideshare.net/PeterCochrane/ai-trust-and-safeguards https://www.slideshare.net/PeterCochrane/digital-destinieshttps https://www.slideshare.net/PeterCochrane/why-industry-40

Chapter 3

Connectivity for Industry 4.0 Kristina Gold, Kenneth Wallstedt, Jari Vikberg, and Joachim Sachs

Introduction In this chapter we introduce a connectivity technology outlook focusing on augmentation of human, machines, and systems capabilities. We provide an overview of relevant industry and manufacturing use cases and corresponding network requirements and outline the network architecture for both local connectivity and wide area connectivity. The proposed wireless network solutions are based on present best understanding when utilizing features and capabilities of cellular connectivity. The cellular technology platform performance is related to 3rd Generation Partnership Project standards such as 5G and LTE. Last, a vision on how to build a fully integrated connectivity solution for industries is presented.

Connectivity Technology Outlook Rapid advancements in the use of machines to augment human intelligence are creating a new reality in which humans increasingly interact with robots and intelligent agents. The ability to transfer human skills in real time to other humans and machines located all around the world has the potential to enable massive efficiency gains. Autonomous operation by machines with self-learning capabilities offers the additional advantage of continuous performance and quality enhancements.

The authors would like to thank all Ericsson colleagues that have contributed to valuable insights in the area of connectivity for industries. K. Gold (*) · K. Wallstedt · J. Vikberg · J. Sachs Ericsson AB, Stockholm, Sweden e-mail: [email protected]; [email protected]; [email protected]; [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_3

23

24

K. Gold et al.

Communication between humans and machines will become more natural, to the point that it is comparable to interpersonal interaction. Internet of Skills-based systems are characterized by the interplay of various devices with sensing, processing, and actuation capabilities near the user. A cyber-physical system is a self-organizing expert system created by the combination of model of models, dynamic interaction between models, and deterministic communication. A cyber-physical system presents a concise and comprehensible system overview that humans can understand and act upon. The main challenge for these systems is the orchestration of the networked computational resources for many interworking physical systems with different levels of complexity. An example of a cyber-physical system is a smart factory where mechanical systems, robots, raw materials, and products communicate and interact. This interaction enables machine intelligence to perform monitoring and control of operations at all plant levels. There is good reason to believe that rapidly increasing data volumes will continue in the foreseeable future (Ericsson 2018). To support new demanding use cases, ultrareliable connectivity with deterministic ultralow latency will be needed. The future connectivity platform delivers truly intuitive interactions between humans and machines. A set of intelligent network applications and features is key to hiding complexity from the users regardless of whether they are humans or machines. The future wireless access network is becoming a general connectivity platform that enables the sharing of data anywhere and anytime for anyone and anything.

Convergence of Technology Platforms The Fourth Industrial Revolution takes the step from mass production to mass customization. This starts with the merger of the latest capabilities within operational technology (OT), information technology (IT), and communication technology (CT). Currently automated factories are based on a hierarchical design with clear separation between shop floor, OT, and the top floor, IT. The interesting observation to make is that at the same point in time, connectivity solutions with low and predictable latency such as 5G, cloud-based technologies, and Industry 4.0 emerge making this convergence possible. Close collaboration is required between all relevant industry parties and standardization bodies to ensure that needs and requirements of a specific industry are adequately understood. This also includes addressing evaluation of relevant technologies and regulatory and business aspects. The communication market for industrial networks is presently dominated (>95%) (HMS Industrial Networks 2017) by fixed access technologies in the OT domain. It is a heavily fragmented market of Fieldbus and proprietary Industrial Ethernet solutions. The fragmentation is also due to noncompatible dialects within each technology domain. A feature being developed now is Time Sensitive Networking (TSN) as extension to the Ethernet standard supporting deterministic

3  Connectivity for Industry 4.0

25

latency requirements. Wireless solutions are only a small fraction of the installed base today; they mainly play a role for wirelessly connecting sensors where communication requirements are not too demanding. One near-term benefit of leveraging wireless connectivity in factories is reducing the amount of cables either because they are impractical or to reduce cost. Also, use cases can be supported that cannot be solved via fixed connectivity such as moving robots and automated guided vehicles or tracking products as they move through the production process. In addition, an increased floorplan layout flexibility and easier deployment of factory equipment can be achieved. Furthermore, it is expected to see a significant increased number of devices, robots, and sensors to support industrial automation. Wireless connectivity offers a flexible and cost-efficient solution to onboard, communicate, and track many different device types.

Connectivity Solutions for Industries Advanced connectivity solutions are key to support the evolution of industries and their mission critical industrial activities as well as less critical communication requirements. Some of the emerging technology tools and their applications are digital twins, smart workspaces, smart robots, and virtual assistants. Examples of implementation technologies are virtual and augmented reality, data augmentation, additive manufacturing, and IoT platforms (Lou et al. 2018). Hence the connectivity platform will need to handle an integration of a variety of connected sensors, machines, equipment, and human operators. Local connectivity will be key, and many of the use cases will require capabilities such as ultrareliable low-latency communication (URLLC), positioning, and life cycle management of both network infrastructure and industrial devices. Other aspects of integration are connecting the enterprises along the supply chain as well as tracking the life cycle management (LCM) of products and systems to support functions as predictive maintenance. Connecting enterprises on a regional, national, or global scale will require a wide area connectivity platform with fast, reliable, and secure connectivity. Common for both the local connectivity solution and the wide area connectivity solution is handling of network access, identities, security credentials, and algorithms of industrial devices, industrial sensors, and network infrastructure. The deployment and operation of connectivity for devices in field require a life cycle management approach covering phases like production installation, provisioning, and maintenance in field. Industrial devices and sensors are expected to be deployed in large numbers, to be in service for over 10 years (industry sector dependent), and not always accessible for (external) maintenance staff. Besides the devices, industries will leverage and deploy applications and its related data across the network. Applications will increasingly be deployed in a distributed fashion for various purposes such as safety and latency requirements as well as autonomous local operation. The connectivity nodes should support edge

26

K. Gold et al.

computing. The ability to automatically deploy and orchestrate both network application and industrial application workloads will be an intrinsic part of a future network solution. Mobile cellular communication technologies are going to play an increasing role for wireless connectivity for industries. In particular, 5G is considered of significant relevance for wireless connectivity for industries (5G Alliance for connected industries and automation White Paper 2018), as it is the first wireless communication standard that addresses the entire range of communication requirements needed for connected industries. 5G is standardized within the 3rd Generation Partnership Project (3GPP). Work on the initial core specification has started in 2017, and a first release has been published in mid-2018. It specifies a novel New Radio (NR) interface, as well as 5G core network functionality. 5G includes capabilities to support three service-type categories: massive machine-type communication (mMTC) for connecting sensors, critical machine-type communication (cMTC) providing ultrareliable low-latency communication (URLLC) for deterministic real-time applications, and mobile broadband (MBB) for high data rate applications (Dahlman et al. 2014a; Sachs et al. 2016). It should be noted that the 4G mobile communication standard long-term evolution (LTE) has been continuously evolved since its original standardization in 2008. In its later releases, also LTE includes support for mMTC and cMTC beyond the original support focused on MBB services.

Requirements for the Wireless Connectivity Solution Network and connectivity requirements for industries are covering a broad set of aspects. These aspects are of both technical and nontechnical in nature. The nontechnical aspects are listed in section “Additional Industry Requirements” below. Furthermore, the requirements are very use case and application specific, and the wireless network should support service differentiation to handle different use cases efficiently and simultaneously.

Technical Industry Requirements The following lists the main identified technical requirements, where several of them originate from many use cases requiring fully deterministic traffic. Latency down to 1 millisecond (ms) for specific use cases will be required. In addition, latency with a guaranteed upper bound, i.e., deterministic latency, is essential for critical automation use cases; packets need to arrive on time; otherwise they are considered lost. It is better to drop packets rather than to deliver them late. High synchronicity may also be required in use cases with less stringent latency requirements.

3  Connectivity for Industry 4.0

27

Throughput, i.e., data rates to be supported are very use case specific and will vary from a few kbps to tens or even hundreds of Mbps. Reliability is defined as the percentage value of the amount of sent network layer packets successfully delivered to a given node within the time constraint required by the targeted service, divided by the total number of sent network layer packets. Availability is typically quantified by the percentage of time during which a system operates correctly. Availability requirements are generally high for use cases impacting the production – production downtime costs money, and many factories operate 24/7. For wireless networks, availability can also comprise the coverage aspect, i.e., in a target area, the percentage of area covered. Positioning can be leveraged in process automation, for diagnostics and condition monitoring as well as motion control. High accuracy, down to a few tens of cm, and low-­latency positioning are seen as an enabler for future smart factories with flexible, modular production systems including more mobile and versatile production assets. Degree of local mobility is the ability to move freely throughout the factory floor. The service should continue during mobility with no interruption time and no packet loss. System determinism incorporates the ability to ensure applications’ adherence and system performance predictability independent of network load. The requirement originates mainly from automation demands requiring deterministic transmission in the order of sub-milliseconds. Wide area coverage and mobility will be required for, e.g., tracking and software updates of produced goods and for manufacturing equipment that moves outside that local industry plant such as trucks and for personnel. The requirement on wide area connectivity can be on regional, national, or global levels. This communication will be on the mobile operator networks, and some use cases will require handover between the local industry communication systems. URLLC performance is rarely required outside the industry plant. Number of industrial devices supported by the connectivity solution needs to be able to scale from few devices to thousands or millions. These devices may be for different use cases, and the number of devices shall not impact system performance in an unpredictable way.

Additional Industry Requirements Connectivity solution for industries needs to support the technical requirements described in the previous section. In addition, there are additional nontechnical industry requirements that should be supported by the connectivity solution. These industry requirements are described in the following: Full control of critical connectivity: The starting point from an industry standpoint is full control of critical connectivity, i.e., connectivity related to the OT parts.

28

K. Gold et al.

This is an inbuilt property of the existing connectivity solutions due to local deployment and is considered essential to fulfill the industry’s internal 24/7 performance targets. Local survivability: The connectivity solution is not to be dependent on any external failures, i.e., it shall be self-contained when it comes to survivability. This also includes support of redundancy and failover times in the order of 100 ms for real-­ time type of control and communication. Failure to fulfill the critical use case requirements may lead to production stop for the machines affected or the whole production cell. Frequent loss of connectivity can therefore have large negative cost impacts. Local content: Production-related data may not leave the industry/factory premises, i.e., all such data needs to be kept locally of, e.g., security and trust reasons. Local management: The management solution including network observability must be suitable for the industry. This implies ease of use for operating the connectivity solution by the industry. In addition, it must be easy to integrate with the industry’s business and operational processes, which includes integration of both existing OT equipment and factories’ IT systems. Life cycle management (LCM): Several industries require LCM in the range of tens of years. This means that long-term availability of industrial devices, network infrastructure, and related services is vital. Industrial devices require an LCM covering initial device configuration, firmware updates, application software updates, provisioning of identity credentials, installation, provisioning, and maintenance in field. This LCM must be automated and remotely managed. The management shall handle deployment and integration of both brownfield and greenfield devices covering connectivity, security, LCM, and application protocol adaptation. Device types range from typical MBB devices such as smartphones to industrial devices such as machines, robots, gateways, and sensors. Industrial devices with an established ecosystem and related economy of scale benefits. Identity management is key to handle, e.g., deploying and revocation of different identities throughout the life cycle of an industrial device including provisioning of an anchor for secure boot for, e.g., software upgrades with signed software. Similar requirements are valid also for network infrastructure. Security: Security has many different functional elements and termination points. The connectivity network shall assure that only authorized traffic is allowed and with the required level of confidentiality protection (e.g., encryption and/or integrity protection) applied. Functionalities like protection against intrusion (hackers) from the Internet, malware reaching devices and servers, tampering of data, etc. shall also be supported but are not typically part of the connectivity solution. As OT in most cases represent critical operations, where any attack can have significant consequences, the OT security requirements are generally much stricter than for IT.

3  Connectivity for Industry 4.0

29

Integration with existing solutions: The connectivity solution needs to be integrated to existing wired OT system as well as to other wireless connected devices through gateways. One example is transport of Industrial Ethernet frames. This requirement is also about integration with the industry’s business and operational processes, which include integration of both OT equipment and IT systems.

Industrial Use Cases The industry domain is very versatile; for instance, take the differences between discrete manufacturing and process industry. Thus, applications and tasks may differ substantially in between different industry sectors but also within each specific industry sector. As an example, within an industry sector, there are applications such as supply management and operation control to robot motion control. In Fig. 3.1 a set of wireless application examples in a factory is outlined. Different aspects of each application should be analyzed and taken into consideration, to mention a few, quality of service, security and safety, availability and reliability, and deployment scenarios. To provide wireless connectivity to these different use cases, a clustering of the requirements gravitates toward three separate service types (5G Alliance for connected industries and automation White Paper 2018). The first set is, in the case of 5G, labeled as evolved mobile broadband (eMBB). These are the use cases typically requiring high data rates, high capacity, and wide area connectivity. Examples are fleet management, augmented and virtual reality, and remote access and management. The second set of use cases is labeled massive machine-type communication (mMTC). In this case the typical use case involves communication requiring low data rates and no stringent requirements in terms of latency. A massive wireless sensor network such as goods tracking is a good example of a typical use case.

Fig. 3.1  Examples of applications and tasks leveraging wireless connectivity. (Source: Ericsson)

30

K. Gold et al.

Examples of cellular access technologies addressing these service types are Cat-M and NB-IoT (Liberg et al. 2017). The third set is the most demanding from a connectivity perspective. These are the critical machine-type communication (cMTC) use cases. As an example, URLLC is an implementation of a cMTC targeting specific manufacturing use cases. These are the use cases having the highest requirements in terms of reliability, availability, deterministic latency, and cycle time. The typical payload size is not very high, and the service area is also limited. Motion control, mobile robots, process control, and assembly line automation are some examples. The connectivity network platform should handle any combination of eMBB, mMTC, and cMTC service types.

Industry Automation The wireless connected use cases are today typically of less critical nature such as monitoring and parametrization. In factories the traffic consists mainly of real-time traffic, which is carried by protocols with highly integrated protocol stack (e.g., PROFINET real-time stack or Ethernet TSN). The TCP/IP protocol stack is mainly used for carrying messages pertaining to start-up configuration, notifications, and non-­critical alarms. A wireless factory network shall support service differentiation to serve the different use cases efficiently, and it is preferable that one network infrastructure can support all connectivity needs of the factory. Coexistence and integration with existing connectivity solutions is necessary, and capillary networks should also be considered, for example, for sensor networks. The wireless factory is also a step closer to a fully flexible production/assembly line, for example, the moveable assembly lines with AGVs which move from assembly station to assembly station. Here reliable wireless technology is a necessity. Use cases considered representative for the manufacturing industry and also as they could get some benefit from wireless deployment are: • Robotized assembly line –– Programmable logic controller (PLC) to robot controller communication –– PLC to PLC communication –– Robot controller to robot communication • Real-time manufacturing process optimization • Automated guided vehicles (AGV) –– Including automated storage and retrieval systems (AS/RS) • Monitoring and preventive maintenance • Critical monitoring and emergency shutdown • Mobile augmented reality (AR)

3  Connectivity for Industry 4.0

31

Isochronous Real-Time applications Typical example: Motion Control

Real-Time applications Typical example: Assembly lines

Non-Real-Time applications Monitoring, parametrization, …

Production line Robot control Packaging Electro-hydraulic cylinder controllers in presses

Electric motor control Laser cutting Printing 100us

1ms

10ms

~50ms

100ms

Latency

Fig. 3.2  Examples of latency requirements for a set of use cases. (Source: Ericsson)

High Multefire Wi-Fi

Reliability (with load)

LTE

NR Licensed

Control loops with guaranteed latency Unlicensed

Non-realtime control

Realtime control, e.g.

Cable replacement

Automated Guided Vehicle

Low

Monitoring

Low

(e.g. seconds)

E2E latency requirement

High

(e.g. milliseconds)

Fig. 3.3  Estimated performance of different wireless technologies with respect to reliability with increasing load (y-axis) and increasing E2E latency requirement (x-axis). Values are estimates. (Source: Ericsson)

The most critical use cases may only need to be supported in certain physical locations of the factory, i.e., only in specific production cells. Deterministic latency is expected to be the dominating deciding factor on whether a use case can be deployed through a wireless setup. See Fig. 3.2. Today wireless solutions are implemented for less latency demanding use cases and hence for non-real-time applications. The wireless segment today is dominated by Wi-Fi solutions and wireless sensor networks in proprietary variants building on IEEE 802.15.4. The next generation of wireless system, such as 5G, will be able to handle deterministic real-time applications down to 1 ms already on 2020. Wireless solutions will in the future be able to handle deterministic latencies below 1 ms. Figure 3.3 depicts an estimated difference between Wi-Fi, MulteFire, LTE, and 5G NR with regard to increasing reliability demands and increasing end-to-end (E2E) latency demands. MulteFire is an adaption of the LTE standard for usage in license-exempt spectrum and is specified by the MulteFire Alliance. Example use cases are placed on the figure.

32

K. Gold et al.

Tactile Internet Tactile Internet is the support of physical interaction with real or virtual objects across geographical distances. There are two central aspects in the development of Tactile Internet and its services: the haptic interactions enabled by numerous sensing technologies and robotics in combination with haptic communications and the transport of physical interactions over long distances. A central use case requirement is to enable real-time interaction for humans with the physical world via the network. The time constants required for “real-time” interactions differ for different types of physiological interactions. Typical response times are described as 1  s (muscular), 100  ms (audio), 10  ms (visual), and 1  ms (tactile). A set of application fields are identified, which all are claimed to require end-to-­ end latencies down to 1 ms: • Remote-controlled robotics • Shared virtual environments as multiple players at different locations performing interacting tasks in virtual reality • Real-time augmented reality for different types of assistance systems • Serious gaming in the area of education, training, simulation, and health, e.g., adaptive and personalized cardio-training All these use cases have a “human-in-the-loop,” and the term “tactile” represents this interaction of the human with a physical – and in some cases virtual – e­ nvironment. The Tactile Internet is sometimes also associated with skillset delivery as it enables the human-in-the-loop to perform tasks from remote as shown in Fig. 3.4. The Tactile Internet is enabled by bringing a variety of technology components together into a complete system for reliable end-to-end real-time interactions. It enables to perform complex tasks remotely, which today require local presence. Important technology components include: • URLLC • Haptic codecs • Networked control and automation Remote control (position / velocity)

Feedback (audio / video / force / position)

Operator (master domain)

Teleoperator (remote slave domain)

Fig. 3.4  Example teleoperation system (see, e.g., Sachs et al. (2018)). (Source: Ericsson)

3  Connectivity for Industry 4.0

33

• Distributed cloud, edge computing, software-defined networking, and software functional mobility • Augmented and virtual reality (AR/VR) • Low-latency video/audio coding and synchronization The Tactile Internet will enable immersive remote operations and interactions with a physical world. Wireless communication will play a fundamental part as it will provide necessary capabilities for the demanding communication requirements in terms of reliability and low latency, for operators or teleoperated systems (Sachs et al. 2018). In the context of industries, Tactile Internet communication plays a role for remote control and operation of industrial equipment, either during a production process or as part of maintenance operations. Remote operations can increase production efficiency and safety; it allows, for example, to involve experts that are not on-site into ongoing operations or to minimize human presence in hazardous environments. Significant research and concept studies have been made, e.g., in remote control in mining.

Local Connectivity for Industries When analyzing both technical and additional industry requirements in depth, the conclusion made is that at least the connectivity solution for the critical parts needs to be fully deployed locally within the industry premises. Hence, a local connectivity solution is required. The solution including functionality and components required is described in section “Network Functionality and Components Required for Critical Use Cases”. In addition, it is beneficial that the same connectivity solution can support also other use cases such as non-critical MBB and mMTC. A solution option enabling support for multiple use cases is outlined in section “Solution Options for Local Connectivity”.

 etwork Functionality and Components Required for Critical N Use Cases The local connectivity solution for critical use cases for industries is shown on high level in Fig.  3.5. The solution is a stand-alone (private) local deployment within industry premises. The main highlights are: 1) Local network infrastructure at industry premises consisting of the following parts: (a) Radio Access Network (RAN): 3GPP LTE and NR are shown as good examples of RAN supporting URLLC. The URLLC RAN connects the industrial

34

K. Gold et al.

Industry premises - OT 3GPP LTE/NR

Licensed spectrum Network Identity

Positioning

OT OSS

Core Control Plane

Industrial Applications

Integration

Core User Plane

Legacy OT

SIM

Fig. 3.5  Local connectivity solution for industries. (Source: Ericsson)

devices to the local network infrastructure and controls the radio links toward the devices. The main features are based on deterministic low latency, network-­controlled QoS, predictable performance even in high-load situations, and multi-connectivity solutions to further improve availability and reliability in the RAN part. (b) Core Network (CN): CN is shown with Control and User plane separation. The Core Control Plane handles industrial device access to the local network and, e.g., authenticates the industrial devices. The Core User Plane part handles the actual user plane traffic, for example, between industrial devices and industrial applications. The CN part shall also support deterministic low latency, high reliability, availability, and redundancy. Redundancy can be achieved either via a single redundant solution or multiple local deployments of the CN parts. Although not shown in the figure, even the transport network between and within RAN and CN parts shall be designed for deterministic low latency, high reliability, availability, and redundancy. (c) Positioning: High accuracy and low-latency positioning services enable future smart factories with flexible, modular production systems including more mobile and versatile production assets. Positioning support is deployed “above” the Core Control Plane. (d) OT Operational Support System (OT OSS): OT OSS is used to monitor and manage the local network infrastructure. The OT OSS should support small footprint and monitoring/management with ease of use. It should also include software management and fault, performance, and configuration management. The OT OSS shall have capabilities to easily integrate both elements of OT systems and the industry IT systems. (e) Cloud Infrastructure: It is used for industrial applications and, for example, for Core Control and User Plane and positioning functions.

3  Connectivity for Industry 4.0

35

2) Spectrum: The technical requirements for very high quality-of-service requirements mean that there must be guarantees toward uncontrolled interference, implying that licensed spectrum is necessary. Spectrum aspects are further described in section “Spectrum Access for Industries”. 3) Network Identity: The local network requires a network identity. One example is a Public Land Mobile Network Identity (PLMN-ID) used also in existing wide area cellular networks, and it has the benefit of wide usage in existing devices. Other network identities are also possible implying that new devices are needed. 4) Industrial Devices: An industrial device is connected via the local network to locally deployed industrial applications. The traffic traverses from the industrial device to the RAN and further to Core User Plane and industrial applications. Another important case is industrial gateways that connect and/or aggregate multiple legacy industrial devices. 5) Security Credentials and Device Identities: The industrial devices and gateways need security credentials for accessing the local network. In addition, each device/gateway needs a unique identity associated with the security credentials. One example is usage of international mobile subscriber identity (IMSI) as device/gateway identity and subscriber identity module (SIM) for storage of the security credentials and related security algorithms. The SIM credentials have tradionally been stored on a removable universal integrated circuit card (UICC), and the physical SIM card must be changed every time the operator is changed. Embedded SIM (eSIM) technology provides new possibilities with a eUICC (embedded UICC) soldered on the device. The eUICC can be remotely provisioned and may contain more than one SIM profile. The active SIM profile can also be changed dynamically. Other identities and security credentials are also possible, for example, private key infrastructure (PKI) with certificates may be used as security credential. The device/gateway identity also needs a related network database (can be seen as part of Core Control Plane) and provisioning system also enabling LCM of these identities. 6) Integration to Legacy OT: The wireless local connectivity solution, e.g., 5G, is part of a complete local solution, and an integration effort is required with other OT connectivity (both traffic integration and integration to management systems).

Solution Options for Local Connectivity The stand-alone local deployment described in the previous section fulfills the current industry requirements for critical use cases. It could also be desired that the same local network infrastructure is used for other non-critical use cases, i.e., to have one network supporting all industry-related use cases. Such non-critical use cases are MBB including voice and many of the mMTC use cases. The stand-alone local deployment is not the only solution option for functionality related to non-­ critical use cases as such functionality can be deployed outside industry premises at, e.g., a mobile network operator (MNO) premises.

36

K. Gold et al. MNO premises or Industry Data Center

Industry premises - OT 3GPP LTE/NR

Licensed spectrum Network Identity

Positioning

Non-critical use cases

Core Control Plane

Industrial Applications Core User Plane

SIM

CentralOSS

OT OSS

Wide-area coverage Core Control Plane

Integration

Industrial Applications

Legacy OT

Transport

Fig. 3.6  Solution options for local connectivity for industries. (Source: Ericsson)

The next solution option is therefore to deploy selected parts of the connectivity solution at a more central network site, for example, at MNO sites/premises or at an industry player data center, as illustrated in Fig.  3.6. Different opportunities are described after the figure, and some of these opportunities imply the need to relax existing additional industry requirements. This solution option may be preferred to support one or more of the following areas: 1) Management of local network infrastructure can be divided between OT OSS and a central OSS. The OT OSS can be used for local monitoring of the connectivity solution, and more advanced connectivity management operations could be performed by the MNO or an industry player. This would decrease the demand for local competence for management of the local network. The main question is the functional and capability division between OT OSS and central OSS, and it will be determined by the business model agreed. Furthermore, if the central OSS would also be used for management of connectivity for critical use cases, then the current industry requirements on “full control of critical connectivity” and “local management” would need to be relaxed. 2) Support for non-critical use cases, such as MBB/voice and non-critical mMTC in the local area. In this case the local RAN would be used to support both critical and non-critical use cases. The local RAN can support the needed mechanisms based on network sharing or network slicing. The CN and industrial applications for critical use cases would still be locally deployed. CN and application servers for MBB/voice and mMTC could be deployed more centrally, enabling the reuse of such functionality already deployed in more central network sites. 3) Support for non-critical use cases in wide area coverage. It is obvious that all use cases requiring wide area coverage can only be supported from wide area ­networks provided by MNOs. For example, during production there is a need for local connectivity meeting critical OT requirements, while the final product potentially requires non-critical wide area connectivity and mobility.

3  Connectivity for Industry 4.0

37

4) Core Control Plane may be placed more centrally. The basic thinking in this case is that the Core User Plane and industrial applications would be handled locally, to support deterministic low latency and other critical requirements. A more centrally placed control plane could have economic benefits as the same network infrastructure can be used for multiple industry premises. Redundancy in the control plane functionality is also likely feasible to support in a more central deployment. In addition, the transport networks between industry premises and more central network sites should be designed for redundancy. If a more central Core Control Plane would be used for management of connectivity for critical use cases, then the current industry requirements on “full control of critical connectivity,” “local survivability,” and “local management” would need to be relaxed. 5) A central cloud for industrial applications. In this case even the industrial applications, and potentially Core User Plane, would be deployed in a more central location. This case could also have economic benefits as the same cloud execution environment can be used for multiple industry premises and even for industrial applications that have logic related to multiple industry locations. Redundancy in the cloud execution environment is likely feasible to support in a more central deployment. In addition, the transport networks between industry premises and more central network sites should be designed for both redundancy and low latency. If a more central cloud execution environment would also be used for critical use cases, then the current industry requirements on “full control of critical connectivity,” “local survivability,” “local content,” and “local management” would need to be relaxed.

Wide Area Connectivity for Industries The local connectivity is strongly coupled with the cMTC use case, while the wide area connectivity is mainly associated with less critical communication needs. Examples for the wide area are related to use cases as smart logistics – end-to-end view of supply chain with logistics data interchange, freight management, connected freight, and product maintenance. The relation to OT may be in the areas of “connected flows” (before OT) and “connected products” (after OT), i.e., these parts will benefit from wide area service and mobility but are seen as non-critical when compared to OT. Such wide area service is typically supported by MBB and mMTC services from the mobile networks. Today, mobile networks cover around 95 percent of the world’s population and continue to expand. For mobile broadband (WCDMA/HSPA or later technology), population coverage is currently at around 80 percent and forecast to grow to around 95 percent in 2022 (Ericsson 2018).

38

K. Gold et al.

Massive Machine-Type Communication Massive machine-type communication is developed to provide wide area coverage and significant signal penetration for thousands of devices per square kilometer of coverage. An additional objective of mMTC is to provide ubiquitous connectivity with relatively low device complexity and low-energy operation. Many of the devices supported are battery powered or driven by alternative energy supplies, have moderate payloads, and might rarely be active, so they tend to some extent be delay tolerant. While the devices typically have a long life span, services and software have to scale and be enabled to be swapped out fast to address new use cases. Examples that fall into this service category include the monitoring and automation of buildings and infrastructure, tracking, logistics, fleet management, and smart agriculture. The following key features are identified for the mMTC use case (Ericsson 2016): • Reduced device complexity and cost  – a key enabler for high-volume, mass-­ market applications. Reduced device size is done by reducing peak rate, memory requirement, and device complexity. • Long battery life (more than 10 years) – the cost of replacing batteries in the field is not viable. More than 10 years of battery life can be achieved by introducing power saving mode and/or extended discontinuous reception (eDRX) functionality. These features allow the device to be contacted by the network – or contact the network– on a per-need basis, meaning that it can stay in sleep for days if applicable. • Extended coverage (15–20 dB) – regional (or even national or global) coverage is a prerequisite for many use cases especially within the transport area. An improvement of 20 dB translates into a significant increase in the outdoor coverage area and significantly improved indoor signal penetration to reach deep indoors. • Support for a huge number of devices – the network capacity must be easy to scale to handle millions of devices. Core network enhancements include signaling optimization, software upgrades for service differentiation handling, and high-capacity platforms. The basic architecture for mMTC may look simple, but challenges arise from the support of large number of devices. This leads to an increase in the control signaling relative the amount of user data traffic. In addition, the mobility tracking of devices can further increase the control signaling on the network. The radio access is executed at the base station close to the devices, while the other main functional components such as the Core Control Plane that also includes the mobility management, subscriber data management (SDM), and the Core User Plane are conducted in the central primary site.

3  Connectivity for Industry 4.0

39

Evolved Mobile Broadband Providing both extreme high data rate and high data capacity, evolved mobile broadband (eMBB) also offers significant extended coverage – well beyond that provided by MBB, e.g., delivered by LTE.  Connectivity and bandwidth are more uniform over the coverage area, and performance degrades gradually as the number of users increases. In the past wide area coverage and MBB, much of the focus has been on the peak data rate that can be supported by a wireless access technology under ideal conditions. However, a more important capability is the data rate that can be provided under real-life conditions in different scenarios. In the case of 5G, the following performance is targeted (Dahlman et al. 2014b): • Data rates exceeding 10Gbps should be supported in certain scenarios, for example, indoor and dense outdoor environments. • In urban and suburban environments, data rates of several 100Mbps should typically be achievable. • Almost everywhere data rates of 10Mbps or more should be accessible; this includes sparsely populated rural areas in both developed and developing countries. One key component of 5G radio access is an innovative air interface called New Radio (NR), which is designed primarily for spectrum assets which span across low-, medium-, and high-spectrum bands. Other key technology components for the 5G radio access solution include (Ericsson 2017): • Advanced multi-antenna technologies such as beamforming with phased antenna arrays and massive MIMO. Beamforming, which focuses radio transmission from multiple antenna elements using narrow beams, reduces interference and improves overall system performance. • Ultra-lean transmission to reduce interference caused by common signaling resources as well as maximize resource efficiency. Cutting always-on transmissions to a bare minimum, so that communication only occurs when there is user data to deliver, allows the transmitter to dynamically – on a millisecond basis – switch off and be silent. Such ultra-lean transmission results in enhanced network energy efficiency and higher achievable user data rates. • Flexible duplex, in which spectrum resources are dynamically assigned to downlink and uplink, enables transmission up to the full bandwidth to be used momentarily in either direction and is a better fit for dynamic traffic conditions. • Access/backhaul integration, where wireless access and wireless backhaul share the same technology baseline and the same spectrum pool, making more efficient use of spectrum resources as they can be shared dynamically. Wireless connectivity between radio network nodes and the rest of the network simplifies deployment – especially in dense deployments that have large numbers of nodes. These technology components will not only apply to the new 5G wireless access systems but also to the evolution of LTE.

40

K. Gold et al.

Wide Area Connectivity Wide area connectivity requirements can be national or global depending on the mobility needs of the devices and products associated with a specific industrial use case. Different solutions are also possible for both cases. The national wide area connectivity is typically supported by one or more MNOs. The global wide area coverage requires support from multiple MNOs. See Fig. 3.7. The different alternatives to achieve wide area coverage are as follows: (a) The existing roaming model is one alternative for global wide area coverage and can be applicable, e.g., for devices mainly used in one area but also moving during their operation. In this case the device is equipped with credentials and identity related to a single MNO and relies on this MNO’s global roaming agreements with other MNOs. (b) Another alternative is that the device is initially equipped with identity and credential for a single bootstrapping operator having global roaming agreements with many MNOs and typically even many MNOs within a country. The initial connectivity provided by the bootstrapping operator can be used to provision the device with the most suitable MNO identity and credentials for the current geographical location of the device using eSIM technology. The identity and credentials in the device may also be updated again during the lifetime of the device. This solution is typically valid for devices that are shipped globally and move between different areas. (c) A third alternative is that the device is equipped with multiple credentials and identities associated with multiple MNOs based on eSIM technology, i.e., even that the identity and credentials are provisioned dynamically into the device. In this case, it would become possible to have new connectivity management logic, to always select the best possible wide area connectivity by changing the current identity and credentials of the device based on, for example, the location of the

Fig. 3.7  Connectivity on different levels. (Source: Ericsson)

3  Connectivity for Industry 4.0

41

device. This alternative can potentially be extended with having the device equipped with multiple modems for connectivity. Each modem can be connected to different networks, potentially even simultaneously. In this case, the connectivity management becomes more advanced as selection of different networks can be more dynamic and aggregation of different networks is also enabled. There are also cases when the device must be connected to both local and wide area connectivity as illustrated in Fig. 3.8. One solution is to have separate credentials and identities for local and wide area connectivity. It is also interesting to use the same credentials and identity for both local and wide area connectivity.

Fig. 3.8  Identity handling for local and wide area connectivity. (Source: Ericsson)

42

K. Gold et al.

Therefore, the local and wide area connectivity solutions need to be integrated. There are different ways to perform this integration as well depending on which entity is responsible for the device credentials and identities. In the first case, the MNO is responsible for these. The MNO’s wide area connectivity and roaming agreements can be used, but a solution is required to support the local survivability requirement of the industry. In the second case, the industry is responsible for the device credentials and identities.

Spectrum Access for Industries The availability of spectrum resources is a key element to support the connectivity requirements for most industries, both for wide area use cases and for locally deployed networks. The usage of the spectrum resources shall be managed such that the service levels can be delivered with certain predictability. For local deployments (e.g., a manufacturing floor), sufficient bandwidth must be available to meet requirements on bitrates and latency; here estimates of spectrum needs are in the range of tens of MHz to hundreds of MHz. Considered frequency bands are in lower bands (below 2GHz), in the mid-band (2–5 GHz) and in higher bands (above 10 GHz). For applications with very high quality-of-service requirements, there must be guarantees toward uncontrolled interference, implying that licensed spectrum is necessary. Unlicensed spectrum may be relevant for non-critical connectivity. To fully address the URLLC/cMTC use cases, 5G/NR on higher bands is considered. Thus, as the solutions and deployments are being built out, there are likely multiple frequency bands being used, e.g., lower bands to cater for the initial deployments of LTE and NB-IoT and higher bands for the roll out of URLLC/cMTC services. In addition, the device ecosystem is dependent on a harmonized spectrum on major markets to meet economy of scale. There are three different ways that licensed spectrum could be put into use: • MNOs use their own licensed spectrum while also providing the connectivity equipment. In this case spectrum is readily available, though a long-term servicelevel agreement between MNO and enterprise is required. • An enterprise leases spectrum from one or more MNOs for the local area of their operation. With the more “localized nature” of radio signals (“confinement” within buildings due to high penetration losses) at higher-spectrum bands, a spectrum leasing arrangement for industry applications has a potential to become more attractive to MNOs for mid-band or mmWave 5G spectrum bands. • Another solution is to set aside dedicated spectrum for licenses for industries that will deploy their own network or have a third party deploy it for them. This would ensure long-term access to spectrum for industries without the need for  agreements with MNOs. Lately there has been considerable attention for licenses for more geographically limited networks, e.g., in the USA (CBRS 3.55–3.7 GHz) and some European countries including Germany and Sweden

3  Connectivity for Industry 4.0

43

(3.7–3.8 GHz). Also, mmWave frequencies are considered for industries, where larger amounts of spectrum may become available and “confinement” within buildings are easier. An alternative for industry spectrum may be to use regionally unused or underused 3GPP spectrum, especially in cases where there is a device ecosystem available, thanks to deployments in other parts of the world, such as 2.3–2.4 GHz.

Vision of Fully Integrated Connectivity Solution for Industries The vision of fully integrated connectivity solution for industries is described in the following and is a summary of the previous sections. First, the industrial design paradigms are described in section “Industrial Design Paradigms”. Section “5G for Industrial Connectivity” gives a view of 5G applicability for industrial connectivity for the different design paradigms. Finally, section “Vision of Fully Integrated Connectivity Solution” provides a high-level vision of fully integrated connectivity solution for industries.

Industrial Design Paradigms The connectivity solutions for industries available today are typically based on a hierarchical design (e.g., according to the ISA (International Society of Automation) Committee for Manufacturing and Control Systems Security and their ISA-99 Plant Logical Framework). Typically, OT is isolated entirely from the IT and Internet via a so-called DMZ (demilitarized zone), and any interactions toward OT are limited to a very controlled set of data over very secure interfaces terminated at the DMZ. The hierarchical design in factory networks is largely due to fixed connectivity and heterogeneous technologies. A fixed Fieldbus network on the shop floor has a physical limitation in size. Many Fieldbus networks exist which are connected via some gateways to the rest of the network. Different parts of the factory may also use different Fieldbus technologies; a gateway is required to make it interwork with the rest of the network. As a result of this, a hierarchical design is achieved, where local Fieldbus islands exist and are connected via gateways. This design puts limitations on digitalization of factories, as information within one part of the factory cannot be easily extracted and exposed to the digital representation of the factory process (the digital twin). This limits the flexible steering and automation of the factory processes. The hierarchical design may however have a benefit regarding security since it achieves some isolation of the different subnetworks. Industry 4.0 is the next era in industrial production, aiming at significantly improving the flexibility, versatility, usability, and efficiency of future smart factories. This is also a move from the hierarchical design of today to a fully connected

44

K. Gold et al.

design paradigm. The fully connected factories become an enabler for full industry digitalization and will provide flexible systems and machines; they will be based on cloud technologies, as well as connectivity based on Ethernet TSN and Internet technology.

5G for Industrial Connectivity 5G-based connectivity solutions for industries will need to support both hierarchical and fully connected design paradigms as the move toward Industry 4.0-based systems will take long time due to typical solution lifetime in factories. The hierarchical design is most commonly used today, and trial deployments of Industry 4.0 are already taking place. 5G characteristics are actually better suited for the fully connected factory. For example, when introducing 5G, it may be difficult to replicate a “hierarchical” connectivity model, since all devices within 5G coverage can in principle access the same network. This means that hierarchical separation needs to be added on top, e.g., by different mechanisms in the network to separate IT and OT traffic. As stated above, the original hierarchical design has a benefit regarding security, since it achieves some isolation of the different subnetworks. For Industry 4.0, the security design is likely to be based on more digital approaches of separating entities and security domains. The telecom industry is evolving wireless connectivity solutions as 3GPP standardizes the new 5G system (5GS) consisting of NG-RAN and 5G Core (5GC) to improve support for MBB, cMTC (incl. URLLC), and mMTC use cases. One main development in NG-RAN, in addition to evolved LTE, is about defining the New Radio (NR) Radio Access Technology (RAT) to support, for example, deterministic low latency, network-controlled QoS, predictable performance in high-load situations, and multi-connectivity solutions to further improve reliability and availability. 5G is therefore the main enabler for meeting the demanding connectivity requirements from industries when moving toward Industry 4.0.

Vision of Fully Integrated Connectivity Solution Figure 3.9 describes a high-level vision of fully integrated connectivity solution for industries. Further description is given after the figure and starts with generic characteristics and then also shows examples of 5G functionality for the different parts. The main aspects of Fig. 3.9 are as follows: 1. The fully integrated connectivity solution supports both wide area and local area connectivity for industrial use cases. The critical use cases are supported in the local area by the local area network infrastructure. This means that all of network infrastructure and support for industrial applications is located within the local area (as described in section “Network Functionality and Components Required for Critical Use Cases”).

45

3  Connectivity for Industry 4.0

Wide area Non-critical use cases Wide area network infrastructure

Local area Non-critical use cases Critical use cases

Legacy OT

Local area network infrastructure

Fig. 3.9  Vision of fully integrated connectivity solution for industries. (Source: Ericsson)

In the wide area, the support for non-critical use cases is solely based on the wide area network infrastructure. For the local area, there are three different options for support of non-critical use cases: (a) Solely based on the wide area network infrastructure. In this case the wireless connection to the devices located in the local area is from the wide area network. This case is also called as “outside in” indicating that there is no network infrastructure deployed within the local area for the non-critical traffic flows. (b) Solely based on the local area network infrastructure in a similar way as described above for critical use cases. (c) Based on combination of both local and wide area network infrastructure. In this case, the wireless connection to the industrial devices is supported by the local area network infrastructure, and parts of the connectivity are supported by the wide area network infrastructure. This means that the local area and wide area network infrastructures are integrated. 2. Therefore, the fully integrated connectivity solution supports flexible deployment of functionality for critical and non-critical use cases. Non-­critical use cases are supported with selected parts, or even all, deployed outside the local area, for example, either at MNO or industry player sites. The support for different use cases is based on network slicing and/or network sharing mechanisms in the wide and local area network infrastructures. Critical use cases can be supported with full local deployment. An interesting future area to investigate is the possibility to relax the additional industry requirements on, e.g., full control of critical connectivity and local survivability while fulfilling the technical requirements on, e.g., latency, reliability, and availability. This may open up for flexible deployment of functionality in which parts of the functionality supporting critical use cases are placed outside the industry premises.

46

K. Gold et al.

3. The industrial devices can, depending on use case requirements, support either both local and wide area connectivity or only local area connectivity. As 5G is used for both local and wide area connectivity, the same device can be used in both cases with an established ecosystem with economy of scale benefits. 4. The local area network infrastructure contains: (a) For critical use cases, a complete local 5G network deployment (supporting cMTC, mMTC, and MBB services), OT OSS, positioning and support for integration to existing OT. For non-critical use cases, the complete local deployment may be used or only parts of it while the rest of the connectivity support is outside the local area. The 5G network architecture supports flexibility in deployment of functionality depending on use case needs and agreements with wide area connectivity providers, i.e., MNOs. (b) The local RAN also supports network slicing and/or sharing to connect the different industrial devices to different core networks for the different use cases. In addition, it may be required that a single device is simultaneously connected to network infrastructure supporting both critical and non-critical use cases. (c) Mobility within the local area is supported by well-proven network-­ controlled mechanisms as standardized in 3GPP. The industrial devices can move freely (handover) within the local area with service continuity and without packet loss. Mobility between wide area and local area may also be supported if required for a specific use case. (d) High-accuracy, reliable, and low-latency positioning is supported by NR in the local area network infrastructure. The related standardization work is still ongoing in 3GPP with the goal to support positioning accuracy of a few tens of cm and reliability of 99%. (e) Spectrum resources with sufficient bandwidth are allocated for the supported use cases. For the critical use cases, licensed spectrum with guarantees toward uncontrolled interference is used. One deployment example is to use multiple frequency bands, e.g., lower bands to cater for the initial deployments of LTE and NB-IoT for non-critical use cases and higher NR bands for the roll out of URLLC/cMTC services (see section “Wide Area Connectivity for Industries”). (f) A network identity that preferably supports unique identification of each local area network. The unique network identity can be used to support, for example, access control to the local area network enabling the devices not allowed to connect to the local area network to not even attempt to connect. Standardization work is ongoing in this area, and different possibilities include usage of PLMN-ID, extensions to PLMN-ID, or other new network identities. (g) The industrial devices are allocated user identities and related security credentials enabling that access of industrial devices to the local area connectivity can be authenticated and any communication can be protected with

3  Connectivity for Industry 4.0

47

encryption and integrity protection based on well-proven 3GPP security mechanisms. One good option is the usage of IMSI as user identity and SIM/eSIM for storage of IMSI, security credentials, and related security algorithms. Standardization work for other user identities and security credentials is also ongoing. To summarize, the vision of fully integrated connectivity solution supports all industry requirements, both technical and additional, for critical use cases. In addition, non-critical use cases are also supported. Local area and wide area connectivity are integrated to support the use cases with need for both. The wide area connectivity is based on reuse of existing MNO networks. The vision provides less cabling in the factory with increased floorplan layout flexibility and easier deployment of factory equipment. The full connectivity solution supports network-controlled QoS and predictable high-load performance with industrial devices based on an established ecosystem. The journey toward successful integration of 5G and Industry 4.0 has started and needs to be a common journey for OT, IT, and CT players. The competence and technology enablers required for successful integration exist in all sides. Close collaboration is required between all relevant industry parties and standardization bodies to ensure that needs and requirements of a specific industry are adequately understood.

References 5G Alliance for connected industries and automation White Paper, “5G for connected industries and automation”, April 2018. Dahlman, E. et al. (2014a, December). 5G wireless access: requirements and realization. In IEEE Communications Magazine, Vol. 52, no. 12, pp. 42–47. Dahlman, E., Mildh, G., Parkvall, S., Piesa, J., Sachs, J., & Selén, Y. (2014b, June 18). “5G radio access”, Ericsson Technology Review Ericsson. (2016, January). Cellular networks for Massive IoT  – enabling low power wide area applications. Ericsson White Paper. Ericsson. (2017, January). 5G systems  – Enabling the transformation of industry and society. Ericsson White Paper. Ericsson. (2018, June). Ericsson mobility report. HMS Industrial Networks, 2017. Liberg, O., Sundberg, M., Wang, Y.-P. E., Bergman, J., & Sachs, J. (2017). Cellular Internet of Things. Academic Press, UK. Lou, D., Höller, J., Whitehead, C., Germanos, S., Hilgner, M., Qiu, W., & Sharma, M. (2018). Industrial networking enabling IIoT communication. Industrial Internet Consortium White Paper. Sachs, J., Popovski, P., Höglund, A., Gozalvez-Serrano, D., & Fertl, P. (2016). Machine-type communications. In 5G mobile and wireless communications technology. Cambridge University Press, UK. Sachs, J., Andersson, L. A. A., Araújo, J., Curescu, C., Lundsjö, J., Rune, G., Steinbach, E., & Wikström, G. (2018). Adaptive 5G Low-Latency Communication for Tactile Internet Services. In Proceedings of the IEEE.

Chapter 4

Wireless Comms. Beyond 2020 William Webb

Introduction This chapter looks at the future for wireless communications. It is an area that has seen huge change and innovation over the last few decades and has become deeply embedded in all our lives. In this chapter we ask whether further change is likely and what this might mean for us all. The chapter starts by explaining all the different sorts of wireless communications and then selecting specific areas such as cellular and Wi-Fi and discussing why others, such as fixed wireless access, will not be considered further. It then looks at the reasons as to why wireless systems evolve, including to add new features, to add additional capacity through better efficiency and to deliver cost savings. It broadly concludes that we are reaching limits in each of these areas, but there are some grounds for optimism. The chapter then discusses one of the growing concerns over wireless – a lack of ubiquitous connectivity – and looks at ways this might be resolved. Finally, the chapter considers whether different models for wireless provision, such as single national networks, might be more appropriate in the future and whether we are reaching a stable utility-like provision of wireless or just pausing for breath before the next expansion.

Defining Wireless There is a huge range of wireless communications including: • Cellular: 2-, 3- and 4G mobile devices, very shortly to be augmented by 5G • Wi-Fi which constantly evolves to add new features, speeds and capabilities W. Webb (*) Weightless SIG, Cambridge, UK © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_4

49

50

W. Webb

• Bluetooth, used for personal areas networks and connecting all our ‘stuff’ together • Internet of Things (IoT) solutions embracing all above plus numerous new forms • TV and radio broadcast and datacomms, from terrestrial towers and satellites • Fixed wireless access (FWA) providing alternative links to home and office • Satellites providing a wide range of communications, sensing and imaging • Specialist solutions, used, for example, by the emergency services The typical standards used for some of these are shown below. Type Cellular Wi-Fi Bluetooth IoT TV FWA

Legacy standards 2G, 3G 802.11a and many others Versions 1–3 None Analogue WiMax

Current standards 4G/LTE 802.11ac Version 4 NB-IoT, LTE-M, LoRa, Weightless DVB-T and similar 4G/LTE

Future standards 5G 802.11ax, WiGig Version 5 Unclear Unclear 5G mmWave

Sometimes the division between these technologies blurs, or they are simultaneously used in combination. In this chapter there is not the space to consider every aspect of every known technology, and so we focus on the most important – those deployed in the largest volume and most critical to the way we live and work today. These are cellular, Wi-Fi, IoT and satellite. It should be noted that we are not discussing the remainder because: • Bluetooth is unlikely to change materially, and without a network element, the drivers for change are much reduced. • TV and radio only provides one-way communications, and its subscriber base is steadily being eroded by IPTV and by entities like Netflix suggesting that wireless broadcast will play a rapidly decreasing role into the future. • Fixed wireless access (FWA) has been tried and has failed many times in the past and is likely to do so again. Even if it succeeds, it simply acts as a short-term substitute for fibre to the home and does not provide new service opportunities. • Specialist solutions tend to be relatively small scale and transitory. So, when discussing wireless, we refer to cellular, Wi-Fi, IoT solutions and some satellite systems.

Reasons for Wireless Evolution In looking to how wireless communications will change in the next decade and more, we need to understand the who, what and why of change. We are used to ever-­ better wireless, but do not expect to see much advance in areas such as provision of electricity or water. Broadly, there are three drivers for wireless evolution: • To provide faster data rates and more features and capabilities such as IoT connectivity

4  Wireless Comms. Beyond 2020

51

60

Exabytes/ month

50 40

2014 2015

30

2016 20

2017

10 0 2012

2014

2016

2018

2020

Fig. 4.1  Historic growth in user data and industry predictions. (Taken from “The 5G Myth”)

• Improved spectral efficiency to enabling more capacity to be squeezed out of systems • Cost reduction sees operators offering more data for no extra cost or even reduced prices These are a reflection of the perceived needs of the end customers who exhibit avid appetites for ever-faster data rates and of ever-greater volumes of data. For nearly a decade now, data usage has grown by over 50% a year,1 a massive increase of around 100-fold since the introduction of the iPhone in 2007. This is shown in Fig. 4.1, which also sets out industry predictions to 2020. And yet subscribers have been unwilling to pay any more and demand operators massively reduce the price per GByte of data. These same general trends are expected to continue indefinitely,2 and new systems such as 5G have been designed to deliver ever-faster, ever-cheaper. However, few trends continue forever, and some are questioning the above premise. They argue that rates are already so high that they are no longer the limiting factor for any applications on the handset and higher data rates would not make any material difference unless new applications emerge that benefit from them. Data volumes are also reaching a point where the highest volume users cannot realistically consume any more video and unless some other, more data-intense, forms of consumption emerge may see a slowing rate of growth and eventually a plateau. Whether we are just a few years from this point, or still many decades, is a matter of intense debate and of critical importance to the future of wireless. We set out in

 For the most widely used discussion of data rate growth and predictions of the future, see the Cisco Virtual Networking Index at https://www.cisco.com/c/en/us/solutions/service-provider/ visual-networking-index-vni/index.html 2  The GSMA provides a view on aspects such as APRUs, network expenditure and more at https:// www.gsma.com/mobileeconomy/ 1

52

W. Webb

120 100

Exabytes/ month

80 60 40 20 0 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 2027 Fig. 4.2  A prediction of plateauing growth. (Taken from “The 5G Myth”)

“The 5G Myth”,3 an argument that it is more likely that a plateau will be reached in the next decade than beyond, as shown in Fig. 4.2, but many disagree. It is also important to understand who drives technology drive and change. Predominantly, it is the equipment suppliers: • The cellular network equipment providers including Nokia, Ericsson and Huawei • The cellular handset providers such as Apple, Samsung and Sony • The Wi-Fi chipset providers such as Qualcomm and Broadcom Standard bodies such as 3GPP for cellular and IEEE for Wi-Fi are the primary vehicles of agreement and change. They aim to be responsive to the predicted needs of their customers while anticipating new generations of technology that will lead to increased equipment sale as devices and networks are upgraded. To return to an earlier point, why does wireless change so dramatically when other important products and services do not? Why, for example, has water not become dramatically cheaper or been delivered at ever higher pressure, with additional features such as pre-softening in areas of hard water? Broadly, this is because technical evolution in the sector is in stasis and wireless is the antithesis with rapidly evolving technologies. In short, other utilities and services do not enjoy the drive of a Moore’s Law.4 Contrast this with the advancement and acceleration of smart screens and AI that have directly improved the wireless environment. In direct contrast, water is constrained by expensive pipes and civil engineering on a massive scale where the economics have not materially changed over the millennia.

 William Webb, “The 5G Myth”, 2016, Amazon  Moore’s Law predicts that the number of transistors on a chip doubles about every 18 months, resulting in ever more capable devices. 3 4

4  Wireless Comms. Beyond 2020

53

Will this disparity continue? Will IT and wireless slow down? There has been decades of discussion and predictions that Moore’s Law is slowing, and this might now appear to be so.5 Further, other constraints such as battery capacity mean that not all the gains in chips can usefully be realised. Over the last few years, the advances in handsets have slowed, making it harder for manufacturers to differentiate new generations. It may well be that it will get progressively harder to realise gains in wireless, and it will gradually tend more towards other utilities, a topic we return to at the end of this chapter.

New Features There has been intense debate about what we might want to use wireless communications for and as a result what features we might need. The manifestation of this debate can be seen in the work on 5G which has also affected not only cellular but also Wi-Fi, IoT and, to some degree, satellite. The 5G community aimed to provide three broad (and new) features: 1. Enhanced mobile broadband, with far faster data rates and greater network capacity 2. IoT connectivity, enabling billions of “things” to be connected 3. Very low latency6 to enable new services such as remote surgery All are to be delivered by the continuing cellular evolution. The Wi-Fi community is similarly considering the same target arena, while satellite solutions are looking at increased speed and in some cases a limited form of IoT connectivity. Convincing demonstrations have already been engineered in labs and field trials. Some see these as solutions in search of problems and hard to justify in practice due to the implied increase in network spend. To others, they are a critical evolution needed to deliver an ever more connected society. These issues were debated at length in “The 5G Myth”; here we discuss what we might expect for each different technology.

Cellular We will see 5G deployed from around 2019 onwards. Bodies such as the GSMA predict a relatively slow pickup7 with perhaps only around 15% of handsets using 5G by 2025 but steady growth beyond this. Hence, by the middle of the next decade,  While there is debate, see sources such as https://www.technologyreview.com/s/601441/mooreslaw-is-dead-now-what/ and others that set out the need for new types of innovation if progress in chip performance is to continue. 6  Latency is the time taken to send a message into the network and get back a response. It is critical where there is rapid interactivity such as in some forms of gaming. 7  See https://www.gsma.com/mobileeconomy/ 5

54

W. Webb

we might expect Gbits/s speeds to be available, latency to fall to imperceptible levels and network capacity to grow by as much as 10x. What is less clear is how extensively these features will be available. At present, it looks like they might be mostly confined to urban areas where the demand for additional capacity forces the deployment of new solutions. If capacity demands grow in rural areas as well, it may be that 5G will be extended across much of the existing cellular footprint, perhaps by 2030. And if new services emerge that require 5G, and for which there is a willingness to pay, that would stimulate rapid extensive deployment. Beyond that, the future is very unclear. Many have said they do not expect a 6G, perhaps on the basis that it has proven hard to show the business case for 5G and perhaps because it feels that the room for improvement is diminishing.

Wi-Fi Wi-Fi continues to evolve in a more chaotic manner, with new enhancements and features appearing at frequent intervals. The core Wi-Fi solution will shortly enhance to 802.11ax with ever-faster data rates and with better support for congested areas where multiple Wi-Fi hotspots overlap. In addition, there is a new variant – WiGig – operating in the 60GHz frequency bands that delivers very high data rates albeit over ranges of only a few metres. Manufacturers at present are unclear as to whether there is a demand for this capability, but it could drive economies of scale and technology for other possible solutions using these frequency bands such as backhaul to trains. We can expect further growth in data rates, in capacity, in ease of deployment and in the ability to manage devices better as they move around the environment.

IoT Perhaps the biggest area of change in wireless connectivity might be “things”. At present, relatively few things are connected, but it has long been envisaged that there are strong advantages in connecting some 50 billion devices. Many of these will be industrial  – in factories, smart cities, smart metres, agriculture and much more. To date, relatively few of these have been connected, partly because of a lack of a suitable wireless technology. Such devices need to operate on batteries for up to 10 years, be extremely low cost, get excellent coverage even in basements and similar, but only need to send tiny amounts of data. Cellular technologies until recently have been optimised for almost completely opposite purposes and are not well suited. In the last few years, some well-designed solutions have started to appear. For cellular operators, technologies like narrowband IoT (NB-IoT) provide a carefully

4  Wireless Comms. Beyond 2020

55

tailored solution and can be deployed on existing base stations. For others, solutions such as LoRa and Weightless can be self-deployed in a similar manner to Wi-Fi, perhaps providing coverage in airports, farms and other discrete areas. Delivering low-cost and long battery life requires quite different design principles than most wireless systems. Costs can be kept low by: • Keeping processing power requirements low, allowing low-cost core processor devices • Keeping transmitter power low, enabling the RF components to be built into the chipset rather than needing discrete amplifiers and filters • Enabling large economies of scale through open standards and fewer, more general-­purpose technologies, which lead to very low chipset costs • Avoiding technologies where there are significant patents in place, resulting in high royalty payments • Adopting ultra-wideband modes based on direct digital signalling technologies with no analogue bands or channels Note, the first two of these also help save power. Power can also be saved by: • Having extended sleep modes, so that the devices only wake up infrequently • Minimising transmissions, for example, by avoiding random access processes as much as possible that require multiple back-and-forth messages and repetition where clashes occur • Scheduling in advance messages as far as possible, which works for devices which send messages periodically (e.g. every hour or once a day) • Designing the network so devices need to listen only for a short time to understand whether they have a message from the network or otherwise need to listen for more information • Avoiding the use of destination addresses which can often be larger in size than the message to be transmitted. This can be achieved by always sending information to the same destination • The employment of mesh nets to reduce communication distances for devices in proximity to each other such as seagoing containers, home, office and health-­ care appliances, vehicle-to-vehicle and retail goods in stores and warehouses When devices do connect, they may well not connect to directly the Internet – so the term IoT is rather a misnomer. Instead, they will connect directly to specific databases: a BMW car to the BMW dealer database, a smart metre to the electricity company’s database, etc. This makes sense logically; it also helps make the system more secure, for example, by preventing devices being used to mount denial of service attacks on other websites. Devices will generally not connect directly to other devices either. Instead, where a connection between devices makes sense, this will occur at the database level, with one database instructing another to communicate to the devices that attach to it. This allows more intelligence and more security to be embedded in the process and overcomes difficulties of one device knowing the address of another. Figure  4.3 shows the architecture of a typical standard  – Weightless – where the simplicity of the architecture and connectivity is clear.

56

W. Webb

Fig. 4.3  An example IoT architecture (Weightless)

There are still some issues to resolve, from cost to security, but these are likely to be overcome in the next few years so that by around 2020 we will start to see strong growth in IoT systems. By 2030, almost everything that can usefully be connected will be enhancing productivity, making life easier and generating new possibilities for products and services.

Satellite Satellite solutions play a small role in our communications environment, predominantly because data capacity is relatively low while latency is high. This means they can only be an accompaniment to terrestrial networks. While satellite capacity has grown dramatically with improving antenna technology which allows satellites to project hundreds of beams within their coverage area, data demand has also grown, keeping the role of satellites relatively minimal. So their role remains limited to: • • • •

Providing a backhaul to planes enabling cabin Wi-Fi services Providing cellular coverage in deeply rural areas Providing IoT connectivity to devices outside of their campus area Continuing to provide broadcast TV services efficiently

However, a number of companies have proposed launching hundreds or thousands of small (very low-cost) satellites into a low orbit. These would provide much higher capacity simply because there would be many more of them. But the difficulties are still extreme, the regulatory issues almost intractable, and basic problems such as getting a satellite signal indoors remain. On balance, they probably will not significantly expand their influence.

4  Wireless Comms. Beyond 2020

57

Heading Towards a Bit-Pipe? Where does this leave us with respect to new features? Wireless technologies essentially deliver bits of data – they are “bit-pipes”. Mobile operators and others strive to add services on top of these to differentiate themselves and to ideally capture more value, but predominantly services are delivered “over the top” (OTT) by companies like Google, Facebook and WhatsApp. Our wireless technologies will get faster, but it is unclear whether this is really needed. They will develop specialised bit-pipes for applications such as IoT. The search for new features, which in 5G has turned into a solution in search of a problem, does not look likely to succeed. Wireless will become a faster, bigger bit-pipe, but the innovation and change that makes a difference to our lives will occur in layers above the wireless connection.

Efficiency: Reaching Shannon’s Limit If evolution in wireless is not to be about new features, will it be more about better efficiency – delivering more bits for every MHz of spectrum? Moving from 2G to 3G resulted in a massive 20x improvement in efficiency. From 3G to 4G, the gain was about 2.5x  – much less but still useful. With data demands doubling every couple of years, all gains in efficiency are important. The much smaller gain from 3G to 4G shows how delivering more in each generation is getting progressively harder. Data transmission, whether wired or wireless, is constrained by a fundamental limit set out by Shannon’s Law. Once a system reaches this point, no further gains in capacity are possible. Wireless systems are very close; hence the difficulties in realising further gains. For basic 5G systems, no gains in efficiency are foreseen, and indeed the need for multiple different signalling channels for 4G and 5G might make matters worse. However, that is not the whole story. Shannon’s limit applies to a single “wired channel”. For a wireless connection, it would be the direct radio signal between transmitter and receiver. But if a separate transmission can be set up between the transmitter and each receiver, then Shannon’s limit can apply to each, rather than to the whole cell. This could be achieved by “beamforming”. At present, most cells are divided into three sectors. Within a sector, a transmission covers the entire sector. This is received by the mobile that it is intended for and also by all other mobiles in the sector. Hence, the sector capacity is the Shannon limit for that channel. But if, instead, the cell could form a narrow beam towards each subscriber, and if each beam was sufficiently narrow that it did not overlap with other beams, then the Shannon limit could apply to each beam. Ten beams per sector would result in a ten-fold increase in capacity. Figure 4.4 shows how successive generations of cellular have approached the limit and how, with multiple antennas, 4G has already exceeded this.

58

W. Webb

Fig. 4.4 Cellular and the Shannon limit. (Source: https://www.slideshare.net/JuanRebes/ path-to-5g-overview-mwc-2015-interdigital)

Beamforming is the plan for 5G. For it to work requires a number of assumptions: 1. Subscribers are distributed around the cell such that nonoverlapping beams can be formed. This is not generally the case in dense areas with clusters down the same road/apartments. 2. Beams can track users as they move and as obstructions suddenly appear, such as buses, in a way that delivers a reliable service. This is far from proven. 3. Antennas that can form beams are big, heavy and expensive and only work in higher frequency bands (e.g. above 2GHz); whether they can be economically and practically deployed on cell sites is unclear. Nevertheless, the basic concept is sound, and the challenges, while severe, will likely be resolved over time. So we might expect to see 5G capacity gains, perhaps not as high as ten-fold but perhaps similar to the 2.5x of 4G. That would be a goal worth striving for, especially in dense urban areas where congestion is most severe. Wi-Fi is also adopting a similar approach, along with solutions that allow Wi-Fi nodes to work better in very dense areas such as shopping centres and high-­ occupancy offices. These include ways to better coordinate systems. Hence, we can expect more capacity from our wireless solutions, both as systems get more efficient and as they utilise the additional spectrum being cleared and auctioned (e.g. the 3.4–3.8GHz band for cellular). But it is also clear that gains are getting ever-harder to find, and beyond 5G it is very difficult to envisage a route to higher efficiency. This may be another reason why 5G is thought to be the last of the cellular generations and why capacity growth in cellular networks cannot continue

4  Wireless Comms. Beyond 2020

59

indefinitely. Best estimates are that there will be sufficient capacity for growth until about 2027, but equally, it may be by this time that demand has stopped growing as users have become satiated with all the data they can consume.8 If systems do not get materially more efficient, or it costs a lot to achieve the efficiency, then the operators will not get cost savings. That looks to be the case with 5G – the extra capacity will be expensive – needing special antennas and complex site upgrades. This has cast doubt recently on the success of 5G and is an important reminder that ultimately deployment is driven by viable business cases and not fashion.

Often Overlooked The drive to ever more base stations and to ever more capacity per base station will impact the backhaul  – the connection from the cell back into the network. Traditionally this has been dominated by optical fibre and microwave point-to-­ point. However, as demands grow, fibre will increasingly predominate and become more widespread. It now appears likely that the convergence of fixed and mobile networks, plus a lot of pending M&A, will most likely realise a focus on the rebalancing of ­investment profiles to enable more and faster fibre roll out. This results in cells becoming smaller as they can readily connect to fibre, eventually leading to widespread self-­ deployment in homes and offices. Antenna wars is another topic seeing scant mention, and it is a non-trivial problem. The real estate in our mobile phones, tablets and laptops is very limited, while the number of required antennas tends to increase. Beyond 3-/4-/5G, Bluetooth and Wi-Fi, the need for another antenna for 60GHz and higher may be beyond our engineering abilities. For sure, the prospect of steerable arrays and all singing, all dancing, fractal antennas is daunting. While these all work perfectly in theory, their practical realisation has met with scant success to date.

Other Ways to Deliver More The ways that we connect are changing, and this may mean that we need different kinds of wireless connectivity. The key trends are for data volumes to be biased ever more towards indoors and for a demand for connectivity everywhere including on trains, planes and deep inside buildings. Data is moving indoors because by far the largest data volumes are associated with video. Video is easier to watch indoors – it is hard to watch video while walking down the street (although some try), and viewing conditions outdoors are often  See “The 5G Myth” for more details on this.

8

60

W. Webb

too bright, too noisy, too wet or otherwise unattractive. As the volume of video grows, so does the percentage of data consumed indoors. It is very hard to know quite what this percentage is; estimates often put it at greater than 70%. Both indoors and environments like planes and trains are best served with indoor (or on-train) cells. This is because it is hard for signals to penetrate a building from outside, and when they do, the loss of signal reduces the capacity available. Conversely, cells inside buildings provide excellent signal levels, and the fabric of the building can shield users from interference from other cells. The future then is clearly in-building small cells widely distributed across all buildings, with multiple cells in larger buildings. Some might argue that this future is already here. Wi-Fi is deployed in almost all buildings, in trains and in planes. All devices have Wi-Fi connectivity. Wi-Fi carries well over 70% of all traffic and is simple, cost-effective and has a strong upgrade path. Others suggest that Wi-Fi is not optimal because it does not integrate tightly with the cellular network and because the unlicensed nature of the frequencies used means that quality of service cannot be guaranteed. These criticisms are valid, but the difficulties in installing cellular solutions at the same level of density as Wi-Fi are so great; it is very hard to see how cellular could play a role in anything other than the largest of public buildings. It would be far easier to fix the problems with Wi-Fi than install something new. Our view is that Wi-Fi is now so deeply entrenched that it will be near impossible to displace it and completely unnecessary. Wi-Fi will evolve to overcome the problems mentioned above, and companies like Google are leading the way with ideas like Project-Fi.9 This fits well with a cellular network where capacity growth will eventually hit limits – Wi-Fi can offload much of the traffic enabling the cellular networks to stay within their capacity limits. This implies that the future of wireless is more about Wi-Fi than 5G. We will aim for a “Wi-Fi first” world where we only use cellular when Wi-Fi is unavailable (e.g. when in a car). Artificial intelligence and widespread data gathering will enable us to coordinate and optimise user-­ provided Wi-Fi nodes around the world. There are ideas for beyond Wi-Fi, such as Li-Fi. Li-Fi uses light rather than radio. Special LED light bulbs in a building are modulated by the building network and devices receive and demodulate these signals. The advantages are a huge capacity increase due to the enormous bandwidth available in visible light. But the costs of adding a backhaul Ethernet or fibre connection to every bulb in a building are substantial, and it is hard to see why it would be needed in most cases. Hence, while options exist for beyond Wi-Fi, in practice Wi-Fi is likely to evolve so that we do not need them. If we do move to a Wi-Fi first world, or some of these other self-deployed technologies come to the fore, then the structure of networks will change. At present large mobile network operators (MNOs) deploy nationwide cellular solutions. In  Project-Fi offers subscribers a cellular-like service. Google steer traffic towards Wi-Fi hotspots as far as possible, with fallback to two different US operators (Sprint and T-Mobile) where there is no Wi-Fi coverage. As a result, subscriptions can be cheaper and coverage and quality better. At present, this is only available in the United States. 9

4  Wireless Comms. Beyond 2020

61

the future we might tend towards a world where most network resources are self-­ deployed by building owners, by companies, by local authorities, or others, and we gain access through an aggregator whose function is to provide seamless service from a patchwork of self-deployed solutions. Cellular might then be a fallback when there is no other coverage. This would lead to a very different structure, where MNOs might no longer be retail operations but instead provide wholesale roaming capabilities to the aggregators. As a result, they would be much smaller and less well-known. All sort of aggregators and value-added entities could be imagined along with various revenue-sharing opportunities for the owners of the Wi-Fi routers and similar. This seems more likely than any dramatic technology change, suggesting it will be the value chain and the existing players that change this time round (compared to previous generations of technology where the players have stayed the same but the technology has been enhanced).

More Radical Innovative ideas continue to proliferate. Google are sending balloons floating across Africa to provide coverage to unserved areas. Facebook has projects for ultrahigh capacity urban base stations working at 60GHz,10 and people like Elon Musk are looking to launch thousands of satellites. Phones might become software-defined devices making them more flexible and upgradable. Might any of these revolutionise wireless in the coming decades? Probably not. As discussed, these ideas are typically not needed or uneconomic. For example: • Balloons across Africa (Project Loon11) have proven very hard to steer to locations where they are useful, instead low-cost Wi-Fi deployments and simple cellular networks are extending coverage. • High-capacity outdoor cells in cities are of little use when most of the traffic is indoors. • Dense satellite networks also struggle to get signals indoors, again where they are most needed. • The idea of fully digital or software-defined radios has been debated for nearly 20 years now. These are devices where all of the radio decoding is performed in software making them highly flexible, able to have new protocols downloaded or adapt to the local environment. However, they inevitably end up as more expensive than devices designed for specific tasks (e.g. 4G reception across a set of defined frequency bands), and since most devices are replaced every few years, the flexibility of being able to change is not particularly advantageous. So while software-defined radios are viable, and while they bring benefits, it seems likely that the costs will outweigh these benefits for the foreseeable future. Only in  See https://code.facebook.com/posts/1072680049445290/introducing-facebook-s-new-terrestrialconnectivity-systems-terragraph-and-project-aries/ 11  See https://x.company/loon/ 10

62

W. Webb

some special cases such as military radio, where costs are less of an issue and the service life is much longer, is this likely not to be true. Having said that, the boundary between the hardware and the software is likely to continue to move towards software over time as the penalties for nonoptimal hardware fall. Indeed, nothing particularly radical has changed wireless communications for decades. Instead, we have seen steady evolution of cellular and Wi-Fi. Perhaps the most radical change will be networks to address IoT connectivity.

Money Makes the World Go Round Ultimately, what transpires in wireless communications will be heavily influence by money. Mobile operators will only deploy new technology if they can make a profit from it. When we deploy Wi-Fi, we may not undertake a business case, but we do seek value for money. One of the reasons IoT connectivity has not grown strongly so far is because it has been too expensive. Despite all the innovations, the growth in data usage, the new OTT apps and the ever-better phones, we typically do not pay more for cellular connectivity year-on-­ year. Monthly expenditure, or average revenue per user (ARPU), has been flat for a decade and is predicted, if anything, to decline. When MNOs introduced 4G, they briefly tried to charge more, but few subscribers were prepared to pay, and so the tariffs were dropped back down again. We also increasingly expect Wi-Fi connectivity to be free and rarely pay for it, expecting that we will be able to find a free alternative if we are charged. We pay less for broadcast as some “cut the cord” and pay only for what they consume or for a relatively narrow selection of material. We are also replacing our phones less often as this feature driven element appears to be approaching a design stasis. There will be new revenue from IoT, but this looks to be small. It is possible to buy an NB-IoT SIM card for around $10 that provides 10 years’ worth of IoT data. That equates to about $0.08/month compared to typical current subscriber levels of around $30/month. Even with about ten times as many devices as people, that is still less than the equivalent of $1/month increase. With little new money, MNOs are inclined to minimise expenditure and have been reducing network investment over the years. That makes it unlikely that there will be dramatic new investment or network change. Indeed, it casts doubt on extensive 5G deployment or at least suggests it will be slow as operators stay within current capital expenditure limits. The current economic model appears to be heading towards commoditised solutions where the lowest price is the most important factor. We noted early on that wireless communications was radically different from utility industries like water where little changed. But as wireless tends more towards a bit-pipe of data, for which we pay a static fee, the parallels with utilities start to become stronger. Wireless has a long way to go before it becomes as boring as water but is headed in that direction. The US administration recently suggested that

4  Wireless Comms. Beyond 2020

63

government might deploy a single nationwide 5G network. While the idea was quickly withdrawn, it was not completely crazy and will likely resurface at some point in the next decade. This might be especially so if the idea of aggregators, discussed earlier, emerges, weakening the position of the current mobile operators and making competition between them less important. It seems likely that the number of mobile operators in many countries will fall, typically through merger,12 and that in some cases, or parts of some countries, a single network might emerge.

There Is Plenty of Room for Exciting New Things That is not to say that the future for wireless is boring. AI, big data, augmented reality (AR), IoT-based services and more can result in a host of new and enhanced applications. Phones will become excellent personal assistants able to predict our needs. Language translation will become near perfect in real time. Phones will help keep us healthy and alert us when medical attention is needed. But none of this is driven by better wireless – it is down to data analytics, covered in other chapters in this book. Wireless is needed to transport the data and will do so superbly. Indeed, the fact that we predict relatively little change in wireless is not because of some problem preventing evolution, it is because wireless already delivers everything we need. That is a good place to be.

Predictions In summary we predict that: • 5G will be introduced in urban areas, where it will deliver valuable capacity increases but no net user benefit. • Wi-Fi will cover all indoor areas, providing ever-better service. • IoT networks will develop and be widely deployed, and billions of things will be connected. • There will be no 6G, or at least if there is one, it will be a relatively minor upgrade rather than wholesale technological change. • Innovative ID wireless systems will emerge but will find it hard to make a business case. • Wireless will soon do everything we need – a good news story. • Innovation will happen, but it will be in OTT, AI, big data and other computing and processing solutions rather than in wireless. • The value chain and industry structure will change, and in particular the number of operators and their role will decline.

12

 At the time of writing Sprint and T-Mobile in the United States were in merger discussions.

64

W. Webb

• The demand for more wireless will see ever more pressure for optical fibre to the node (FTTP) to provide the necessary bandwidth to every building. • In a wireless future dominated by low-power, low-cost, very short-distance communications, mesh nets and dense FTTP are not only a need they are a natural outcome. • Far simpler wireless such as UWB variants will also be a necessary outcome along with the notion of bands and channels at >30GHz.

Chapter 5

Digital Transformation Paul Graham

A goal of the next industrial revolution (4.0) is the creation and use of technologies able to perform tasks that are beyond human ability. This is central to the creation of many new industries and the longevity of many that are long established and is only possible with total digital transformation. An extreme example of this is the introducton of self-drive vehicles which will require digitalization that not only involves the car but the transport environment. The challenge for self-drive car is that a typical urban landscape is very hostile and highly unpredictable. One (very expensive) solution is to adapt the environment with the introduction of sensors in the landscape which are available to the decision-making processes in the car. An alternative is to develop a hive mind between all the self-drive vehicles, so they share information and experiences to facilitate future critical decision-making and preventing collisions.

An Empowering IoT This will see direct device-to-device and device-to-Internet connections beyond the PC, smartphones and personal computation devices. Machines, appliances, devices, components, vehicles, containers and pallets are a small subset of what we might see on the IoT which will typically perform singularly simple tasks, such as sensing and providing performance and ownership data back to central and/or distributed systems. The essential power of IoT is down to the number of devices connected to perform basic tasks including the collection of data from numerous sources for wider analysis and application.

P. Graham (*) Matrixx Software, Saratoga, CA, USA e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_5

65

66

P. Graham

Healthcare Transformation People are gradually subsuming more responsibility for their own health by purchasing and wearing monitoring devices to track their activity levels and vital signs such as heart rate, respiration, blood glucose, etc. which can be stored and processed for later use by a doctor or specialist. The possibilities here are obvious: to constantly monitor patients for any unusual trends or anomalies. Data containing the data of millions of people has not yet been collected on a real-time basis and processed, nor have the hospital records of complete hospitals been digitized and analysed. But what is happening is the aggregation of data from an increasing population on new and sophisticated scanner and robotic and augmented reality surgeons. Human surgeons may perform 20–100 k operations in a lifetime, but that experience can be gathered by a large population of robots in a day. And if they are all networked, then they all gain from that sharing in a way humans cannot.

Digital Transformation Network Billing In the early days of telecommunication, charging was based on a record made by the switch board operator. The next advancement was metered calls; this would still result in a bill for usage, but the charges on individual calls would be unknown. Electronic switching for voice calls enabled more detailed records to be made, and the concept of an itemized bill was introduced. But in terms of the experience of the user, the person would get a bill at the end of the month, and the payment would be collected. But none of this fits the need of modern world where distance, time and

5  Digital Transformation

67

bandwidth are now increasingly irrelevant factors, and the focus is on devices provided and their quality. Modern billing systems significantly change the experience and opportunity space for the provider and the consumer. Operators can retire legacy systems and services to reduce operating costs and improve services and efficiency to give consumers a better overall experience whilst receiving usage and payment feedback in real time. Smaller payments and the sense of trust encourage consumers to use and demand more. Ultimately billing systems hang between two extreme corner posts: flat fee and micropayments! And where this is ultimately going is still being played out by rapidly evolving technologies and other influencers such as over-the-top providers/services/applications. Total digitization also results in real-time analytics which can be used to tune the business, network and services and provide feedback to the consumer. Telcos have traditional sold products which rely on static plans and relied on “breakage” to boost revenues. Breakage is the term used where a consumer uses a plan out of the bundle, so, for example, their monthly plan may include 2G of data; however the consumer may exceed this and pay for data out of plan at an inflated rate. This can cause the consumer to be more cautious with their usage and untrusting. Real-time analytics can be used to predict usage and offer the consumer bundles of data before they exceed the quotas.

5G The telecoms industry is now introducing 5G technology – a move from one generation of network technology powered by:

68

• • • • •

P. Graham

Core network with radio stations and network Mobility including HLR/HSS, voice switching, messaging and data OSS BSS Consumer-purchased devices

The development of 5G commenced with the radio network and an increase in data rates which in turn necessitated the building of new networks designed to deliver a large number of new and varied/dynamic services across a diverse set of devices from high-definition television to the IoT. This advanced wireless technology is divided into three main service types: enhanced mobile broadband (eMBB), ultrareliable and low-latency communications/mission critical control and massive machine-type communications (mMTC) or massive IoT. Enhanced mobile broadband (eMBB) is to provide for data-intensive applications that are bandwidth hungry like gaming and video streaming the same level of service as provided by fixed-line broadband services. Ultrareliable and low-latency communications (uRLLC) or mission critical control is for services that require low latency, extremely high reliability, availability and security, such as autonomous driving and Internet applications that provide feedback to the user. Specific standards are being developed to exploit this service such as cellular vehicle-to-everything (C-V2X) and real-time command and control. Massive machine-type communications (mMTC) or massive IoT for low-cost, low-­ energy devices with small data volumes on a mass scale, such as smart cities, Narrowband IoT will be enhanced with further capabilities such as voice, location and low latency. The increased bandwidth and lower latency offered by 5G will permit the processing of masses of data in the cloud for almost instantaneous deployment and use. Examples are augmented reality services, gaming, medical diagnostics, industrial VR and control, etc. – all under customer control. This all creates opportunities for the transformation of businesses. 5G introduces new requirements that also apply to digital transformation: Real-time provisioning and service assurance. Order management should no longer take hours or days but should be completed in seconds by the customer and not the provider. Multi-tenant and cloud-based online charging, something that is critical for 5G slicing. Multi-domain and multi-partner improvements to BSS. Real-time orchestration which must be highly scalable and distributed. QoS management in real-time with AI-driven service assurance. End-to-end management of hybrid virtual and physical networks across access, aggregation and core infrastructure including all previous generations (4G, 3G, Wi-Fi) and fixed networks. 5G infrastructure as a service will eventually cover the entire network offering dynamic network and service capabilities with new and novel applications.

5  Digital Transformation

69

Network Slicing A fundamental concept of 5G is a clear separation of the control plane and communication planes which allows network slicing/sharing or multi-tenancy which in turn demands real-time charging and usage monitoring. Network monitoring tools already determine resource usage with a high-level view but are unable to provide the granularity for individual services and segments. A real-time charging system that has a view of segments and individual services would be able to provide the necessary analytics to adapt resources provided to network slices and guarantee the quality of service. AI and machine learning can continuously examine to provide a solution here, but this will also put increased pressure on network resources as more people choose to create and use their own services and micro-services on demand. IoT devices represent another extreme in the spectrum of potential 5G services with applications such as remote monitoring and utility metering requiring a reduced level of service through to vehicle telemetry and communications, mobile health professional diagnostics and conferencing. So, some services will need a guaranteed level of service the whole time, whilst others will be sporadic with scant needs.

Energy The first step in the digital transformation of energy metering is the introduction of smart meters which can send back usage to the supplier. It also permits consumers to view their usage and see how they are charged. It should be possible to adjust user demand to reduce the peak-to-mean ratio of the demand curve experienced by the grid. For example, appliances such as freezers and fridges could turn off at peak demand, whilst electric car ownership and local energy storage means grid smoothing can also be facilitated.

70

P. Graham

Customer Care • An example of customer care digital transformation can be found with the Renault group in the auto industry. To maintain customer engagement, they employ a fully connected car experience which supplies telemetry and data back to the customers smartphone app. This vehicle connection is continuous for the life of the vehicle and incorporates business analytics, IoT, artificial intelligence, mobile networking and cloud computing. • This allows owners to keep their completed maintenance logs via their smartphone app. It also facilitates them to book new appointments by tracking the next scheduled service, and an integrated dealer locator makes it simpler to find the closest dealership when traveling from home. There is also a remote dashboard that alerts for safety and efficiency issues, which provides alerts to the driver for important issues, such as low tyre pressure, and can also assist with scheduling a dealer visit, which again increases customer engagement and can also potentially increase revenue. • This digital transformation is also planned for Renault to use this expanding database of vehicle information to target additional higher profit aftersales and more connected services. It provides an excellent foundation for even more growth, and opportunities for a better social media presence, whilst the potential for customer retention via more frequent and higher-quality engagement.

5  Digital Transformation

71

• Electric cars have made it increasingly important to have a fully connected experience. As well as being able to have full control over charging, the manufactures are able to monitor all aspects including battery health. This serves two purposes: it provides the customer with feedback regarding the status of the vehicle and battery, and it also provides invaluable data back to the manufacturer to improve the development of electronic cars which are still being understood. The industry is still at an early stage of development so requires information regarding battery performance and vehicle aging to further the understanding of this new field.

Consumer Experience Company workers/staff are now monitored for movement, speech patterns and digital communication. Although this can be used for security, training and malevolent reasons, its primary use is to improve communication and eliminate bottlenecks caused by individuals. Workers stress levels can be monitored, so patterns of duress in the work place can also be identified and eliminated. On another level, video streaming services such as Netflix have complete consumer monitoring feedback loops which determine viewing behaviours and preferences of the kind of programs that they should commission. Users are segmented by preferences such as crime, science fiction, documentaries, etc. Viewing habits determine the popular plotlines, characters, genre and length of episodes which is fed back into the scripts of future productions. It really is about pleasing the customer and keeping them contented and happy! In contrast customer care has been a neglected aspect of the telco/netco experience. In this industry it all started with the switch board operators interacting directly with the customers, which was rapidly negated by the earliest of digital transformations and to be further degraded by the introduction of mobile devices and services. Over a period of time, the pendulum moved from less care to contact centres designed to provide a more tailored experience for the customer with human and then a mix of machine and human operatives. Customer relationship management also became more refined with closer relationship between customer and agent.

72

P. Graham

For the technophiles this has now started to migrate to self-help, user groups and DIY solutions with contact between customer and provider being a last resort! This story/evolution is far from over as AI is now moving into the field and proving very effective! An indication of where this is all going can be glimpsed by looking at the services and digital experiences provided by Amazon, eBay, Apple, Facebook, Netflix, etc.

Customer Engagement The key is to engage with the consumer in a manner that is compulsive, but not annoying or intrusive, and ideally initiated by the user. People want to be informed, and they want good service; indeed they are prepared to pay for it  – this is the Amazon experience. Looking for products to purchase can take two approaches. There are products that are similar to what already exist or the new and never tried before. Maintaining a history of use, preferences, behaviours and suggestions is necessary. It maybe that a certain sales strategy works over and over, or it may be time to try something different. Ultimately, AI can be used to create the products based on the successes on the basis of a global or selected population whilst filtering out all previous failed or less likely to succeed approaches.

5  Digital Transformation

73

General Digital Transformation From the study of telecoms and the utilities industry, we can conclude that there’s a familiar pattern to how legacy systems operated and the components of an ideal system that has undergone digital transformation. A traditional system contains: • • • • • • • •

The service Usage recording or metering Toll tickets to billing Bill construction and rating Invoicing Payments and recovery of debt The catalogue of products Customer care

The key question is: is it possible to create a generic digital transformation model that can be applied across many industries? We can identify several elements which seem to exist in different industries; these are: • • • • • •

The consumer The service that the consumer is using The catalogue that displays the available products The charging system for payments Digital billing for invoicing and debt collecting The digital channel for self-care

The industries that we have looked at so far have taken legacy processes and adapted them with modern technology. A prime example is healthcare which has traditionally been sick care – it is all curative with almost no preventative measures. There is no attempt to predict when somebody becomes sick or track their health to predict when they will become sick or suggest life style corrections to prevent a potential sickness or disease occurring. Care for the average sees crude methods such as measuring a person’s weight or basic fitness but nothing that collects health data and activity to provide feedback to improve life style to prevent ill health for the individual. Technology is slowly enabling a better healthcare experience with the introduction of digital medical records and monitoring devices to wear and in the home. Robotic surgery has transformed some aspects of operations. 3D printing technology has been developed to enable the printing of human organs and other body parts. However, the digital user experience is still sadly lacking. • Digital transformation can help improve the survival rates of patients, and provide the ability to monitor them and provide feedback to the health care workers. It has the ability to provide a knowledge base that is spread over the globe. Patients symptoms could be managed against a database looking for similar symptoms and able to consult with a highly specialized physician who has dealt

74

P. Graham

with similar symptoms. This can be made possible with telehealth and technology that connects all the sources and physicians. • AI would eliminate the need to consult with the physicians and can be used to hunt through all the health data looking for the symptoms and a diagnosis. • By placing priority on the user experience, this type of technology will change the care base that is available for all patients, not just patients who can afford it. The implementation of fast-moving collaboration technology and AI may also open new doors for smaller facilities to get the care their patients’ need fast. It will increase workforce productivity and decrease costs for the industry. • The importance of the digital transformation and the digital user experience in healthcare was nicely summed up by the International Data Corporation (IDC): –– “The emerging consumerist market character points toward the need for new service delivery models and product offerings, powered by digital capabilities”, IDC said. “Furthermore, future challenges such as rising levels of obesity, patients living with one or more chronic conditions, an aging population, and a widely unchecked healthcare spend are emphasizing the need for U.S. healthcare organizations to think digital and get digital done”. –– The five steps towards the digital user experience should be top priority to achieve the healthcare your patients dream about. From expanding your care footprint, offering telehealth services to patients, using data to learn and ­delivering the omni-experience, your facility will be offering the best of the best in patient experience.

Health The current generation of health wearables take very crude measurements such as activity and unreliable heart rate readings with the goal to provide motivation rather than to provide accurate health feedback to the user. Mass heart rate readings covering millions of people and the feedback using AI and machine learning are still an opportunity which can be exploited. People’s day-to-day health needs have never been digitized. Computers and software have not yet been used as a means to monitor and provide preventative advice to users, so the digital transformation process in health will be a leap straight to full digitization, rather than the progressive use of technological advancements as have been seen in the telecommunications world.

5  Digital Transformation

75

Blockchain Beyond IoT and dynamic billing, blockchain is another key enabler. As a highly secure distributed ledger of records connected by massive peer-to-peer networks, it is ideally placed to satisfy the needs of many industries including healthcare. Cryptographic hashing is used to ensure the blockchain is immutable and cannot be modified and smart contracts can be used to confirm each transaction that is stored using digital signatures and public/private key cryptography. There is no centralized authority, so the records can be made public and shared with all parties across the global telecoms network. Blockchain is also ideally suited to future billing, charging and all forms of transaction for telcos, netcos, mobile operators and many more. It is also ideal for identity ownership and retaining subscriber’s details and their management, but it is all something that a telco would be very reluctant to surrender. Not so the new industries that pose a threat to this sector or the dynamic market leaders! It is possible to have identity blocks beyond mobile numbers that can be used to move all of a subscriber’s details between operators. This supports churn and number portability. So when the subscriber initiates the port process, the nodes in the blockchain peer-to-peer network will firstly check that the subscriber owns the number; it will ask permission from the number ported from network, which can check the ownership of the number in the blockchain. Once the nodes have checked the ownership, it can then create a transaction which ports the number to the chosen new network. This transaction will be recorded in the blockchain and made immutable due to the checking and cryptographic hashing. The blockchain forms the centralized number portation database which can be used by each network to construct their routing tables for requests to subscribers in other networks. Each mobile network operator will own nodes in the peer-to-peer network and so will get notification of the port numbers in the form of the latest

76

P. Graham

block in the blockchain. The information can be fed into the order management and the subscriber provided with an eSIM which is loaded into the phone using the mobile app. Having a system like this in place enables the goal of rapid movement of subscribers between available networks. It can be extended to include all forms of identity as well as payment methods. It is also applicable to those elements of the IoT reliant and mobile networks.

Summary This chapter has discussed digital transformation and application in telecommunications systems with some lessons applied to other industries such as healthcare. 5G and the concept of slicing will mean that the infrastructure itself will be presented as a service via APIs. It will be necessary to have real-time OCS that serves a number of functions: to offer a virtual dynamic resource charging for the services and to monitor 5G usage and infrastructure. We have looked in some detail at the prospect of blockchain introducing mass disruption through decentralization and outsourcing with the conclusion that traditional business models will be affected and telecoms and wider business will have to look at new approaches. Similarly, the embracing of AI and machine learning will remove human interaction from the system and constitutes a primary aspect of the fourth industrial revolution.

Chapter 6

Big Data, Small Data, and Getting Products Right First Time Human Ramezani and Andre Luckow

Introduction Data in its various shapes is the foundation of Industry 4.0 and has become a critical component for many aspects of advanced manufacturing (Hellinger and Stumpf 2013; Bruner 2013; Evans and Annunziate 2012). The term Industry 4.0 encompasses a broad set of technological, organizational, and societal changes along the entire value chain of industrial corporations. Industry 4.0 promises to shorten development cycles and improve flexibility and the ability to customize products while benefiting from higher efficiencies (Lasi et al. 2014). In the following we focus on data-related aspects. With the increasing deployment of connected machines in industrial environments, the opportunities for utilizing data in conjunction with ubiquitously available data storage and processing capabilities and novel advanced techniques, such as machine learning, are immense. The result will be the transformation (and potential disruption) of existing industries and the creation of new industries and business models. The following technological trends drive the advances in smart manufacturing: (i) Pervasive and affordable sensors and low-cost computers enable data collection at unprecedented scale. (ii) Scale-out storage on premise and in the cloud provides cost-efficient storage for this Internet-of-Things (IoT) data. (iii) Flexible on-demand compute resources in the cloud enable us to extract insight from data.

H. Ramezani (*) Den Haag, Netherlands A. Luckow Munich, Bavaria, Germany © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_6

77

78

H. Ramezani and A. Luckow

(iv) Machines (e.g., robots) are increasingly automated and can act autonomously on real-time data feeds and insights. During the manufacturing process, data arises at different levels and granularity: Internet-of-Things sensors, e.g., vibration, temperature, and camera-based sensors, and at higher levels process-related data, such as work orders and part movements. Big data does not mean the right data for a specific use case – in many cases it is instrumental to align and integrate big, high-fidelity sensor data and small process data. This is the prerequisite for interpretation of the data and the usage of machine learning, an important tool to identify patterns, make predictions, and derive prescriptive measures. The goal of this chapter is to develop a common understanding of characteristics of big and small data, technologies, and applications to drive modern manufacturing processes with the objective to support the creation of highly customized and personalized products. A main target  of modern manufacturing is to produce such products in the right quality the first time. Organizations that utilize big and small data will be able to generate valuable insights for their business to achieve this objective. But, in order to achieve optimal outcomes, various challenges need to be addressed: data is a challenge to collect and manage these appropriately and find the precise drivers, the right interdependencies, and the correct conclusions. Old paths of interpretation need to be abandoned in order to find new, unforeseen, and beneficial opportunities. This chapter describes both technologies and processes as well as their relationship. Particularly, we focus on data-related aspects and refer to the following: (i) Big data as the usage of large volumes of data in conjunction with statistical and machine learning techniques. (ii) Small data as a highly curated, actionable data. Further, small data often refers to a specific object under investigation. While small data is typically collected using rigorous scientific methods, such as control group studies, surveys etc., big data is often generated by sensors and the data exhaust of existing systems. Techniques for curating actionable small datasets out of big data are commonly referred to as data science (Dhar 2013) and include, e.g., different statistical methods and machine learning. Table 6.1 summarizes the technologies discussed in this paper with respect to their impact on manufacturing processes. We will explain the usage of different kinds of data and machine learning techniques to support different aspects of manufacturing, such as assembly, quality management, logistics, maintenance, and supply chain management. The chapter is structured as follows: In section “Small Data in Manufacturing” we start with “small data” in manufacturing investigating RFID, ergonomics, and augmented reality. In the next section, we discuss the usage of AI-based computer vision techniques for various processes, e.g., quality management and logistics. We continue with a discussion of natural language processing, present application of advanced printing solutions, and introduce finally blockchain technology in section “Blockchain”.

6  Big Data, Small Data, and Getting Products Right First Time

79

Table 6.1  Technologies and their impact on manufacturing processes

Technology RFID Augmented/ virtual reality AI computer vision Natural language Advanced printing Blockchain

Description Radio frequency-based smart tags enable the detailed tracking of assets Novel immersive user experience technologies that enable to visualization of complex datasets, models, etc. Usage of machine learning techniques (e.g., deep learning) to detect complex patterns and conduct advanced computer vision tasks Machine learning-based capabilities for extracting knowledge from textual data and creating new conversational user interfaces Additive techniques for creating physical objects from digital models Blockchains or distributed ledger technologies are decentral, distributed systems for maintaining and sharing data in a trustless environments

Application/ process Logistics

Big vs. small data Small

Design and engineering, assembly Quality, logistics

Small

Assembly, aftersales

Big and small

Component Production Logistics

Small

Big and small

Small

“Small Data” in Manufacturing The development in manufacturing has already led to highly efficient processes, e.g., in quality and supply chain management. One example is the concept of Just-­ In-­Time (JIT) manufacturing that was first introduced by TOYOTA in Japan (Ohno 1988). This enables effective production and minimizes the storage of required material. Warehousing has always an impact on costs, complexity, and efficiency. Getting third-party parts exactly at the time of need helps to build cost-effective production cycles but also has some very important prerequisites. Since the incoming inspection of the delivery is almost relinquished, the quality assurance needs to be ensured through on-site audits at supplier sites.

RFID With the improvements in the field of radio frequency identification, the sensors become smaller, cheaper, and less sensitive to environmental influences; RFID tags in many different shapes/costs are used in modern smart manufacturing plants. Thus, RFID tags became more and more popular for tracking assets, such as parts, and support processes, such as inventory management, goods receipts, etc. Another

80

H. Ramezani and A. Luckow

use case was tracking of special containers. These containers are used for specific tasks and usually quite expensive. Therefore, these are rented and should be utilized efficiently. With the use of RFID tags, the status, location, and utilization could be planned much more effectively and enabled a plan vs. actual comparison which resulted in a much better overall planning with reduced costs. All these use cases were far away from big data applications as we know today. A step from small toward big data was taken when a better monitoring of parts list was investigated. Complex products consist of a huge variety of parts. These might have different status as well as being delivered from different suppliers. It is beneficial to the manufacturer to have a complete and up-to-date view of the product, not only during production but also during regular life cycle. This would not only support maintenance and warranty aspects, but it would also help with fraud prevention. At any given point in time, the precise component and their history/status could be derived using an onboard parts list. RFID enables manufacturer to monitor parts, goods, equipment, and the supply chain in real time at an unprecedented accuracy. To utilize its full potentials, RFID must be tightly integrated with an Internet-of-Things platform (IoT platform) and advanced analytics capabilities. Further, this IoT platform will be responsible for managing further sensor data, e.g., from electrical machine sensors, cameras, microphones, etc.

Ergonomics Ergonomics is an important aspect in developing products. Digital methods are widely used to improve product design from an ergonomics point of view. RAMSIS (Seidl 1997) is 3D tool to design vehicle cockpits in an ergonomic way using a realistic human model based on international data which is derived through an anthropometrical evaluation. Aspects like reachability, visibility, and comfort can be investigated and assessed for different body types and percentiles. Subsequently, cockpits can be designed regarding sex, age, or national requirements and also consider the change in size based of the planned release of the product in a market. Besides the workstation-based system, the ergonomic design can also be developed and verified in a virtual reality (VR) environment. Furthermore, this application is not limited to the product alone. Production-­ related topics like assembly line design, employee stress, and other health-related aspects could be investigated thoroughly involving all relevant parties (e.g., specialist in occupational medicine) together to fulfill labor and occupational health and safety legislation. One important aspect of this methodology is using the experience of highly skilled experts which not necessarily perform a task in the same way as planners have thought of. The whole planning process for the production of new products is verified step by step involving departments in the headquarters with ­on-­site production workforce and off-site experts. Thus, potential issues are discovered at an early stage of development making required adaptations much more cost

6  Big Data, Small Data, and Getting Products Right First Time

81

effective. Simply use available data and don’t wait until shortcomings are apparent on the assembly line. While the topics mentioned above optimize the production cycle from an ergonomic or logistics point of view, some other aspects like quality management, sustainability, and customer involvement offer still some potential for improvement.

Augmented Reality Virtual reality (VR) and augmented reality (AR) technologies provide novel ways to interact with data through virtual models, providing, e.g., the ability to investigate different design trade-offs. VR provides fully immerse virtual environments to users, while AR focuses on augmenting real-world, physical objects with enhanced information (Burke et al. 2018). Video and image data processing were always of high importance. Potential use cases for an augmented reality approach in development, production, and service were investigated and introduced in early 2000 (Friedrich 2002). Display of production, maintenance, or repair instructions in the field of view were among the investigated use cases. Especially the support in the production of special products like jet fighters which are more like building one prototype after another instead of walking along a traditional assembly line would benefit from augmented manufacturing information. One of the most relevant shortcomings of that era was the insufficient (general) pattern recognition capabilities, which would have resulted in too many application-based interventions into the product and/or environment in order to guarantee adequate object detection/recognition. Another important aspect was the creation of the context sensitive information in the field of view of personnel. These aspects of “explicit programming” requirements made it difficult for AR to gain a major breakthrough at that time. Automotive industry introduced another AR feature into the market: Head-up Displays. BMW introduced as the first automotive company a first generation of color Head-up Display in 2004. Real world is augmented with additional information in the field of view of drivers, but since no object recognition was required, this function was easier to implement for the bigger market. Furthermore, the displayed information was anyhow available via the onboard navigation system. This technology will further improve, and additional information might be superimposed on real objects which enable a wide set of useful data for drivers. The development in the gaming industry and mobile communication yield to an unprecedented trend of cost-effective and high-performance solutions. Cluster of standard PCs equipped with high-performance graphic cards or even smartphones are used for industrial applications these days. The shortcomings of the “small data” era (e.g., object recognition) were targeted by artificial intelligence. The new opportunities enabled a wide variety of new applications of which some examples are shown below.

82

H. Ramezani and A. Luckow

AI for Computer Vision Video and image data is becoming the most prevalent and important data type in industrial applications. Approaches for extracting knowledge and guiding business decisions from these data are becoming essential. Artificial intelligence (AI) methods like data science, machine or deep learning can be used to improve products, processes, and services. These tools help us to better understand issues and derive more accurate predictions. Machine learning describes the extraction of knowledge and insights based of experience and available data without explicit programming. Patterns and regularities should automatically be detected and used to derive predictions. Deep learning (Goodfellow et al. 2016) refers to a set of machine learning techniques that utilize neural networks with many hidden layers for tasks, such as image classification, speech recognition, and language understanding to name a few. Luckow et al. (2016, 2018) studied the development of AI applications and proposed a holistic system architecture, which was used to investigate a set of manufacturing applications like: (i) Visual inspection of parts and components (quality management process) (ii) Barcode detection (iii) Truck trailer identification numbers In the following, we review the results and lesson learned.

Visual Inspection The visual inspection of parts is a core quality management process in manufacturing. Typically, multiple instances of this process exist in different production phases. The visual inspection has two major subcategories: detection of (i) the presence of parts and (ii) the presence of defects. Examples of type (i) systems are the Visual Inspector (Ramaraj 2015) and Q-Gate (Luckow et al. 2016). The Visual Inspector is a tool developed to identify missing parts on the assembly line using fixed-mounted cameras in conjunction with different traditional computer vision techniques, e.g., image similarity metrics. Q-Gate utilizes different object classification networks (e.g., GoogLeNet/Inception (Szegedy et al. 2014)). Deep learning techniques offer improvements by providing a better generalization and more complex applications in this area. More advanced approaches are required for type (ii) systems. The key challenge lies in the (typically) uniqueness of defects and the scarcity of available training data. It is straightforward to collect data for training the object detector for part presence identification. However, collecting substantial population of defects requires a huge initial investment in gathering and labeling high-resolution images. Alternative approaches are required for the detection of unknown defects such as unsupervised

6  Big Data, Small Data, and Getting Products Right First Time

83

learning systems based on auto encoders or anomaly detection. The successful deployment of an AI solution for visual inspection requires also higher maturity of the model to provide value and prevent rejection by end users. Maintainability and flexibility of the deployed model are also critical. To address these issues, it is crucial to focus on hardware for data collection (e.g., high-resolution cameras) and the software for processing the images. Since defects may be only visible in three-­ dimensional image (e.g., dents or flushness of the surface) or require high resolution (e.g., scratches on a shiny surface), the selection of the hardware that will maximize the information for training is essential. Many solutions utilize models deployed on the edge, but in the future the use of cloud-based approaches will be necessary in order to combine images from different quality stations and to frequently retrain complex models to include new inspections and changes in design and provide support for new products.

Logistics One of the most important portions of maintaining a smooth manufacturing process is keeping the assembly line fed with the correct materials at all times. This process starts by delivering the correct materials to the plant and supplying them to the correct locations at the right time and sequence. Instead of manually searching the storage warehouse and scanning each barcode separately, a deep learning model can be utilized to detect and read barcode labels on all boxes received at the production facility. By supplementing the manual work with an AI system, associates become more productive in a shorter amount of time. The model detects entire barcode labels, which can then be used with a barcode reader (e.g., Tesseract) to provide necessary information about the contents of the box as well as its source and destination. Another promising application is the management of trailer yards. Trailer IDs printed on the trailer can easily be read using AI-based computer vision; trailer can be identified using cameras mounted to yard management vehicles. Typically, it is the task of an associate to locate each truck and provide its parking location. For this purpose, a combination of object detection and OCR is required. It is important to keep track of all delivery vehicles and their locations in a trailer yard.

Infrastructure for AI Computer Vision Designing the right setup for the AI system is of huge importance, since AI infrastructure and applications are far more complex than traditional IT systems. Requirements range from the integration of various components including data sources, batch and streaming processing, data management systems, model management, and device management. To cope with these needs, edge computing gains

84

H. Ramezani and A. Luckow

more and more importance. Gartner subsumes edge computing as solutions in which data processing is performed at or near the source of data generation, in particular embedded devices or sensors, for example, using a large number of camera systems in the assembly process for conducting visual inspections (Hung 2018). Edge deployments are subject to various trade-offs related to model complexity, privacy, manageability, flexibility, and performance that need to be considered (Srivastava et al. 2018).

Natural Language Processing (NLP) NLP research started in the 1950s, but the major breakthrough was achieved within the last decade when machine learning methodologies became widespread and helped NPL find its way into the mass market. Apple introduced Siri in October 2011 as an integrated feature in the iPhone 4S. Nowadays services like Siri, Amazon Alexa, or Google Assistant are fundamental parts of our daily lives. Personal assistants are integrated into many consumer products. But the crucial success factor is not only the correct language processing on its own; it is also very important to deliver the requested feedback based on the provided input. BMW introduced a digital personal assistant for its cars that is mostly related to driving experience, but as the system gets to know the driver and the habits better, it improves over time (BMW 2018). Not surprisingly, NLP plays not only an important role in the consumer area. In order to improve production processes, hand-free applications – and thus language processing included – gain more and more importance. In order to make manufacturing processes more effective, employees need to get more flexibility to perform their tasks. Quality assurance activities are just one example. Instead of verifying topics manually step by step introducing feedback manually into a computer located next to the assembly line, wearable solutions can be used to (a) describe next task and (b) gather the observed feedback giving the worker flexibility with this hand-free approach. More value is delivered, when multi-input channels (e.g., voice) could be utilized. The vocal interaction should not be limited to a specific individual and/or specific set of keywords. Therefore the NLP improvements can bring major breakthroughs in this area. Combining such conversational user interfaces (CUI) with the object detection applications via deep learning mentioned earlier, an integrated system to enable an effective interaction with robots performing a variety of tasks within an unpredictable environment is just the natural next step (Ma et al. 2018). CUIs are also relevant for agent-based systems that are designed to allow complex communication between the system, the users, and the application in order to perform complex tasks. A use case based on a manufacturing process lies in support of the associate working on an assembly line. They are in constant need of a variety of different parts (option/objects). For each build, the associate must pick and choose the correct part to avoid any errors. Additionally, this must be done in a timely manner in order to prevent delays in the process. While many parts of manu-

6  Big Data, Small Data, and Getting Products Right First Time

85

facturing have been automated, assembly remains a challenge as the adaptability and dexterity of a person cannot be easily replaced via robot. Instead, there is an opportunity in helping the associate through a collaborative robot that can assist in tasks rather than remove the associate from it. This is most effective for process steps that are repetitive or error-prone.

Advanced Printing Another opportunity lies in the availability and evolvement of 3D printing. Raw materials can be used to create required components demand driven on-site instead of carrying the assembled parts through the logistic chain. This flexibility will open new planning opportunities within the production process. BMW Motorcycle division introduced iParts (BMW iParts 2018) as an optional equipment item in 2018. Even though this was an April Fools’ Day joke, it might give a hint for future opportunities. Availability of spare parts in remote locations or dealing with very rarely required parts might be solved more elegantly. The impact of advanced printing on logistics lies in provision of parts through data instead of physical handling of such over the product life cycle. This approach is naturally limited to a (smaller) subset of parts, but nevertheless, it will help to improve logistics and reduce the long-term carbon footprint. Besides supplier chain implications, 3D printing offers also a big potential for product customization and, hence, customer satisfaction. Clients can use this technology to create – to a certain extent – self-designed accessories and parts. “MINI Yours Customised” is an example (MINI 2018). Customers can design specific parts and transfer their car to something unique (Fig. 6.1). Taking this approach further, parts will be defined through data which is in direct control of customers. Certain geometrical boundaries are given, but a large degree of freedom is left to the control of the end user. The technologies previously mentioned might also bring a whole new production strategy into consideration. Instead of building a production facility at an isolated place, corporations might build a facility exactly at the place where it is required and disassemble when the need is satisfied. Project MILESTONE (Vanwijnen 2018) might be a very first example for

Fig. 6.1  Examples of MINI Yours Customised – led door sills and side scuttles

86

H. Ramezani and A. Luckow

such an approach. The construction company Van Wijnen uses 3D printers in the City of Eindhoven in the Netherlands to “print” houses on-site which will not only cut costs but also reduce environmental issues. A robotic arm with a nozzle uses a special cement to print walls, layer by layer, creating the house based on the architectural design. Another beauty of this method is the ability to create buildings of any desired shape, because this solution – in contrast to traditional concrete – offers a wide range of flexibility. Customers design their houses based on their own needs and personal taste. Furthermore, this technology allows sensors to be placed at any desired location on the property resulting in a totally smart home.

Blockchain Blockchain is another technology that might have an immense impact on the industrial value chain. As defined by Gartner, blockchains or distributed ledger technologies are decentral, distributed systems for maintaining and sharing data in a trustless environment. Each data record is cryptographically secured ensuring that it can be verified by every participant of the network. Using blockchain, business processes across organizational boundaries can be tightly coupled and thus provide immense potentials to improve process efficiency, transparency and provide the ability to explore new business models. Further, blockchains provide the ability to decentralize large parts of the IT infrastructure to address performance and privacy needs, overcoming issues such as trust and the ability to conduct secure transactions. Supply chains are a prime example of cross-organization collaboration that benefits from blockchains. Cabalero and Hamilton (2018) investigated the improved supply chain visibility by providing verifiable authenticity and traceability of parts. Moreover, blockchain enables the verification of original parts and, thus, represents an efficient counterfeit protection. PartChain (Miehle et al. 2018) is an example for creating, monitoring, and sharing part information across automotive OEMs and suppliers. Considering business functions that cross organizational boundaries, further applications would benefit from this approach, e.g., cross-company payments, improved transparency in joint ventures, and improved risk management. Further, customers can also benefit from blockchain-based systems in unforeseeable ways as new ideas and services that do not require intermediaries emerge, e.g., in the domain of mobility, energy/charging, finance, etc. The landscape of public and enterprise blockchain systems is evolving rapidly. Three main areas can be thought of: 1. Public Blockchains: Cryptocurrencies, such as Bitcoins (Nakamoto 2008), were the first successful application that utilize the concepts now referred to as blockchain. With the success of Bitcoin, the new possibilities of the technologies became apparent. Projects like ColorCoins (Rosenfeld 2012) provided some innovations on top of the Bitcoin protocol and enabled rudimentary more advanced transactions by associating metadata with coins. Ethereum (Buterin

6  Big Data, Small Data, and Getting Products Right First Time

87

et  al. 2018) generalized the ledger concept further and utilizes the decentral blockchain protocols to manage the execute of smart contracts. While Ethereum was originally designed as public, permissionless blockchain, it can also be deployed in private and permissioned settings. 2. Private and Consortia Blockchains: In addition to these public, permissionless blockchains, several systems suitable for smaller semipublic deployments, such as consortia, emerged. These blockchains particularly address enterprise requirements, such as scalability, confidentiality, and governance. In contrast to public blockchains, private and consortia blockchains assume that identities of participants are known and that nodes are run on managed IT environments. Ethereum can be used in public and consortia/private deployments. Blockchains, such as Chain Core, Corda (Hearn 2016), HyperLedger Fabric (Androulaki et al. 2018), and Hyperledger Sawtooth (Intel 2018), are primarily designed for consortia. 3. Data Markets, IoT, Machine-to-Machine Transactions: A main limitation of current public and private blockchain is scalability. An emerging set of use cases requires a more advanced computational capabilities and the management of high volume and velocity data. An example is tracking of the provenance of IoT sensor data and the ability to support high-frequency machine-to-machine transactions. Emerging blockchains, such as IOTA (Popov 2017), Hedera Hashgraph (Baird et al. 2018), and XAIN (XAIN 2018), aim to provide the necessary scale to support data-intensive blockchain applications. To support a wide variety of blockchain application, a scalable and flexible infrastructure is required. Figure  6.2 illustrates a Blockchain-as-a-Service reference architectures comprising of several modular building blocks of core and management infrastructure. An often neglected aspect are off-chain capabilities, such as off-chain data management and security services (key management, identity management). Different solutions present different abstractions for smart contracts, consensus algorithms, and the enforcement of privacy and security policies. The technological choice strongly depends on the specific requirements, e.g., a non-­

Fig. 6.2  Blockchain as a Service

88

H. Ramezani and A. Luckow

Fig. 6.3  Blockchain market and enterprise adoption roadmap

byzantine consensus algorithm is suitable for many medium-sized permissioned blockchain deployments that require higher throughputs. Blockchain technologies are still in its infancy. We expect that the technologies will be adopted in multiple steps as depicted in Fig. 6.3. The first generation of use cases focuses on the utilization of blockchains to track digital and physical objects in cross-organizational settings. The establishment of such track and trace systems provides the basis for more advanced services, such as the conduction of secure value exchanges of both monetary values and goods. As the technology matures, a larger amount of data (including IoT sensor data) and transactions (including machine-to-machine transactions) will be managed by blockchain-based systems.

Discussion Industry 4.0 is shaped by advances in data processing technologies, in particular the Internet of Things (e.g., RFID), large-scale data processing, and machine learning enabled AI and blockchain. While big data is a key ingredient for advanced machine learning capabilities in the domain of computer vision and natural language, small data is the key for actionable decisions. For example, AI systems must provide an explainable output that is related to the object it investigates (small data). In the case of visual inspection, it needs to provide a sufficient amount of details on the location of the detected error. Another example are blockchains, which enable the digitization of cross-organizational value chains, such as supply chains. They are a critical component for maintaining high-quality, small datasets across organization boundaries. In summary, big and small data systems play a crucial role in modern Industry 4.0 value chains enabling the design and production of high-quality products. The confluence of blockchain, IoT, and AI enables every entity, i.e., person, machine, or thing, to have an identity, be intelligent, and transact with other entities.

6  Big Data, Small Data, and Getting Products Right First Time

89

References Androulaki, E., Barger, A., Bortnikov, V., Cachin, C., Christidis, K., De Caro, A., Enyeart, D., Ferris, C., Laventman, G., Manevich, Y., ralidharan, S. M., Murthy, C., Nguyen, B., Sethi, M., Singh, G., Smith, K., Sorniotti, A., Stathakopoulou, C., Vukolic, M., Cocco, S. W., & Yellick, J. (2018). Hyperledger fabric: A distributed operating system for permissioned blockchains. In Proceedings of the thirteenth EuroSys conference, EuroSys ‘18 (pp. 30:1–30:15). New York: ACM. Baird, L., Harmon, M., & Madsen, P. (2018). Hedera: A governing council & public hashgraph network. https://s3.amazonaws.com/hedera-hashgraph/hh-whitepaper-v1.1-180518.pdf BMW. (2018). Intelligent personal assistant. https://www.bmwgroup.com/en/company/bmwgroup-news/artikel/IPA.html BMW iParts. (2018). https://www.press.bmwgroup.com/global/article/detail/T0279598EN/ bmw-motorrad-iparts-revolutionises-spare-parts-management?language=en Bruner, J.  (2013). The industrial internet  – the machines are talking. http://radar.oreilly. com/2013/03/industrial-internet-report.html Burke, B., Cearley, D., & Blau, B. (2018). Top 10 strategic technology trends for 2018: Immersive experience. G00344889. Buterin, V., et al. (2018). Ethereum: A next-generation smart contract and decentralized application plat-form. White Paper: https://github.com/ethereum/wiki/wiki/White-Paper Caballero, G., & Hamilton, S. (2018). Blockchain in supply chains: Looking beyond the hype. MIT https://ctl.mit.edu/events/tue-10242017-1730/blockchain-supply-chains-looking-beyond-hype Dhar, V. (December 2013). Data science and prediction. Communications of the ACM, 56(12), 64–73. Evans, P. C., & Annunziate, M. (2012). Industrial internet, ge technical report. http://www.ge.com/ sites/default/files/Industrial_Internet.pdf Friedrich, W. (Oct 2002). Arvika-augmented reality for development, production and service. In Proceedings of international symposium on mixed and augmented reality (pp. 3–4). Goodfellow, I., Bengio, Y., & Courville, A. (2016). Deep Learning. MIT Press. http://www.deeplearningbook.org. Hearn, M. (2016). Corda: A distributed ledger. https://www.corda.net/content/corda-technicalwhitepaper.pdf Hellinger, A., Stumpf, V., & Kobsda, C. (Eds.). (2013). Umsetzungsempfehlungen fur das Zukunftsprojekt Industrie 4. Hung, M. (2018). Iot implementation and management — from the edge to the cloud: A gartner trend insight report. Gartner. Intel. (2018). Hyperledger sawtooth. https://sawtooth.hyperledger.org/docs/core/releases/latest/ Lasi, H., Fettke, P., Kemper, H.-G., Feld, T., & Hoffmann, M. (Aug 2014). Industry 4.0. Business and Information Systems Engineering, 6(4), 239–242. Luckow, A., Cook, M., Ashcraft, N., Weill, E., Djerekarov, E., & Vorster, B. (2016, Dec). Deep learning in the automotive industry: Applications and tools. In 2016 IEEE international conference on Big Data (Big Data) (pp. 3759–3768). Luckow, A., Kennedy, K., Ziolkowski, M., Djerekarov, E., Cook, M., Duffy, E., Schleiss, M., Vorster, B., Weill, E., Kulshrestha, A., & Smith, M. (2018, Dec). Artificial in- telligence and deep learning applications for automotive manufacturing. In 2018 IEEE international conference on Big Data (Big Data). Ma, M., Luckow, A., Kennedy, K., & Schleiss, M. (2018). Voice bot system design for appli- cation with collaborative robotics in manufacturing. in submission. Miehle, D., Stroebel, M., Henze, D., Seitz, A., & Bruegge, B. (2018). Partchain: A blockchain-­ based traceability system for supply chain networks. in preparation. MINI. (2018). Yours customised. https://yours-customised.mini Nakamoto, S. (2008). Bitcoin: A peer-to-peer electronic cash system, http://bitcoin.org/bitcoin.pdf Ohno, T. (1988). Toyota production system: Beyond large-scale production. Taylor & Francis. Popov, S. (2017). The tangle. http://iotatoken.com/IOTA_Whitepaper.pdf

90

H. Ramezani and A. Luckow

Ramaraj, M. K. (2015). A training assistant tool for the automated visual inspection system. https:// tigerprints.clemson.edu/cgi/viewcontent.cgi?article=3290&context=all_theses, 9. Rosenfeld, M. (2012). Overview of colored coins. https://bitcoil.co.il/BitcoinX.pdf Seidl, A. (1997). Ramsis-a new cad-tool for ergonomic analysis of vehicles developed for the german automotive industry. Technical report, SAE Technical Paper. Srivastava, A., Nguyen, D., Aggarwal, S., Luckow, A., Duffy, E., Kennedy, K., Ziolkowski, M., & Apon, A. (Dec 2018). Performance and memory trade-offs of deep learning object detec- tion in fast streaming high-definition images. In 2018 IEEE international conference on Big Data (Big Data). Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S. E., Anguelov, D., Erhan, D., Vanhoucke, V., & Rabinovich, A. (2014). Going deeper with convolutions. CoRR, abs/1409.4842. Vanwijnen. (2018). https://www.vanwijnen.nl/actueel/wereldprimeur-wonen-in-een-3d-geprinthuis-van-beton/ XAIN. (2018). Xain: The trusted access control protocol for machine networks. https://xain.io/

Chapter 7

The Internet of Things and Sustainable Manufacturing David Heatley and Mohamed Abdel-Maguid

In the relatively short time since the term “The Internet of Things (IoT)” was first coined by Kevin Ashton in 1999 (Ashton 2009) and the opportunities of an Internet which interconnects billions of network-enabled devices began to be seriously explored, much progress has been made developing enabling technologies and service platforms and progressively integrating the IoT into the fabric of everyday life. This chapter reviews the current state of the art of the IoT and what can be expected in years to come, with a focus on the manufacturing sector and the key enabling role of the IoT in Industry 4.0. This chapter begins by describing what the IoT is and just as importantly what it is not. The many sectors in which it is already having and will continue to have a profound impact are outlined to give a flavour of the scope of applications of the IoT and the nature of the benefits. The remainder of the chapter then focuses on the manufacturing sector, which of all the sectors alluded to is the one that promises to benefit most immediately and most profoundly from the IoT. Particular attention is paid to the new capabilities that the IoT is uniquely equipped to deliver. Example use cases are presented that highlight the scope and impact that the IoT is already having in manufacture, particularly in the context of supporting growth and sustainability. The chapter concludes by offering a brief vision of how the IoT, together with cloud computing and big data, are underpinning the realisation of Industry 4.0 – the next evolutionary phase that embraces manufacturing and all other industry sectors in a digital world.

D. Heatley (*) · M. Abdel-Maguid University of Suffolk, Ipswich, UK e-mail: [email protected]; [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_7

91

92

D. Heatley and M. Abdel-Maguid

The Internet of Things The IoT alludes to the ubiquitous interconnecting of all kinds of network-enabled electronic devices, machines and systems via the Internet. It is not in itself a new piece of stand-alone technology destined to have a profound impact on society in the same way as, for example, smart phones or flat high-definition display screens. Rather, the IoT is a conceptual architecture – a high-level abstraction – that naturally arises from the astonishing growth in network-enabled mechanical and technological entities that carry out functions in their own right and share information about those functions and their outcomes with other entities via the Internet. The IoT promotes a vision in which ubiquitous interconnectivity across all kinds of network-­enabled entities and the sharing of information creates new value chains through better ways of interacting, communicating, gathering data, processing and applying information and knowledge gained, across all sectors. Interestingly, Kevin Ashton, the originator of the title IoT (Ashton 2009), admitted during a radio interview in 2015 (Ashton 2015) that he wished he had proposed the title “Internet for Things” which more accurately conveys his original vision in which the Internet is purely an enabler of the connectivity between all things which then gives rise to the new functionality, rather than the Internet in itself being the new functionality. That title, had it been chosen, would also have more clearly conveyed his intended meaning as a conceptual framework, made real by drawing together and integrating a broad array of stand-alone technological things. Nevertheless, through time the title “Internet of Things” has become embedded in today’s vernacular and so it is adopted in this chapter.

Architecture Traditionally network-enabled devices and machines have been deployed to fulfil specific tasks required by their users and the sector in which they operate. For example, production lines within a manufacturing plant have been instrumented (i.e. equipped with network-enabled sensors and other monitoring devices) in a manner that is specific to that plant’s product line and the operational requirements of the organisation. Similarly, road haulage and goods delivery vehicles have been instrumented to enable the host organisations to track their location and keep customers appraised of when delivery can be expected; and clinical appliances within a large hospital have been instrumented to enable the staff to quickly locate shared items (e.g. crash trollies) and enable the hospital administrators to managed their inventory. Deployments such as these and numerous other examples that need not be described are highly effective within their respective organisations but are fundamentally isolated from each other. They can be viewed as island deployments, each implementing a sort of “INTRAnet of Things”. By not coming together as an “INTERnet of Things”, opportunities are missed. For example, if a manufacturer of

7  The Internet of Things and Sustainable Manufacturing

93

Fig. 7.1  Architecture reference model for the IoT

clinical equipment is experiencing a delay on the production line and that information can be shared with the manufacturer’s subcontracted delivery agency, that agency can update its schedule accordingly and then alert the destination hospital of the delay and the new delivery schedule. All of this could be done automatically with little if any involvement from human operators in any of the organisations involved. The key is sharing information between each of these organisations’ operating systems via the Internet. Figure  7.1 illustrates the Architecture Reference Model (ARM) that has been designed to support this kind of information sharing and other core attributes that are fundamental to the realisation of the IoT (Bassi et al. 2013). The ARM in Fig. 7.1 is sometimes referred to in the technical literature as the Functional Model for the IoT since it summarises the key top-level Functionality Groups (FG) within the ARM and their interactions with each other. The device FG specifies the type of network-enabled devices and their functionalities that produce the data required by the application FG. The network connectivity to these devices is specified in the communication FG which accommodates the plethora of connectivity technologies that can be employed, fixed/cabled and wireless/mobile. This FG also specifies the various networking and other nodes (e.g. hubs, routers, data stores, etc.) located between the devices and the Internet. Collectively the FGs between the communication and application FGs specify the requirements of the stakeholders. For example, the clinical equipment manufacturer in the above example will specify: • The types of raw materials and other supplies that are critical to the production lines. • How the progress of the lines is monitored and how that data needs to be presented to human operators.

94

D. Heatley and M. Abdel-Maguid

• The kind of decisions that need to be taken and must be facilitated by the deployed system. • The manner in which the product is delivered to customers and how that is monitored. • How the manufacturer interfaces with their suppliers and customers in all interactions. • The kind of service and capabilities that the manufacturer expects from the IoT. • The ability to build additional services and applications on top of the IoT to respond to new business opportunities. All of these details and more are expressed within the specifications of these FGs. The security FG enforces the specified policies concerned with data security, privacy and trust in the IoT. Crucially, it ensures no activity within any FG can circumvent the policies and gain unauthorised access. Lastly, the management FG is responsible for managing all of the FGs and the interactions between them. The arrows in Fig. 7.1 depict where interactions are permitted by the security and management FGs. The security and management FGs directly interact with all the other FGs; however to avoid clutter in the diagram, these interactions are not depicted as arrows. Instead, these ARM-wide interactions are implicit.

Connectivity The IoT accommodates all forms of fixed/cabled and wireless/mobile technologies to connect network-enabled devices and machines to the Internet. Installations that use fixed/cabled connections will most likely exclusively employ Ethernet technology. However, where wireless/mobile connectivity is required, the options are more extensive and still expanding as new technologies emerge. The common schemes are summarised in Fig. 7.2 (Woolhouse 2016). The wireless technology most suited to the application is primarily determined by the data rate that must be supported and the distance over which it must be communicated. Other influencing factors might include the link robustness required by the application (i.e. the availability of the link), the state of maturity of the technology and ultimately cost. For in-building deployments where the distance is just a few metres up to perhaps 100 m, data rates up to 100 kbit/s can be supported at low cost by Bluetooth and Zigbee. Data rates extending to 10Mbit/s over the same distance can be supported by WiFi and the LTE schemes but at a progressively higher cost. If however operation over a longer distance is required, for example, in outdoor deployments where the network-enabled devices might be a km or more from the receiving station/node, data rate is typically limited to a few hundred kbit/s, and the wireless technologies are potentially costly. Cost comparisons must take into account whether the required connectivity infrastructure already exists. For example, IoT deployments that use Bluetooth or

7  The Internet of Things and Sustainable Manufacturing

95

Fig. 7.2  Wireless connectivity technologies for the IoT

Zigbee might prefer to provide their own end-to-end installations since they are low cost and doing so avoids any possibility of incompatibility between the system being deployed and the local infrastructure. In contrast, IoT deployments that use WiFi might wish to capitalise on existing infrastructure (e.g. hubs and access points) to reduce installation cost in the confidence that the maturity of WiFi standards minimises the possibility of incompatibility issues. IoT deployments that use LTE will benefit from its existing large-scale infrastructure (e.g. base stations and backhaul connections) as well as its carrier-grade link availability because it occupies a licenced portion of the electromagnetic spectrum which ensures minimal interference from other types of wireless systems, e.g. emergency services and commercial broadcast channels. In contrast, Bluetooth, Zigbee and WiFi occupy a licence-­ exempt portion of the spectrum which is subject to interference from other systems, and hence link availability is variable. Further details of the wireless connectivity technologies for IoT are given elsewhere in this book.

Growth of Interconnected Devices According to data published by Statistica (2018a), the number of globally interconnected devices is projected to triple from 23 billion in 2018 to 75 billion in 2025. Figure 7.3 shows that the growth is predicted to accelerate year on year, driven by the increasing rate of uptake of the IoT as the associated technologies mature and the benefits to users become compelling must-haves. Underpinning as well as driving the acceleration in Fig.  7.3 is the associated growth in the number of machine-to-machine (M2M) connections (Statistica 2018b) and the types of deployed sensors (Statistica 2018c) shown in Figs.  7.4 and 7.5, respectively.

96

D. Heatley and M. Abdel-Maguid

Fig. 7.3  Installed base of interconnected IoT devices worldwide from 2015 to 2025 (Statistica 2018a)

Fig. 7.4  Number of M2M connections worldwide from 2014 to 2021 (Statistica 2018b)

Fig. 7.5  Projected global IoT-enabled sensors by market segment in 2022 (Statistica 2018c)

7  The Internet of Things and Sustainable Manufacturing

97

All of the above charts serve to emphasise that the uptake of the IoT is a major, global trend, driven by the compelling benefits that the IoT is uniquely equipped to deliver and also by the growing realisation that these benefits are absolutely vital to growth and sustainability. This close interdependency between the IoT, growth and sustainability is certain to strengthen over time. It is important to note that the numbers in the above charts are baseline estimates that do not take into consideration upcoming advances in the Internet or other IoT enabling technologies that deliver game-changing improvements far beyond what was anticipated. The actual numbers in due course could therefore be substantially higher.

Global IoT Market The functionality made possible by the IoT brings benefits to multiple sectors: public and private and consumer and industrial. According to an analysis carried out by Growth Enabler in April 2017 (Growth Enabler 2017) and referenced by Forbes (Forbes 2017) and other sources of statistical data, the global IoT market, valued at $157 billion in 2016, is predicted to rise in value to $457 billion by 2020, as shown in Fig. 7.6. That equates to a compound annual growth rate (CAGR) of 28.5%. This figure is notably higher than the CAGRs for other rapidly growing markets such as mobile apps (CAGR of around 14% over 2016 to 2022 (Forward Geek 2016)) and automotive telematics (CAGR of 20.4% from 2016 to 2024 (Variant Market Research 2017)). The global IoT market in 2020 is anticipated to be dominated by manufacture (24%), smart cities (26%) and connected health (20%), followed by smart homes (14%), connected cars (7%), smart utilities (4%) and wearables (3%). This predicted view of the near future is consistent with the data map in Fig. 7.7, extracted from a Forrester report (Stroud 2017), which shows the sectors (highlighted in black) where there was the greatest interest in 2017 in integrating the IoT into their operations going forward.

Fig. 7.6  Anticipated growth of the global IoT market across multiple sectors (Growth Enabler 2017; Forbes 2017)

98

D. Heatley and M. Abdel-Maguid

Fig. 7.7  IoT opportunities against the level of interest shown by sectors in 2017 (Stroud 2017)

Manufacture is a clear early adopter along with transport and logistics. Health is not yet showing a similar commitment, but there are indications in today’s media and health sector publications of a trend that should see IoT substantially adopted in that sector by 2020, as predicted in Fig. 7.6.

Impact of the IoT on the Economy The IoT will have a profoundly beneficial impact on the economy of individual nations and worldwide because of its enormous scale and depth of penetration into all aspects of living. Not only does it promise a world of uniquely identifiable interconnected smart objects but also entire smart cities and more, all interoperable across platforms that support intelligent communication, resource identification and information gathering and sharing. This will create new employment opportunities and new ways of economic growth, which will lead to a more efficient utilisation of resources and reduce the burden on economies to sustain vital processes that are inherently costly, particularly in developing countries. Taken as a whole, it is estimated that these opportunities, enabled by the IoT, will have a total potential economic impact of up to $11.1 trillion per year by 2025 across all major sectors (up to $3.7 trillion for manufacture alone) (McKinsey & Company 2015). That is equivalent to around 11% of the projected world economy at that time.

7  The Internet of Things and Sustainable Manufacturing

99

Achieving this level of impact is dependent upon technical, organisational and regulatory challenges continuing to be overcome and solutions implemented. In particular, organisations that use the IoT will need better tools and methods to extract insights and actionable information from IoT data. It will take time for systems to be created that maximise IoT value, as it will for management innovations, organisational changes and new business models to be developed and implemented that accurately relate the projected benefit to a resulting real operational improvement. This could lead to a “productivity paradox” – a lag between investing in new technology and receiving irrefutable evidence of consequential operational gains. This phenomenon was first observed by economists Robert Solow and Stephen Roach, who in 1987 noted that despite the widespread adoption of computers to automate office functions, there was no clear evidence of their improvement on productivity (Brynjolfsson and Hitt 1998). Subsequent research discovered problems in how government statistics measured the impact of computers and a lag between investment in technology and the organisational adjustments required to realise significant productivity gains. Although the productivity paradox is now a well-understood phenomenon the problematic consequences of which can be guarded against, the IoT might be particularly susceptible to its effects, while the tools to meaningfully measure operational improvement are still in development. It is therefore advisable that organisations that have invested in, or plan to invest in the IoT, have appropriate contingencies in their business plans. Significantly, in the fullness of time, the IoT might have a similar impact on the economy of developing countries as well as advanced economies. The IoT is highly likely to return economic value more rapidly in advanced economies compared with developing countries because of the extensive infrastructure already in place and the higher value associated with each deployment. However, in contrast there is potential for a greater number of IoT uses in developing economies, and the absence of legacy technologies enables IoT implementations to be deployed at a faster rate. Consequently, in due course the rate of economic return from the IoT in developing economies could match or even exceed that in advanced economies. Current studies estimate that around 62% of the potential annual economic impact of IoT applications in 2025 will be in advanced economies, with the remaining 38% in developing economies (McKinsey & Company 2015). The higher value in advanced economies mostly reflects the higher wages and costs in those regions, which raise the economic value of the benefits delivered by the IoT, for example, in improving productivity and the utilisation of assets. As these same values rise in developing economies, the economic return from the IoT in those regions will also grow.

Technology Standards for the IoT The growth of the IoT across multiple sectors is fundamentally reliant on the creation of appropriate standards for the technologies and operating practices employed. Given the vast numbers of connected devices predicted in Fig.  7.3, each one of

100

D. Heatley and M. Abdel-Maguid

which must be uniquely addressable via the Internet, it is essential that the IP (Internet Protocol) address field is sufficiently wide to cope with these requirements and have ample headroom for future expansion. The latest IP addressing protocol, IPv6, which was released in 2010 and formally launched in 2011 (Wikipedia 2018) was specifically designed with that and other requirements in mind. It accommodates 3.4 × 1038 addresses which is many orders of magnitude more than the 7.5 × 1010 addresses predicted to be required by 2025. Indeed, it is considered highly unlikely that the address field of IPv6 will ever run out through the needs of the IoT and other address-hungry developments that might come along. To put the scale of the IPv6 address field into an easily visualisable context, every grain of sand on planet Earth (NPR 2012) and a further 1019 Earths could each have its own unique IPv6 address. The address field is truly vast! A further crucial requirement is a common communication language that all interconnected devices can understand and use to share information. MTConnect (2018) has been developed for that purpose, specifically to promote interoperability between the diverse range of manufacturing equipment and supporting IT infrastructure typically found on the shop floor. It provides a clearly defined vocabulary with which machine tools can express themselves in a common language that is interpretable by software applications. Crucially it is built upon standard Internet conventions (HTTP, TCP/IP, XML and Ethernet) which ensure its full compatibility with all Internet-enabled devices. While MTConnect specifically facilitates the exchange of information between machine tools and related manufacturing equipment, standards such as the OPC Unified Architecture (OPC-UA for short) (OPC Foundation 2018), released in 2008, address the interoperability required for plant-wide data collection. It specifies an open, non-proprietary set of specifications that ensure that automation systems in manufacturing are compatible with one another and can communicate with a high-­ level, plant-wide control system. It provides the architecture for the bridges that connect these individual automated systems to the plant-wide data highway that interconnects all the systems within the factory and connects the factory to the outside world via the Internet. Crucially, the organisations that oversee MTConnect and OPC-UA are working together to ensure these standards are fully compatibility with each other. This in turn should ensure that no manufacturing machines nor their supporting IT are excluded as the IoT continues to grow in scale and penetration at an ever-­accelerating pace.

People Empowerment It is often argued that because the IoT reduces the scale of human involvement in certain activities, particularly in manufacture and other industrial sectors, it will have a detrimental impact on employment and the growth of workforce skills. However, research by the World Economic Forum (2015) has concluded that the IoT “… will drive growth in productivity by presenting new opportunities for

7  The Internet of Things and Sustainable Manufacturing

101

people to upgrade skills and take on new types of jobs that will be created. While lower-­skilled jobs, whether physical or cognitive, will be increasingly replaced by machines over time, the IoT will also create new, high-skilled jobs that did not exist before. Companies will use IoT technologies to augment workers, making their jobs safer and more productive, flexible and engaging. As these trends take hold, and new skills are required, people will increasingly rely upon smart machines for job training and skills development”. As has been the case with many other new productivity-improving technologies, the IoT will impact workers in different ways. The value of some types of knowledge workers will increase since the IoT will create new needs for human judgement and decision-making. In contrast, the demand for workers in some services such as food preparation, office and home cleaning services and retail checkout could fall as such tasks become predominantly automated, particularly in economies where wages are high. Certain kinds of manual work will inevitably come under increasing pressure from the IoT; however it will also open up new employment opportunities. For example, skilled workers will be needed to install and maintain the physical elements of IoT systems, while other skilled workers will be needed to provide specialist support for installed IoT systems as well as design and develop new IoT systems. With all things considered, the IoT has the potential to have a net positive impact on employment. Indeed, the IoT might prove to be vital to the long-term sustainability of a skilled, engaged workforce.

Social Impact of the IoT The sharing of data and information, which is a fundamental attribute of the IoT, will impact society in many ways, both beneficial and potentially detrimental (BCS 2013). The IoT has the potential to revolutionise the way people and businesses communicate, interact, learn and go about their daily activities. It will have a profound impact on socially critical sectors such as health, finance, homes, transport, power/ energy, education and retail/distribution as well as manufacture and other industry sectors. It will also facilitate the next generation of socially critical applications, from dynamically managing traffic flow and monitoring the impact on air pollution to assisted living applications for the elderly, tracking tagged clinical assets in ­hospitals and various objects in shops and stores and more. These innovations and the new knowledge gained from the captured data will help to raise the standard of living and make societal resources more accessible and widely utilised. However, these and other benefits of the IoT must be balanced against the attendant risks to privacy, data protection and security through the sharing of data and information and their detrimental impact on society. More is said later about these risks. Additionally, inequalities in the accessibility of data of value to individuals and communities must be avoided; else they might create parallels with the problematic digital divides that already exist within society.

102

D. Heatley and M. Abdel-Maguid

Environmental Impact of the IoT Like many new technological innovations, the IoT has the potential to have both a positive and detrimental impact on the environment. Crucial to the sustainability of the IoT is ensuring that the positives outweigh the negatives. Consider first some examples of the positives. The IoT would reduce the carbon footprint of commercial road transport and logistics by enabling road traffic flow to be managed more efficiently, which has already been alluded to. By gathering data from roadside traffic flow meters as well as metrological systems and nationwide transport authority data bases, a more informed view of the real-time traffic conditions on the ground can be obtained and used to determine the optimum route diversions to put in place should the need arise, aimed at minimising the disruption and delays experienced by road users. By helping to keep journey time to a minimum for as many road users as possible, the IoT can help to reduce the volume of fuel emissions into the atmosphere (Pazvakavambwa 2018). A further positive arises from the ability to use the IoT to track the location of objects that have specialist disposal requirements such as harmful chemicals, clinical waste and more. Improper disposal activities such as fly-tipping can be detected and action taken against the perpetrators. Crucially, the waste material can be recovered and correctly disposed of. Interestingly a further positive arises from the need to provide electrical power to the sensors and other items of technology that make up the IoT. Power provision, be it from a mains supply or internal battery, generally comes with negative environmental credentials for several widely reported reasons. However, in applications where sensors are deployed in huge numbers over vast geographic areas, such as in nationwide traffic monitoring and meteorology, these sensors can scavenge power from solar cells or heat exchangers. They therefore have a zero footprint compared with pre-IoT capabilities. Even allowing for the negatives that may arise in the manufacture of power scavenging devices and the raw materials they use, there is still likely to be a net positive gain because of the vast numbers of sensors involved. On the negative side, the IoT has the potential to generate vast volumes of data which needs to be stored in data centres which are unavoidably power hungry because of the vast scale of their IT systems and the expansive cooling required. When considering the environmental impact of such centres, a key metric is the “information-to-kW ratio”. For any given kW of power used by the data centre, how much information can be stored? The objective is to maximise the information-to­kW ratio. Given that the carbon footprint of a data centre, excluding the proportion attributed to cooling, is predominantly governed by the total “digital” data storage capacity of the centre, which is a fixed quantity, the ability to store more information within that capacity results in fewer additional data centres being needed, in which case the net carbon footprint is reduced. Consequently, a key goal in IoT data management is to develop lossless data compression techniques that make maximum use of the available storage capacity.

7  The Internet of Things and Sustainable Manufacturing

103

A further negative, albeit a relatively short-lived one, is the potential need for newly adopted IoT users to dispose of their current pre-IoT items of technology which for one reason or another are not compatible with the IoT.  Consequently, there will be an outflow of elderly machines and devices, only some of which will be upgradable or recyclable to some degree. The rest will need to be scrapped, which has environmental implications. This trend will reduce as the penetration of machines and devices that are fully compatible with the IoT grows. Generally speaking the IoT promises to have a positive overall impact on the environment. The likelihood of this will increase as the coverage of the IoT increases and IoT-enabled devices become more widespread and embedded in everyday activities. This trend will serve to promote and underscore the sustainability credentials of the IoT.

The IoT in Manufacturing: The Opportunities The previous sections have considered the attributes of the IoT for multiple application sectors. The remainder of this chapter focusses on manufacturing. When the IoT is deployed in manufacture and other industrial settings, it is often relabelled as the Industrial IoT (IIoT for short). This helps to draw a distinction between the group of IoT applications that are specific to manufacture and industry in general and the many other groupings of IoT applications which include healthcare, smart domestic appliances, automotive and more. The IIoT label also categorises the kind of information that is produced by and shared between devices in industrial settings, as distinct from information concerned with healthcare, the status of domestic appliances and so on. Interestingly, the underlying technologies of the IoT and IIoT are often the same, particularly with regard to sensors, wireless connectivity, network routing and more. Indeed there is no difference at all in the IP addressing used in both cases, nor in the underlying Internet technology (HTTP, TCP/IP, XML and Ethernet). Generally the only difference worthy of note between, for example, smart sensors designed for manufacture applications and those for healthcare applications is the degree of physical robustness, size and weight. For the purposes of this chapter, it is not necessary to draw a distinction between IoT and IIoT since the focus is on manufacture. However, because IoT is by far the more common vernacular, it will continue to be used explicitly throughout this chapter. The relevance of the IIoT to this chapter is implicit. The opportunities afforded by the IoT in manufacturing can be visualised as the layers of an onion, as illustrated in Fig. 7.8. At the core is the ability for manufacturers to make better, more informed decisions. That leads to improvements in overall productivity which then increases revenue, which is spent on improving customer satisfaction and service delivery, which attracts investment to expand the business, all of which ultimately underpins the sustainability of the business.

104

D. Heatley and M. Abdel-Maguid

Fig. 7.8  The opportunities afforded by the IoT in manufacturing

Decision-Making The IoT delivers the means for manufacturers to have a detailed view of the real-­ time status of their entire production line, from the sourcing of raw materials to processing and machining, product assembly and testing, through to delivery to customers and after sales services. The Internet-enabled devices that are present in every part of this chain of activities deliver real-time data about the status of each stage. By collating that data from all devices and importing it to software designed to deliver an end-to-end visualisation of the status of the whole suite of operations, manufacturers can see at a glance whether issues might be arising at that very moment, what they are and where they are, and then make decisions based upon a complete and up-to-the-minute knowledge of the status of their operations. Software that uses machine learning and AI techniques can be used to assist the decision-­ making process, particularly in instances when multiple interdependent variables are involved. This same software allows manufacturers to test a variety of possible decisions in a safe virtual environment, to identify the decision that is most likely to deliver optimum results prior to making a firm decision in the real world. By understanding the consequences of decisions more fully, manufacturers can achieve strategic objectives and benchmark performances because the final decisions are based on real knowledge and wisdom, not theory or guesswork. Better decisions mean fewer mistakes and less waste, which reduces operating costs and streamlines productivity, all of which improves the sustainability of the business. A smart manufacturing

7  The Internet of Things and Sustainable Manufacturing

105

environment such as this enables managers to determine whether every element of a production line is performing at an optimal level. For example, machining parameters such as cutting speeds and the resulting quality of the product will reflect the settings that are most effective, as proven by operator experience, cutting tool manufacturer, machine builder and the online community of end users. Using that feedback, real-time updates to the settings can be automatically applied to maximise productivity, minimise waste and energy consumption and ensure that products are consistently at a high standard. Implicit in the above is that effective decision-making relies on the correct, most relevant data being captured and presented to decision-makers in a timely fashion. The IoT delivers the means for all available data to be captured and made available. Ultimately it is the manufacturers who make the decisions, assisted by software tools.

Inventory Control and Supply Chain Management Striking the optimum balance between the arrival of raw materials and other components from suppliers, and the burn rate on the production line, ensures that the line is operating at peak efficiency and the inventory of locally held stock is large enough to cope with unforeseen events while avoiding an unnecessarily high storage cost. Achieving and maintaining this balance relies on manufacturers having a detailed view of all the operations within the bounds of their plant as well as the activities of their suppliers. This is the same requirement highlighted above for the purposes of decision-making, which serves to emphasise the wider importance of this key IoT-enabled capability. Figure 7.9 depicts how the ubiquitous connectivity of the IoT facilitates inventory control and supply chain management across the whole production line, including external suppliers and other agencies. Consider a use case in which a raw material supplier runs their internal operations with the aid of IoT technologies. This is depicted in a highly abbreviated form at the left side of Fig. 7.9. Data from the sensors and other IoT-enabled devices and machines on the supplier’s production line is captured and displayed in a form that visualises the flow of the line and enables effective decision-making. The firewall ensures that all data sharing between the various IoT-enabled devices and machines is restricted to within the supplier’s plant and that only approved data is made available to other organisations via the Internet. Next, consider that the supplier does not have their own transport capability and instead they contract an external agency. That agency will have their own internal systems to log new orders arriving from their clients and those that have been delivered, as well as to track the progress and location of their delivery vehicles. IoT-­ enabled devices monitor these activities, the data from which is gathered and visualised as described before. An approved subset of this data is shared with the materials supplier and product manufacturer according to the policies defined by the firewall.

106

D. Heatley and M. Abdel-Maguid

Fig. 7.9  Inventory control and supply chain management

The IoT gives the product manufacturer in Fig. 7.9 access to the data made available by their supplier and the transport agencies involved in delivering the supplies as well as the final product to customers. That data is combined with the manufacturer’s internal data from the production line to display the status of the whole production line to the decision-makers, including an inventory of the current and predicted stock levels. Each vertical column in Fig.  7.9 is intentionally depicted to contain the same generic horizontal structure: the activity layer, the instrumentation layer, the flow control layer in which the status of the activities and inventory of the stock is visualised to the decision-makers and the firewall layer which manages the data sharing policies. This generic structure can be used to represent the role and functionality of any participant in a manufacturing chain. Additional columns can be readily integrated into the existing structure to extend the scope of suppliers and other organisations that the manufacturer might require in order to respond to new business opportunities. This highlights the intrinsic scalability of the IoT, which in turn attests to the sustainability it brings to the business activities of the organisations involved.

Predictive Maintenance Predictive maintenance is a further compelling opportunity afforded by the IoT. Sensors within the IoT-enabled machines and devices deployed throughout the production line will periodically notify the plant-wide central control system about the condition of the machines and devices from a maintenance perspective. The sending of these notifications can be preprogrammed in the individual machines, or they could be initiated by commands issued from the central control system. In either case the intent is to inform the manufacturer on a periodic pre-planned basis of the condition of the machines and devices employed on the production line and

7  The Internet of Things and Sustainable Manufacturing

107

whether a routine scheduled inspection is imminent. This differs from the information presented to the manufacturer for the purposes of decision-making, which is concerned more with unplanned changes that can occur at any instant. The real power of predictive maintenance is in alerting the manufacturer to a catastrophic event, such as an equipment failure, before it has happened, enabling pre-emptive remedial action to be taken. By continuously archiving the data from the IoT-enabled devices deployed throughout the production line, a historical perspective of the throughput of the line is built up over time, including instances when the throughput was disrupted and the reasons why. Software algorithms can be trained to recognise the signature in the data that usually precedes a disruption, for example, an unusual energy pattern when a machine tool becomes unexpectedly blunt ahead of its routine replacement period. When such a signature is detected, the manufacturer is promptly alerted. Predictive maintenance in one form or another has always been embedded in the culture of manufacture. The additional benefits delivered by the IoT that make today’s class of predictive maintenance so compelling include: • Breadth of view. Manufacturers can monitor the real-time status of the whole production line, including the movement of materials from suppliers who are at different geographic locations and the delivery of product to end users that is undertaken by independent delivery agents. Uniquely, every activity and dependency across the whole production line can be monitored. This enables predictive maintenance to be applied to the whole production line, not just the portion that is under the direct control of the manufacturer. • Speed of response. The manufacturer is alerted within moments of an unplanned catastrophic event occurring or a failure that is predicted to imminently occur with a high probability. • AI support. Software algorithms can assist the manufacturer to interpret the data and extract valuable meaning that can be promptly acted upon. • Minimal down time. The above attributes are key to minimising the duration, impact and cost of any down time, be it planned or unplanned. The IoT is uniquely equipped to reduce down time to a level previously unattainable. • Profitability. Ultimately the IoT gives manufacturers the means to increase business profitability and the productivity of machines as well as the workforce, by streamlining production processes and automating plant machinery. • Sustainability. The above attributes are key to the sustainability of manufacture going forward.

Resource Utilisation IoT-enabled production lines give the host organisation the means to monitor the utilisation of individual machines on the lines and their energy consumption. An analysis of this data reveals where there might be pinch points on the line, where throughput is slowed because insufficient machines are installed at that part of the line, or the

108

D. Heatley and M. Abdel-Maguid

operating speed of machines at an earlier part of the line is set too high which is causing too much product to reach the machines at the pinch point. Conversely, an analysis of this data might reveal that too many machines have been installed at a certain point on the line, some of which are very rarely used yet they are still constantly drawing power and hence are needlessly expending energy. In either case, this insight enables decision-makers to implement informed changes to the production line that ensure the best throughput while maximising the utilisation of all machines on the line and managing their energy consumption more efficiently. This in turn maximises production efficiency while minimising the running costs of the line.

Cloud Computing and Big Data Traditionally in manufacturing, data about the status of a production line is stored and analysed on local IT facilities in the host organisation’s own premises. Consequently, the onus is on the host organisation to provide and maintain these facilities and implement appropriate backup and data security measures. If that organisation wishes to benefit from the IoT, they are responsible for implementing upgrades and expanding their IT facilities in order to cope with the potentially significant volume of additional data that will be produced that needs to be stored and analysed. All of this incurs a cost that traditionally is entirely met by the host organisation. Cloud computing enables IoT-enabled manufacturers to manage their operations and running costs more effectively by off-siting their data storage and computational analysis requirements to IT facilities held in the Cloud. In addition, big data has become synonymous with the ability to identify significant trends and patterns embedded in vast volumes of unstructured data, which is the basis of decision-­making, predictive maintenance and resource utilisation described earlier. The wider benefits of cloud computing and big data to IoT-enabled manufacture include: • An IoT-enabled manufacturer that requires additional or less data storage capacity and analysis computational power for a period of time needs only to access what is required from the Cloud. Fluctuations in the manufacturer’s storage and analysis requirements are readily accommodated, which minimises the cost of these provisions to the manufacturer. • Space on the shop floor is prioritised for production line equipment since data storage and plant-wide computational facilities are largely absent. This enables the upper ceiling on productivity relative to floor space to be more readily achieved. • Powerful software tools and analytical platforms are hosted in the Cloud and regularly updated as new versions are released. IoT-enabled manufacturers have access to facilities such as these rather than having to purchase their own licences and cover the cost of updates. • Best in class data security enables IoT-enabled organisations to protect their own data as well as specify the data sharing policy they wish to have with other IoT-­

7  The Internet of Things and Sustainable Manufacturing

109

enabled organisations that contribute to production lines (e.g. material suppliers and transport agencies). • Data backups, software upgrades and system maintenance are managed by the Cloud operator(s) and included in the service that the manufacturer has signed up for. • Ubiquitous presence. Cloud computing gives all IoT-enabled organisations a presence in the Cloud regardless of their location. • Responsivity. Data sharing connections between IoT-enabled organisations can be quickly created to respond to new business opportunities as soon as they arise. Through these attributes and others not listed, cloud computing and big data are key to the sustainability of manufacturing by enabling IoT-enabled manufacturers to increase their productivity, reduce costs and rapidly respond to changes and new business opportunities that fundamentally influence their long-term prosperity. Sustainability is considered more widely in the next section.

Sustainability The sustainability credentials of the IoT in manufacture have been highlighted at several points thus far in this chapter. They can be grouped in two main categories: business and environment. Business: the sustainability credentials of the IoT in a business context include: • Growing the profitability of a business and safe guarding its longevity. This is achieved through the IoT enabling production lines to be streamlined, productivity increased, wastage minimised and product quality kept consistently high. • The almost limitless scalability of accessible storage and analytical resources in the Cloud enables manufacturers to respond to new business opportunities and continually grow their business. • Predictive maintenance minimises the impact of equipment failures and other production line issues by using extensive data gathered over time from IoT-­ enabled devices to identify trends that normally precede a failure of some sort. Pre-emptive remedial action can then be taken when such trends are detected in the data, thereby minimising or even completely avoiding production down time. • Using the introduction of IoT-enabled technologies in the work place to grow the skill base of the workforce. This in turn increases the qualifications held by individuals and improves their wider employability. Manufacturers who can offer these opportunities to their workforce increase their commitment to their employees and their appeal to potential new recruits, all of which contributes to the growth and sustainability of a skilled population. • IoT-enabled production lines can respond quickly and more flexibly to customers’ specific requirements, particularly requests for customisation. This encourages a closer engagement between manufacturers and customers which can lead to a long-term, mutually beneficial relationship between the parties. This in turn increases the sustainability of the manufacturers’ business.

110

D. Heatley and M. Abdel-Maguid

Environment: the sustainability credentials of the IoT in an environmental context include: • Minimising the wastage of raw materials on production lines and their energy expenditure by using IoT-enabled devices to constantly monitor the status of the lines and energy usage and alert decision-makers when the measures of these activities fall outside the acceptable range. • Tracking the disposal of materials that have specialist disposal requirements such as harmful chemicals and clinical waste. Guarding against improper disposal activities and enabling corrective action to be taken when such activities are discovered. • Continually monitoring activities that have a large carbon footprint and contribute to atmospheric pollution, such as road traffic, power generation, steel works and manufacture. Data on these activities is captured and analysed by a range of government bodies that are empowered to take action if the measures fall outside the acceptable range. The breadth of the areas listed above speaks to the scale of the influence that the IoT has on sustainability.

The IoT in Manufacturing: The Challenges Every new technology has to overcome its share of hurdles that often inhibit widespread adoption. The IoT is no different. The typical challenges faced by the IoT are not just limited to technical or financial issues, such as the cost of sensor devices and instrumented machines, or the investment in network and computing infrastructure. The progress of the IoT is also influenced by the complexities associated with governance, security, interoperability, privacy and standardisation. Moreover, behavioural and organisational factors such as outdated mind-sets, budget constraints, cultural change, day-to-day decision pressures, changing business priorities and whether there is an appetite for risk also play a significant role in promoting or hindering the widespread adoption of the IoT. There is however a way forward for early adopters. Business and technology leaders who invest in research and intelligence and learn by interacting with key IoT experts, investors and startups can identify and pursue opportunities to drive up customer experience, optimise costs and grow profits, all while outpacing market competitors. Some of the challenges faced are considered below.

Security and Privacy The current lack of comprehensive network security protocols has left every connected IoT device potentially exposed to cyberattacks, data-breach threats and identity theft vulnerabilities. The news media is full of instances of hacks and data

7  The Internet of Things and Sustainable Manufacturing

111

breaches, the seriousness of which is on the rise. Inadequate security of IoT devices and networks is the most pressing challenge faced by the IoT industry as it continues to compound the risk of data vulnerability for both businesses and individual consumers. Any IoT connected device or machine on a production line presents an opportunity for hackers to steal and share manufacturers’ confidential information. Even seemingly innocuous data such as the speed of a cutting blade at a particular point on the production line can provide valuable intelligence to a competitor. The value of the intelligence increases exponentially with the sophistication of the information contained within the hacked data. Hacking can also take over the control of devices connected to the IoT to, for example, alter the settings of machines on a production line to cause the resulting products to fail final tests or even to shut down the whole production line and lock out the local workers, preventing them from restarting the line. Such attacks could also raise major safety concerns for workers if the effected equipment controls very high temperatures in the production process, such as in steel production, or very high voltages and other high-risk factors. In addition to installing robust data security and anti-hacking measures within the Internet and IoT platforms, steps must also be taken to embed suitable protection within the individual devices and machines on the production line that are ­connected to the IoT. Furthermore, these measures must be continually updated to stay ahead of hackers.

High Implementation Costs The global adoption of the IoT in a business to business (B2B) environment is impacted by the high cost of implementation associated with IoT-enabled products and solutions. For instance, a manufacturing company with multiple plants, workflows and varying equipment types, seeking to modernise its operational infrastructure for the IoT, will have to consider upgrading its legacy infrastructure and systems to achieve a fully standardised and interoperable IoT environment. Policymakers, such as industry-facing standards associations and governments, have a significant role to play in addressing this issue by drafting key regulations and standards that reduce costs and drive market adoption.

Adaptability and Interoperability Interoperability is a fundamental attribute of IoT-enabled machines and devices. All IoT devices and platforms must also be highly adaptable to cater for the widest possible range of applications. For example, in order to implement an IoT solution in a warehouse to track products, monitor inventory and map disparate delivery locations, it is essential that the platform is interoperable to ensure that the various

112

D. Heatley and M. Abdel-Maguid

logistics and warehouse management systems can communicate with each other and share their information. Some standards are already in place (MTConnect 2018; OPC Foundation 2018), and wider progress is being made by the IEEE (IEEE Standards Association 2018), the Industrial Internet Consortium (2018) and Open Connectivity Foundation (2018). Nevertheless, manufacturers and companies across the industrial space are understandably hesitant at this early stage to bear the cost and business risk of replacing existing equipment to accommodate an interoperable IoT world while the development of standards is still under way.

Compatibility and Longevity Although IoT-enabled machines and devices have been a technological reality for some years, their adoption in manufacturing has been relatively slow and limited in scope. Factors include the additional time required for integration, the cost of upgrading legacy equipment and retraining staff. Manufacturers are uncertain when forming a view of the total cost structure and revenue potential of their IoT implementations and meaningfully comparing that with what they currently have. Will the perceived benefits be achieved in reality? The confusion created by standards that are still emerging and might be subject to more changes has further exacerbated things. Consequently, manufacturers have delayed making the big decisions and committing to the necessary investments in IoT.

The Role of the IoT in Industry 4.0 The Industry 4.0 initiative (Rostetter et al. 2016), introduced in 2011 by the German Federal Government, provides a framework that enables manufacturers to achieve previously unattainable efficiencies in operations, productivity, energy savings and more, all with sustainability in mind. These objectives are achieved firstly by mining the data from all kinds of network-enabled devices to extract meaningful information about the status of production lines and other key assets. That information is then used to inform decisions aimed at maximising the throughput of the production lines and the entire supply chain while attending to alerts caused by delays in the delivery of raw materials, machine failures and so on. These activities exactly map on to the capabilities of the IoT as described thus far in this chapter. The importance of the IoT in Industry 4.0 cannot be overstated. It is a fundamental prerequisite to the realisation of Industry 4.0 (Werr 2015), as is big data and cloud computing. It provides a base for what is often referred to as the digital factory (sometimes the Smart Factory), which is a key evolutionary stage on the path to Industry 4.0. A digital factory runs and monitors its own processes using IoT-­ enabled technologies with little if any human intervention, and where applicable it

7  The Internet of Things and Sustainable Manufacturing

113

autonomously takes action in response to alerts that impact on these processes. In addition, a digital factory exploits the flexibility and adaptability of the IoT to facilitate new manufacturing practices that are more responsive to changes in customer requirements and the need to offer a higher degree of product customisation than ever before. Indeed, whereas product customisation has been almost exclusively the domain of small specialist businesses that can offset the high tooling cost of their small production runs with a premium price to their highly selective customers, the IoT enables digital factories that might normally only undertake large-scale runs of a product with a fixed design to also produce customisable products on a large scale as well as in quantities as few as 1, all at a competitive price. As a consequence these digital factories are able to engage with a far wider range of customers than has been previously possible, from individuals and small enterprises who need a small number of components to construct a single prototype of what might become their next product up to airplane manufacturers who wish to subcontract the fabrication of a new design of landing gear to upgrade an entire fleet. All of this can be accommodated by digital factories, enabled by the IoT. Examples like these give a sample of the future that Industry 4.0 promises to deliver, enabled by the IoT, cloud computing and big data.

Outlook In less than 20 years, the IoT has evolved to a level that is beginning to have a major impact in manufacture and other sectors on a global scale. The functionality that can be achieved by sharing data between IoT-enabled devices far exceeds what has been feasible to date. The opportunities for manufacturers that use the IoT and the sustainability it brings to their businesses are potentially profound and will redefine the way in which they work. For these organisations to fully capitalise on these opportunities, it is essential that industry leaders come together to influence, promote and embrace how the IoT is reinventing the industry. This is especially important in the age of Industry 4.0 to ensure that the growth of IoT-enabled devices, cloud computing and big data is optimally co-ordinated to deliver the best overall outcome for all interested parties. Interestingly, while the IoT concept is still relatively recent and its rollout is at an early stage, the IoT will continue to have a presence in the social consciousness, as well as in the media and technical press for some years to come. However, as the emergence of Industry 4.0 gathers pace and IoT deployments in the manufacturing sector become common, the IoT along with cloud computing and big data might become viewed as mere enablers for Industry 4.0 and hence might fade into the background. Industry 4.0 might then become the common vernacular for all things involving the ubiquitous sharing of information, big data and cloud computing in manufacture.

114

D. Heatley and M. Abdel-Maguid

Bibliography The online material listed below was successfully accessed during March 2019. Ashton, K. (2009). That ‘Internet of Things’ thing. Available online at: https://www.rfidjournal. com/articles/view?4986#back-from-modal Ashton, K. (2015) Peter Day’s World of Business. Available online at: http://downloads.bbc.co.uk/ podcasts/radio/worldbiz/worldbiz_20150319-0730a.mp3 Bassi, A., et al. (2013). Enabling things to talk – Designing IoT solutions with the IoT architectural reference model. Springer. BCS. (2013). The societal impact of the Internet of Things. BCS, Springer. Brynjolfsson, E., & Hitt, L.  M. (1998). Beyond the productivity paradox  – Computers are the catalyst for bigger changes. Communications of the ACM, 41(8), 49–55. Forbes. (2017) 2017 Roundup of Internet of Things forecasts. Available online at: https://www. forbes.com/sites/louiscolumbus/2017/12/10/2017-roundup-of-internet-of-things-forecasts/#269994481480 Forward Geek. (2016) Mobile app development market share, CAGR growth, Expected Revenue Structure Forecast to 2022. Available online at: https://www.forwardgeek.com/article/ Mobile-App-Development-Market-Share-CAGR-growth-Expected-Revenue-StructureForecast-to-2022-20161115 Growth Enabler. (2017) Market pulse report, Internet of Things (IoT). Available online at: https:// growthenabler.com/flipbook/pdf/IOT%20Report.pdf IEEE Standards Association. (2018) Internet of Things. Available online at: https://standards.ieee. org/initiatives/iot/stds.html IIC. (2018). Industrial internet consortium. Available online at: https://www.iiconsortium.org/ index.htm McKinsey & Company. (2015). The Internet of Things  – Mapping the value beyond the hype. Available online at: https://www.mckinsey.com/~/media/McKinsey/Business%20Functions/ McKinsey%20Digital/Our%20Insights/The%20Internet%20of%20Things%20The%20 value%20of%20digitizing%20the%20physical%20world/Unlocking_the_potential_of_the_ Internet_of_Things_Executive_summary.ashx MTConnect. (2018). MTConnect  – A free, open standard for the factory. Available online at: https://www.mtconnect.org/ NPR. (2012). Which is greater, the number of sand grains on earth or stars in the sky?. Available online at: https://www.npr.org/sections/krulwich/2012/09/17/161096233/ which-is-greater-the-number-of-sand-grains-on-earth-or-stars-in-the-sky OCF. (2018). Open connectivity foundation. Available online at: https://openconnectivity.org/ OPC Foundation. (2018). The OPC unified architecture. Available online at: https://opcfoundation. org/about/opc-technologies/opc-ua/ Pazvakavambwa, R. (2018). IoT helps transport sector reduce carbon footprint. Available online at: https://www.itweb.co.za/content/lLn14Mmy3n8MJ6Aa Rostetter, C., Khoshafian, S., & Adams, K. (2016). The adaptive digital factory. Available online at: https://www1.pega.com/system/files/resources/2018-04/proucts-manufacturing-digitalfactory-whitepaper.pdf Statistica. (2018a). Internet of Things (IoT) connected devices installed base worldwide from 2015 to 2025. Available online at: https://www.statista.com/statistics/471264/ iot-number-of-connected-devices-worldwide/ Statistica. (2018b). Number of Machine-to-Machine (M2M) connections worldwide from 2014 to 2021. Available online at: https://www.statista.com/statistics/487280/ global-m2m-connections/ Statistica (2018c). Projected global Internet of Things enabled sensors market in 2022, by ­ segment. Available online at: https://www.statista.com/statistics/480114/ global-internet-of-things-enabled-sensors-market-size-by-segment/

7  The Internet of Things and Sustainable Manufacturing

115

Stroud, R.  E. (2017). IoT opportunities, strengths and momentum. Available online at: http:// isacasfl.org/wp-content/uploads/wow/presentations/6.%20Robert-Stroud%20-%20IoT_ Overview%20-%20ISACA-South-Florida-Chapter.pdf Variant Market Research. (2017). Automotive telematics market overview. Available online at: https://www.variantmarketresearch.com/report-categories/automotive/ automotive-telematics-market WEF. (2015). Industrial Internet of Things: Unleashing the potential of connected products and services. World Economic Forum report, January 2015. Available online at http://reports.weforum.org/industrial-internet-of-things/ Werr, P. (2015). How Industry 4.0 and the Internet of Things are connected. Available online at: https://www.iotevolutionworld.com/m2m/articles/401292-how-industry-40-the-internetthings-connected.htm Wikipedia. (2018). IPv6 deployment. Available online at: https://en.wikipedia.org/wiki/ IPv6_deployment Woolhouse, A. (2016). The great IoT connectivity race. Available online at: https://community. arm.com/iot/b/blog/posts/the-great-iot-connectivity-race

Chapter 8

Security Challenges in the Industry 4.0 Era Mohammed M. Alani and Mohamed Alloghani

Introduction The fourth industrial revolution, commonly referred to as Industry 4.0, is a trend that is changing the way industry works. Where computers and systems automation are thought of as the third industrial revolution, a group of new technologies represent Industry 4.0. These technologies include, but are not limited to, cyber-physical systems, Internet of Things (IoT), cognitive computing and artificial intelligence (AI), and cloud computing. It is indeed a revolution because it is changing how different industries operate, introduces entirely new industries, and make other industries obsolete. Let’s look at an example of the impact of these changes. As the year 1886 is mainly considered the birth year of the first automobile, when Karl Benz registered his patent for a moving vehicle. The whole industry for over 100 years relied on fairly similar concepts of internal combustion engines and gasoline fuel. Now in the year 2018, electric cars are gaining popularity. According to the International Energy Agency, over three million electric vehicles are now operating around the world, with the expectation of the number to grow to over 200 million vehicles in 2030 (Cazzola et al. 2018). However, the automotive industry is not our example. Our example is actually the taxi industry. The taxi industry started declining since the birth of services like Uber, Careem, and Lyft. These services have caused severe decline in the industry of taxi driving. These services were not possible without the developments in different areas of technology like networking, mobile

M. M. Alani (*) Khawarizmi International College, Abu Dhabi, UAE e-mail: [email protected] M. Alloghani Liverpool John Moores University, Liverpool, UK e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_8

117

118

M. M. Alani and M. Alloghani

Internet connectivity, Global Positioning Systems (GPS), cloud computing, and distributed database management, just to name a few. Another hit to the taxi driving industry came soon after with the emergence of self-driving vehicles. Self-­driving cars are becoming a reality due to developments in many technologies like networking, big data, artificial intelligence, and sensor networks, among many others. One of the earliest applications of self-driving vehicles are self-driving taxis. With the technology starting to be actually implemented in cities like Dubai, the whole taxi driving industry might come to an end, in the near future. However, these developments are not all negative to the job market. Many new jobs are available now that were not even heard of 10 years ago. Areas like big data analytics, and cloud computing networking, among many other new areas are in high demand for specialists. The following subsections will give a brief definition of the technologies that comprise the fourth industrial revolution.

Cyber-Physical Systems Baheti and Gill define cyber-physical system in (Baheti and Gill 2011) as “a new generation of systems with integrated computational and physical capabilities that can interact with humans through many new modalities.” In 2009, Wolf mentioned the term “cyber-physical systems” and stated that you will be hearing the name more in the coming few years (Wolf 2009). Wolf also states that cyber-physical systems are the next step because they provide improved performance and efficiency. In a simple definition, a cyber-physical system is a system that combines physical components with computer-based systems and algorithms. The physical component is controlled or monitored by the cyber component. An example of that is an autonomous vehicle. The autonomous vehicle takes physical inputs from various types of sensors and feeds that input into a cyber system that controls the vehicle physical parts that in turn controls the vehicle motion.

Internet of Things The IoT is the connections of physical things to the Internet in a way that provides the capability to read sensors data and control the physical thing from distance (Kopetz 2011). The idea of controlling physical systems using computers is as old as computers. However, the main development that IoT brought as compared to embedded systems is the capability to access sensors’ data and control systems remotely. The first Internet-connected smart device goes back to the year 1982 when a group of researchers at Carnegie Mellon connected a Coke vending machine to the Internet (The only coke machine on the internet 2018). The machine was altered to share information like inventory and whether a newly brought Coke is cold or not. In 1991, Mark Weiser from Xerox labs introduced the concept of ubiquitous computing and identified the computers of the twenty-first century as “Specialized

8  Security Challenges in the Industry 4.0 Era

119

elements of hardware and software, connected by wires, radio waves and infrared, will be so ubiquitous that no one will notice their presence” (The only coke machine on the internet 2018). IoT can be thought of as the bridging of all of the developments mentioned above. IoT enables devices like home appliances, vehicles, and other devices to network and exchange data. Hsu and Lin stated that the IoT market is on its way to hit $7.1 trillion by 2020 (Hsu and Lin 2016). This indicates that IoT applications are becoming more and more acceptable publicly and more businesses are adopting IoT-based solutions.

Cognitive Computing As per John Kelly in (2015), the term cognitive computing includes a combination of artificial intelligence and technologies of signal processing. This includes natural language processing, speech recognition, human-computer interaction, machine learning, and reasoning, along with other technologies. These technologies can be briefly identified as: • Natural language processing is “a theoretically motivated range of computational techniques for analyzing and representing naturally occurring texts at one or more levels of linguistic analysis for the purpose of achieving human-like language processing for a range of tasks or applications” as identified in (Liddy 2001). • Speech recognition is the technology used to capture and analyze human speech so that it becomes recognizable by computer-based systems (Graves et al. 2013). • Human-computer interaction deals with the study of designing computer systems that are interactive, productive, and user-friendly (Smith-Atakan 2006). • Machine learning is “a set of methods that can automatically detect patterns in data, and then use the uncovered patterns to predict future data, or to perform other kinds of decision making under uncertainty” (Murphy 2012). • Reasoning involves inference, planning, and learning to have a system that can be used in making decision (Hertzberg and Chatila 2008). In summary, these technologies help in simulating, emulating, translating, and predicting human thinking, interaction, and behavior. Although many of these technologies were available a long time ago, with the rapidly increasing popularity of mobile devices, these technologies have found a much larger scope of applications since the introduction of the term cognitive computing back in 1993 (Johnson 1993). Applications of these multiple technologies emerged in a variety of areas like health, education, entertainment, military, etc. The most known example of cognitive computing is IBM Watson platform (2018). The company defines it as “AI for Business.” It is designed in such a way that can mimic human cognitive capabilities and make decision in a similar way. The main difference is that such systems are capable of processing huge amounts of data in relatively short period of time compared to humans.

120

M. M. Alani and M. Alloghani

Cloud Computing The main concept of cloud computing, although named differently, goes back to the year 1961 (Erl et al. 2013). In that year, John McCarthy stated, at the MIT Centennial “computers of the kind I have advocated become the computers of the future, then computing may someday be organized as a public utility just as the telephone system is a public utility... The computer utility could become the basis of a new and important industry.” The initial term introduced then was utility computing, which referred to a computer- on-demand service that can be used publicly with a pay-for-what-you-use model. Before the end of the 1990s, Salesforce.com introduced the first remotely provisioned service to the enterprise. Near the end of the 1990s, the concept started to take a different direction. The main focus was on presenting an abstraction layer used to facilitate data delivery methods in packet-switched heterogeneous networks. In 2002, Amazon.com introduced Amazon Web Services (AWS) platform. The platform, back then, provided remotely provisioned computing resources and storage. In its commercial sense, the term cloud computing emerged in 2006 when Amazon launched its Elastic Compute Cloud (EC2) services. The service model was based on “leasing” elastic computing processing power and storage where enterprises can run their apps. Later that year Google also started providing Google Apps. Cloud computing was identified by NIST in (Mell and Grance 2011) as: a model for enabling ubiquitous, convenient, on-demand network access to a shared pool of configurable computing resources (e.g., networks, servers, storage, applications, and services) that can be rapidly provisioned and released with minimal management effort or service provider interaction.

In its simplest sense, cloud computing is dynamic provisioning of computing resource where customers are charged per use, similar to the idea of electricity or tap water services.

Industry 4.0 Components According to Gubán and Kovács (2017) and Zezulka et  al. (2016), Industry 4.0 integrates three or more mutually interconnected factors including digitization and integration of simple technical-economical networks transforming them to complex ones, product and service digitization, and implementation and integration of new market models. The three comprise of different technologies and concepts, which are generalized as components of Industry 4.0. Zezulka et  al. suggested that the components of Industry 4.0 are based on either Reference Architecture Model Industry 4.0 (RAMI 4.0) or Industry 4.0 Components models (Zezulka et al. 2016). The RAMI 4.0 model presents a three-dimensional conceptualization of the interconnection of technical-economical features. In general, RAMI 4.0 states that Industry 4.0 consists of business, functional, information, communication,

8  Security Challenges in the Industry 4.0 Era

121

integration, and asset layers of development and maintenance or usage type and meant to improve production instance (Zezulka et  al. 2016). Such a system also considers the connection to the rest of the world, enterprise operations and prospects, work units and stations, and control of devices. However, the elements identified in the RAMI 4.0 model can be categorized into big data, Smart Factory, Cyber-Physical Production Systems (CPPSs), and IoT technologies and concepts. An overall visualization of the components of Industry 4.0 including the three major components is shown in Fig.  8.1. The four factors explicate the core roles of Industry 4.0 and the likely business applications although the platform generally intends to automate production processes. As per Fig. 8.1, big-data-related technologies serve as the interface between the three primary functions and the different applications of Industry 4.0 concepts. Big data encompass data generated from within and outside the business environment, and as a technology, it might involve machine learning and other data analytics techniques geared toward business intelligence as well as security measures against fraudulent activities (Yan et al. 2017; Lee et al. 2014). The CPPS can also be an active part of big data analytics because it focuses on all cyber computational techniques and communication technologies in the context of people and devices. The CPPS integrates computation, networking tools, and technologies and physical processes that execute commands to interpret and track results (Herterich et al. 2015). The concept of Smart Factory as a component of Industry 4.0 arises from IoT interconnectivity inside the factory and the resultant network application and process alignments (Sadeghi et al. 2015). It is imperative to note that a Smart Factory consists of several CPPSs and it is also important to note that Smart Factory

Fig. 8.1  The components of Industry 4.0 including the three major functional factors (Geissbauer et al. 2016)

122

M. M. Alani and M. Alloghani

presents a large number of attack surfaces. An attack surface can be identified as the area of a system that can be subject to an attack. Whenever attack surfaces increase, threats increase as well and become less manageable (Alani 2016). The concept of Smart Factory revolves around connection and integration of industrial devices and processes so that they support automation of either productive or manufacturing tasks. Smart Factory is associated with restructuring of production systems based on emerging technologies of computer integrated enterprise, manufacturing, and automation techniques that lower costs of production, shortens product life cycle, and prioritizes improvement of product quality (Balador et al. 2017). It is paramount to note that Smart Factory disregards market boundaries and as such it leverages Internet-based technologies to develop new structures for production systems (Zawadzki and Żywicki 2016). It is critical to note that new production systems and structures are adaptive and distributive with a mandatory overlay of automation of all processes and subsystems. On the other hand, the concept of big data, as explained in the previous section, has close association with sensors and devices that collect and store data over the platform. Besides the sensors, new Industry 4.0 technologies include developments in augmented reality (AR) devices and wearables that can be used to collect data in the healthcare sector among other experimentally based industries (Balador et al. 2017). However, big data’s most noticeable application area is data analytics and its specific applications in business intelligence and fraud detection (Fernández-Caramés and Fraga-Lamas 2018). Big data analytics can also contribute to process discovery and product life cycle improvement besides being critical to fraud detection and prevention efforts. Regarding the human-technology interaction in the platform, the CPPS accounts for physical interaction between business processes and physical and mechanical systems as illustrated in the next section. Arguably, the CPPS is the basis for the new production era as it involves integration of wireless control systems with sensors and intelligent systems that control the production lines (Cassandras 2016). In general, cybernetic systems are elements of the CPPS that accounts for the sensors and logic units that are implemented alongside actuator units. The CPPS interact with the physical world using computing capabilities and communication tools e­ mbedded with control mechanisms that include feedback loop interactions and process monitoring (Chaâri et al. 2016). It is imperative to note that CPPS plays a vital role in the collection and storage of big data because it contains the sensor technologies that capture and transform physical data into signals. Ideally, the CPPS consist of the integrated computers, networks, and physical systems. The rest of the devices that do not fit within the category of CPPS are considered as part of the IoT. The concept of IoT in Smart Factory promotes practices related to quality control, production efficiency, throughput, and speed of production because the devices support realtime tracking of milestone achievements regardless of the location of the network servers (Botta et al. 2016). The importance of IoT on a smart factory floor includes supporting real-time evaluation of processes based on automated decision-making approaches as opposed to human-based approaches. However, for IoT to support such automations, it requires historical data. This might include operating in data-­ collection mode before the full automation of the factory.

8  Security Challenges in the Industry 4.0 Era

123

Interoperability is a concept of Industry 4.0 that results from integrating the previously mentioned components together. It covers the essence of considering big data, Smart Factory, CPPS, and IoT as the components of Industry 4.0, and its mandates cover the scope of Industry 4.0. The concept of interoperability focuses on the minimization of the errors associated with transmission and translation of data and information (Bahri et al. 2012). In specific, interoperability ensures efficiency, reliability, and accuracy of information meant to further digitization of product and services, development of digital business models and customer access services, and digitization and integration of both vertical and horizontal value chains. The components mentioned in this section are far from being a comprehensive list of the Industry 4.0 components (Bahri et al. 2012).

Security Threats to Industry 4.0 As Waidner and Kasper (2016) posit, mistakes and errors involving people, infrastructure, equipment, and machines within the CPPS have operational safety implications, and authors identify “wide attack surface of CPPS” as the amalgamated major threat to Industry 4.0. The CPPS and the threat surfaces are as summarized in Fig. 8.2. From Fig. 8.2, the major sources of threats include electronics, operating systems, and other third-party software, networking, machines, and people. The vulnerabilities associated with the threats include physical attack and cyberattacks Social engineering, Phishing, ...

Eavesdropping, man-in-the-middle, denial of service, ... Cyber stack

Humans Engineers, operators, ...

Machines Computers, CPS, cloud, ...

Networking

Software

Ethernet, WiFi, ...

OS, applications, ...

Electronics CPUs, microcontrollers, actuators, sensors, ...

Physical stack

Runtime-attacks, reverse engineering malware, ... Side-channel attacks, reverse engineering, invasive attacks, ...

Physics Mechanics, physical principles, ...

Energy, matter, ...

Production process

Product

Cyberphysical production system (CPPS)

Fig. 8.2  The CPPS architecture showing attack surfaces as applicable to Industry 4.0 (Sadeghi et al. 2015)

124

M. M. Alani and M. Alloghani

including man-in-the-middle, man-in-the-browser, and social engineering and phishing attacks (Sadeghi et al. 2015). As specified in Fig. 8.2, the attacks are applicable to specific parts of the CPPS. In cloud computing, especially in the context of big data, it is increasingly becoming difficult to secure data transmission between physical and virtual assets, and authentication is even challenging when disseminating information over multiple vulnerable interconnected devices (Yan et al. 2017). Some of the threats to availability, confidentiality, integrity, and authentication of big data include wide attack surface, availability of global data in the local context, user behavior manipulation, and complexity and misconfiguration. Threats posed by having a wide surface attack are associated with the single point of failure, and most of the reports have cited denial of service (DoS) as the leading mode of attack. Further, exposure in the layer handling sensors has often led to information leakage affecting business processes in cloud computing environments (Antao et al. 2018). Finally, in most cases, complexity and misconfiguration within CPPS, as well as other components of Industrial 4.0, encourages bigger scope attacks, most of which are attributed to privilege escalation. More importantly, all the components of Industry 4.0 have different attack surfaces although CPPS and IoT have shown many vulnerabilities compared to the rest of the components. Industry 4.0, just like any other group of integrated systems, faces both insider and external threats, especially given the complication and advancing protection-­ evasion techniques available in the cyber-sphere. Furthermore, the number of attack surfaces increases with the increasing number of Industry 4.0 technologies that an organization implements, and the subsequent unexpected number of attacks or points of attacks also depends on the number of extended systems and the attack surfaces unique to them. For instance, implementing IoT besides some other aspects of Smart Factory increases the number of vulnerable points, but interconnecting the two makes the process of vulnerability identification even more complex. This complicates attacks detection and prevention. The following subsections discuss different aspects of Industry 4.0 security threats.

Business Cyber-Espionage At the core of Industry 4.0 is the pursuit of smartness. Due to that, devices, machines, and services are integrated and interconnected with the aim of working seamlessly. For instance, in Fig. 8.2, the production processes are integrated with physical systems which in turn interacts with software and networking devices and the entire architecture. The same architecture supports different computer systems and workstations and connections to external users through the Internet. All the mentioned points have sensitive data or require service overlay to data centers and as such, act as potential points of attack. Given the current growth in the number of threat actors, especially those associated with advanced persistent threat (APT) groups, sensitive

8  Security Challenges in the Industry 4.0 Era

125

information contained within the difference department data centers are increasingly becoming exposed, and organizations risk losing trade secrets. Regardless of the motivation of the threat actor groups, Baena et al. assert that the levels of skills that cyberattackers have make it easier for them to target specific industries and organizations and extract sensitive information including intellectual properties (Baena et  al. 2017). The challenge with business cyber-espionage is the likely involvement of governments. For instance, Poindexter, in (2018), provides a detailed account of several incidences involving corporate espionage between the United States and China. In most of the cases, the US federal government through its cyber task forces and intelligence bodies has accused China of perpetrating multiple repeated intrusions into organizational networks aimed and stealing intellectual property as well as proprietary business information (Poindexter 2018). Industrial espionage affects companies in various ways like massive decline in sales and weaker financial position. In addition, the theft of sensitive information, particularly intellectual property of products and production processes, jeopardizes the competitive ability of the affected firm in the market. Corporate theft and industrial espionage continue to increase as many firms pursue more technologies to achieve the smart factory setup (Akaev et al. 2018). In the context of Industry 4.0 and given the rate at which organizations and companies across industries are adopting some of the concepts and technologies, the threats related to espionage and theft of confidential information are much greater because of the interaction of the partners and devices within the networks. Hence, it is crucial to establish security techniques and control mechanisms for ensuring transparency and trust within all types of Industry 4.0 platforms, and such measures should focus on the protection of intellectual property and confidential information (Deloitte industry 4.0 report 2018). It would be prudent for individual organizations to pursue some of the modern and advanced data protection and encryption technologies including the possibility of implementing quantum cryptography at production, manufacturing, and distribution levels.

Denial of Service A denial-of-service (DoS) attack is an attack that targets the availability of a system or service by flooding it with requests that renders the service unavailable for its legitimate users (Alani 2016). The specific type called distributed DoS (DDoS) is a type of DoS attacks in which the flooding requests come from multiple sources on the network simultaneously. In conventional DoS attacks, the perpetrator bombards the target server with enormous requests forcing it to commit all the available computing resources, and the input data forces the server to malfunction leading a crash. The technique can also be used as a supporting tool to eavesdropping in man-in-the-­ middle attacks. In general, the buffer overflow that occurs within the server might lead to disabling of security systems, especially network intrusion detection system (NIDS) and firewall (both hardware and software). Industry 4.0 depends on

126

M. M. Alani and M. Alloghani

interconnection and integration of many systems and processes, which creates environments that can be suitable for the launching of successful DoS attacks (Waslo et al. 2017). Furthermore, the increasing use of cloud computing-based techniques in manufacturing processes and smart factories will expose many organizations to DoS attacks if the cloud system is not well-designed to defend against these attacks. As organizations advance the applications, many other unknown vulnerabilities emerge (Waslo et  al. 2017). The consequences and impacts of DoS attacks can severely affect operations of the organization in the context of Industry 4.0 because of the use of sensors, and other automated machine parts can receive most of the impact. Besides the material damage, DoS impact can be costly because the damages may necessitate replacement of servers and sensors as well as forcing for redesigning of systems and implementing a different architecture to make it difficult for the same attack techniques to succeed again. Nonetheless, a manufacturing or production plant that has invested in Industry 4.0 technologies can endure interruption of services, implementation of complex protocols to restore services, and embarking on intense training on machine operations after the redesign. The primary challenge with DoS is that they are difficult to detect and as such their risks cannot be easily quantified and planned for and the constraint extends to creation of controls within the systems and processes to minimize the impacts.

Extension of Services and Systems It is fairly correct to say that Industry 4.0 encourages distribution and extension of services across the departments of an organization, and the approach has been regarded as the ultimate means of improving the efficiency of the supply chain (Rennung et al. 2016). According to Masoni et al. in (2017), Industry 4.0 and the concept of AR have made it easy for organizations across industries to maintain services and processes remotely. The authors also noted that the success of AR in industrial applications is ascribed to Industry 4.0 because the architecture of the application requires skilled operators and control rooms that are found in most integrated and distributed manufacturing and production settings. However, it has been noted the origin of vulnerabilities in the supply chains can be traced back to suppliers where possibilities of portal hacking, man-in-the-middle attacks, DoS, and side channel attacks that result from unencrypted connections and data transmission remain the major source of concern in integrated and extended systems used in Industry 4.0. The organizational structure adopted in Industry 4.0 also encourages the spread of any type of the threat once it finds its way through of a vulnerable point. In specific, the supply chain has proven infectious to other dependent systems. Liu and Xu in (2017) and Ustundag and Cevikcan in (2018) suggest that the transformation to Industry 4.0 requires plans for creating security awareness and developing control mechanisms and policies for authentication including encryption technologies and behavior-analysis tools for preventing hacking of supply chains and their dependent processes.

8  Security Challenges in the Industry 4.0 Era

127

Smart Factory and Security As Waslo et al. note in (2017), the transformation from Industry 3.0 to Industry 4.0 was associated with changes in technology and subsequent cyberthreats. For instance, Industry 3.0 relied on serial and ladder logic systems that depended on LAN, TCP/IP, and vendor scripting language that exposed the systems to threats such as DoS, system failure due to packet malfunction, and man-in-the-middle parsing false information to operators. Conversely, Industry 4.0 is mostly based on the industrial Internet and relies on ad hoc connections facilitating rigorous data collection and monitoring for improving the life cycle of the product (Waslo et al. 2017; Wang et al. 2016a, b). The application has been extended to market-based planning and supply chain management, and as a consequence of the approach used in the platform, organizations are exposed to malicious promotional campaigns driven by botnet farms, product purchase by botnets, and the extensive breach in product usage (Waslo et al. 2017; Pereira et al. 2017). All these threats are consequences of the new technologies that have been employed in Industry 4.0, and they are complex and more advanced in comparison to those associated with Industry 3.0.

Security Challenges for IoT-Enabled Manufacturing Industry 4.0 has made inspired intelligent manufacturing through automation of production processes besides other managerial applications during manufacturing. According to (Tupa et al. 2017), both IoT and CPPS are at the core of intelligent manufacturing, and they aid in manufacturing process management, planning of human resource and role allocation, and management of machine and manufacturing systems and technologies among other applications. The challenges experienced in IoT-enabled manufacturing are associated with this specification automation and managerial application areas. Meany (2017) terms such risks and challenges as functional and attribute to failures or mistakes in networks, robots, software, and semiconductors used in implementing automated and intelligent manufacturing systems (Stock and Seliger 2016; de Man and Strandhagen 2017). He et al. (2016). demonstrate the implementation of IoT infrastructure in manufacturing is associated with the elements of CPPS and its objective is to create synergy between the manufacturer and suppliers. The challenges experienced within IoT-enabled manufacturing environment can be summarized as follows. Figure 8.3 shows the different applications, business processes, service entities, IoT services, communication, and IoT devices in an elaborate cyberspace showing all the fundamental networking, hardware and software requirements. The flow of data involves collection using sensors, communication over the Internet, computation, and analysis, and actions at the service layers and the challenges can be experienced at any of the stages. Firstly, the data flow over the system has innumerable vulnerability points, and attackers can steal data at any of the points and without

128

M. M. Alani and M. Alloghani

Fig. 8.3  Challenges of IoT cyberspace in smart manufacturing (He et al. 2016)

appropriate protection measures (He et al. 2016). In specific, the attackers can gain physical or logical access to CPPS or programmable logic controllers within the IoT. Additionally, attackers can read and modify software information and, in the process, obtain pertinent information for counterfeiting as illustrated in Fig. 8.3. In some cases, it could also be possible to manipulate either the firmware or software with the aim of weakening the production protection mechanism. A successful creation of a counterfeit within any of the CPPS constituting the IoT can split the system into individual components and identify the user processes and components. Such information can be used to analyze the system and rebuilt the circuit or the targeted components thereby compromising the manufacturing system. Similar actions can lead to replication of original firmware. It is apparent that hardware is the ultimate line of protection against data espionage although, given the trends in BOYD and COYD besides poor legal frameworks, it is quite difficult to guarantee physical protection to the systems.

Security Challenges in Industry Control Systems Some of the devastating impacts of gaining control of the logical and physical systems within the IoT framework is that information can be manipulated without the responsible parties knowing (He et al. 2016). Under such circumstances, it is quite difficult to maintain the quality of service as well as the quality of the product. The IoT and its incalculable CPPS interfaces have increased the threat surface associated with it because attackers continue to improve evasive attacks that are largely

8  Security Challenges in the Industry 4.0 Era

129

passive in nature (Leitão et al. 2016). The major concern is that most manufacturers are adopting IP-based CPPS and they are further exposed to the different attacks illustrated in Fig. 8.3. More importantly, most manufacturers are making the change while still using the old and obsolete equipment. The movement exacerbates the problem because the traditional systems did not have some of the attack surfaces that are available IP-based CPPS, while implementation of the current industrial development platforms does not consider additional security features (Helo et al. 2014). Unfortunately, besides the control of physical and software systems, most intangible assets, especially customer data, are found within the IoT and CPPS, and as such corporate value of IoT-enabled manufacturers can drastically deteriorate in case of a successful attack (Zuehlke 2010).

Security of Industry 4.0 Versus Security of Legacy Systems Just like any other new technology, Industry 4.0 brings along new security challenges. What is unique about Industry 4.0 is that it is not a technology per se; rather it is a combination of various technologies. Each of these technologies represents an attack surface. Innovations in the emerging trends of Industry 4.0, IoT, AI, etc. are not going to stop any time soon. All of these innovations will bring new security challenges. Security threats have become a non-avoidable concern when it comes to developing new systems. Security threats to Industry 4.0 can be thought of as the combined threat of the underlying technologies like cloud computing, sensor networks, mobile ad hoc networks, and IoT in addition to new threats that are generated by interfacing all these technologies together. These interfaces that connect all the technologies together can be targeted by malicious attackers as well. A similar example can be found in cloud computing. Cloud computing is based on several technologies like grid computing, virtualization, and server operating systems. Threats to cloud computing go beyond the combined threats to the underlying technologies. As the interface between these technologies is server through an abstraction layer, usually referred to as the hypervisor, this new layer generates a new attack surface and has its own security threats. On the other hand, IoT also brings a whole new list of challenges. In (Shahzad et al. 2017), a thorough discussion of IoT security threats is presented with focus on different IoT applications like applications in oil fields and smart grids. These particular two examples are of high importance because they can have massive impact on people’s lives. The paper discusses different threats and attack scenarios from insider attacks and unauthorized access, to different types of attacks. Smart grids have been a target for many attacks in the past few years (Wang and Lu 2013; Liu et al. 2013; Aloul et al. 2012). Another one of the most challenging aspects of IoT is the exploitation of IoT devices to perform attacks on other systems, such as the attack on Dyn (IoT devices used in DDoS attacks 2018). The Internet service provider Dyn was subject to a

130

M. M. Alani and M. Alloghani

massive DDoS attack back in 2016. The attack on Dyn was particularly interesting because it was the first large-scale attack that employed a vulnerability in IoT devices and employed these devices to attack Dyn. This means that security challenges in Industry 4.0 are not limited to protecting the confidentiality, integrity, and availability of your systems, but also protect other systems from the harm that can be used by your systems for malicious purposes. In 2018, Boye, Kearney, and Josephs provided a detailed survey of industrial IoT cybersecurity threats (Boye et al. 2018). With their survey, they suggested a method of continuous security assessment. This method can be useful with rapidly developing technologies like IoT and other Industry 4.0 components. Another thorough survey of cybersecurity issues in industrial critical environments was presented by Ani et al. in (2017). The change in the way manufacturing processes are conducted, due to the gradual growth of Industry 4.0, requires that governments introduce suitable legislation to keep up with the change. Countries should not end up in situation where they are late in issuing laws to regulate the new technologies, like what happened with electronic commerce. It took some countries over 20 years to issue laws that regulate electronic commerce. This has caused misuse that led to people losing faith in new technologies, and this affected the acceptance and hindered the development of new technologies.

Knowledge Protection and Anti-piracy Protection Big data has several dimensions, and among them is velocity, which according to Fallis (Barwick 2012) refers to the speed at which the large-volume diverse data is transferred between the source and the applications. The required high speed and necessity to maintain the integrity and confidentiality of the information poses unique challenges during transfer and distribution. In specific, the knowledge embedded in smart products must be transferred carefully, and process knowledge, especially about production methods, must be treated as trade secrets and protected (Waidner and Kasper 2016). Some of the emerging concern is how organizations and firms will endeavor to their intellectual property and knowledge system within appropriate jurisprudence of data management. The concern, as Pereira et al. (2017b) note, is having an efficient framework for legitimizing copyrights when concepts such as Bring Your Own Device (BYOD) and Choose Your Own Device (CYOD) will expose CPS and other control systems within the smart factory abstract. The other concern is having a standard format for big data because the increased use of sensors and actuators continue to lead to the creation of more diverse (variety) data. The existence of this diversity and its continued development imply that more information will become available and the ease of access would render difficult to control the usage or the nature of the information that can be retrieved from the data (Vaidyaa et  al. 2018). For instance, the production process captured as event logs when modeled using tools such as process mining framework (ProM) or

8  Security Challenges in the Industry 4.0 Era

131

Disco can lead to inferences and conclusions about product designs although such logs may not necessarily require protection. Hence, as formats of big data evolve, and the information collated become diverse, it will become quite difficult to protect knowledge just as it will be harder to implement privacy measures. Nonetheless, the data storage systems within Industry 4.0 require protection against manipulation and access by unauthorized parties. As Magruk (2016) infers, the level of uncertainty associated with the major subsystems of Industry 4.0 calls for techniques of incorporating copyrights on both product and sensitive data exported for use in other CPPS and smart production systems. The event logs data is one area that may benefit from embedded copyright and intellectual property as part of the metadata may alleviate theft or unapproved use of such information. Antao et al. (2018) state that passive attacks that lead to access to private is a growing concern in Industry 4.0 and suggest that file verification techniques such as the use of HASH alongside the integration of cryptographic should be prioritized. However, Industry 4.0 lacks the expertise for implementing such techniques, and future endeavors of integrating them may equally prove challenging. It would be prudent for experts to work toward knowledge and copyright protection mechanisms that are inextricable from the data so that information from production systems and other sensitive information are segregated and access provided according to clearance levels (Magruk 2016). According to Ehrlich et al. (2018), security levels besides dynamic configuration protocols of security credentials is a first step in ensuring that processes within Industry 4.0, as well as people responsible for the processes, are cognizant of demands as per the authorization and access policies. The suggested security levels and measures also consider the physical substrate hardware used in the different aspects of the system (Ehrlich et al. 2018).

Legal Certainty and Data Protection The legal certainty or lack of it thereof is a subject that the very objectives of Industry 4.0 complicate because its concept is based on the distribution of services among different stakeholders. As such, the legal certainty and the subsequent enforcement of data protection laws or breaches in a legal context are dependent on whether the legal requirements were considered at the time of designing the system. Otherwise, failure to establish a legal framework from the beginning may result in uncertainties and innumerable risks that may fatally damage industrial development in Industry. 4.0. Of the many legal challenges, the unclarity of legal framework used in smart and service-oriented platforms is the main problem. For instance, the generation of event logs is inevitable, but most systems lack the legal framework for sharing them, and as such, it becomes difficult to limit their usage as well as the extent at which retrieved information can be shared. Given the unclarity of the legal frameworks used within industrial development platforms, it is also unclear how to legally protect and defend the generated data.

132

M. M. Alani and M. Alloghani

Data-Driven Cybersecurity The number of threats to Industry 4.0 is rising, and one of the ways getting ahead of it is implementing cyber intelligence techniques for tracking, analyzing, and identifying digital security threats (Falco et al. 2018). Some of the aspects of security that would benefit from data computational approaches include authentication and secure analytics based on machine learning algorithms. One of the emerging areas of authentication is the use of genetic algorithms that use biometric data to authorize access and maintain records of such access. Moving forward, it would be prudent for IoT-based manufacturing and Industry 4.0 technologies to rely on biometric cyber-based metrics for developing cyber profiles and verification protocols as a security measure (Osorno et al. 2011). Further, manufacturing industry must protect their cloud-based applications and all other systems that communicate over the Internet, and with developments in machine learning, it would be reasonable to implement techniques that rely on user behavioral patterns to identify malicious access activities (Liu et  al. 2018). With the advancements in data science, datadriven cybersecurity will alleviate most of these challenges although a gap between data science and Industry 4.0 exists, especially when it comes to implementation.

Security Recommendations for Industry 4.0 The following points are general security recommendations for Industry 4.0 that address security concerns in Industry 4.0 or in one of its underlying technologies: • The use of software-defined networks (SDN) in industry 4.0 can reduce security risks and make security policies more controllable (Khondoker et al. 2017). • Small-scale industry 4.0 solutions can benefit from centralized security solutions to reduce cost and improve security (Heikkilä et al. 2016). • Adopting standards that identify security frameworks and controls can be very useful. Standards such as ISO2700X, SANS critical security controls, German IT-Grundschutz, and NIST’s famous framework for improving critical infrastructure cybersecurity (Hartmann and Halecker 2015; Weyer et al. 2015). • Keep all systems patched and updated. This can protect your systems from many zero-day attacks. • Always have proper policies set. Security policies and other policies that govern the use of the systems (Kagermann 2015). • The use of advanced IoT gateways can reduce the security risks significantly (Fraile et al. 2018). When using smart cards, use smart cards that have proven security and implement proper cryptographic algorithms (Malina et al. 2018). • Industrial systems that were built a long time ago have definitely been target of many attacks throughout the time. This should encourage organizations to adopt newer technologies or at least very well tested old technologies (Shahzad et al. 2017).

8  Security Challenges in the Industry 4.0 Era

133

• A proposal was made by Hoeller and Toegl, in (2018), to use trusted platform modules to improve security in industry 4.0 smart environment. Although this has impact on dependability, this direction looks promising. • Safeguarding endpoints in smart industrial environments is equally important to protecting the central systems (Zhou et al. 2018). • Investment in cybersecurity is vital for the survival of the organization (Beissel 2016). • The organization must demonstrate its capacity to protect its technological resources. This gives more trust to your clients (Pereira et al. 2017). • Do not apply old solutions to new challenges in security. With the new threats emerging rapidly, it is important to keep up and not rely on older security controls (Khan and Turowski 2016).

References Akaev, A., Rudskoi, A., & Devezas, T. (2018). Digital economy and the models of income distribution in the society. In SHS Web of Conferences, Vol. 44. EDP Sciences, 2018, p. 00005. Alani, M. M. (2016). Elements of cloud computing security: A survey of key practicalities. Springer. Aloul, F., Al-Ali, A., Al-Dalky, R., Al-Mardini, M., & El-Hajj, W. (2012). Smart grid security: Threats, vulnerabilities and solutions. International Journal of Smart Grid and Clean Energy, 1(1), 1–6. Ani, U.  P. D., He, H., & Tiwari, A. (2017). Review of cybersecurity issues in industrial critical infrastructure: Manufacturing in perspective. Journal of Cyber Security Technology, 1(1), 32–74. Antao, L., Pinto, R., Reis, J., Gonçalves, G. (2018). Requirements for testing and validating the industrial internet of things. In 2018 IEEE International Conference on Software Testing, Verification and Validation Workshops (ICSTW). IEEE, 2018, pp. 110–115. Baena, F., Guarin, A., Mora, J., Sauza, J., & Retat, S. (2017). Learning factory: The path to industry 4.0. Procedia Manufacturing, 9, 73–80. Baheti, R., & Gill, H. (2011). Cyber-physical systems. The Impact of Control Technology, 12(1), 161–166. Bahri, A., Bouaziz, R., & Gargouri, F. (2012). “Towards an efficient datalog based evaluation of the FSAQL query language. In International Workshop on Resource Discovery. Springer, 2012, pp. 150–180. Balador, A., Ericsson, N., & Bakhshi, Z. (2017). Communication middleware technologies for industrial distributed control systems: A literature review. In Emerging Technologies and Factory Automation (ETFA), 2017 22nd IEEE International Conference on. IEEE, 2017, pp. 1–6. Barwick, H. (2012). “The ‘Four Vs’ of big data.” Implementing information infrastructure symposium, Computerworld Australia. http://www.computerworld.com.au/article/ 396198/iiis_ four_vs_big_data/. Beissel, S. (2016). Cybersecurity investments. Springer. Botta, A., De Donato, W., Persico, V., & Pescapé, A. (2016). Integration of cloud computing and internet of things: A survey. Future Generation Computer Systems, 56, 684–700. Boye, C. A., Kearney, P., & Josephs, M. (2018). Cyber-risks in the industrial internet of things (IIoT): To- wards a method for continuous assessment. In International Conference on Information Security. Springer, 2018, pp. 502–519. Cassandras, C.  G. (2016). Smart cities as cyber-physical social systems. Engineering, 2(2), 156–158.

134

M. M. Alani and M. Alloghani

Cazzola, P., Gorner, M., Schuitmaker, R., & Maroney, E. (2018). Global ev outlook 2018. France: International Energy Agency. Chaâri, R., Ellouze, F., Koubâa, A., Qureshi, B., Pereira, N., Youssef, H., & Tovar, E. (2016). Cyber-physical systems clouds: A survey. Computer Networks, 108, 260–278. Elsevier. Deloitte industry 4.0 report. https://www2.deloitte.com/content/dam/insights/us/articles/4323Forc esofchange/4323Forcesofchange_Ind4-0.pdf. Accessed 22 September, 2018. Ehrlich, M., Wisniewski, L., Trsek, H., & Jasperneite, J. (2018) Modelling and automatic mapping of cyber security requirements for industrial applications: Survey, problem exposition, and research focus. In 2018 14th IEEE International Workshop on Factory Communication Systems (WFCS). IEEE, 2018, pp. 1–9. Erl, T., Puttini, R., & Mahmood, Z. (2013). Cloud computing: Concepts, technology, & architecture. Pearson Education. Falco, G., Caldera, C., & Shrobe, H. (2018). IIoT cybersecurity risk modeling for SCADA systems. IEEE Internet of Things Journal. Fernández-Caramés, T. M., & Fraga-Lamas, P. (2018). A review on human-centered IoT-connected smart labels for the industry 4.0. IEEE Access, 6, 25939–25957. https://doi.org/10.1109/ ACCESS.2018.2833501. Fraile, F., Tagawa, T., Poler, R., & Ortiz, A. (2018). Trustworthy industrial IoT gateways for interoperability platforms and ecosystems. IEEE Internet of Things Journal, 5(6), 4506–4514. https://doi.org/10.1109/JIOT.2018.2832041. Geissbauer, R., Vedso, J., & Schrauf, S. (2016). Industry 4.0: Building the digital enterprise. Retrieved from PwC Website: https://www.pwc.com/gx/en/industries/industries-4.0/landingpage/industry-4.0-building-your-digital-enterprise-april-2016.pdf Graves, A., Mohamed, A.-R., & Hinton, G. (2013). Speech recognition with deep recurrent neural net- works. In Acoustics, speech and signal processing (ICASSP), 2013 IEEE international conference on. IEEE, 2013, pp. 6645–6649. Gubán, M., & Kovács, G. (2017). Industry 4.0 conception. Acta Technica Corviniensis-Bulletin of Engineering, 10(1), 111. Hartmann, M., & Halecker, B. (2015) Critical infrastructure security in smart ecosystems,” in Future City Challenges Conference, HTW Berlin & Metropolia, Berlin, Germany, Vol. 15, p. 16. He, H., Maple, C., Watson, T., Tiwari, A., Mehnen, J., Jin, Y., & Gabrys, B. (2016) “The security challenges in the IoT enabled cyber-physical systems and opportunities for evolutionary computing & other computational intelligence. In Evolutionary Computation (CEC), 2016 IEEE Congress on. IEEE, 2016, pp. 1015–1021. Heikkilä, M., Rättyä, A., Pieskä, S., & Jämsä, J. (2016) Security challenges in small-and medium-­ sized manufacturing enterprises. In Small-scale Intelligent Manufacturing Systems (SIMS), International Symposium on. IEEE, 2016, pp. 25–30. Helo, P., Suorsa, M., Hao, Y., & Anussornnitisarn, P. (2014). Toward a cloud-based manufacturing execution system for distributed manufacturing. Computers in Industry, 65(4), 646–656. Herterich, M. M., Uebernickel, F., & Brenner, W. (2015). The impact of cyber-physical systems on in- dustrial services in manufacturing. Procedia CIRP, 30, 323–328. Hertzberg, J., & Chatila, R. (2008). Ai reasoning methods for robotics. In B. Siciliano & O. Khatib (Eds.), Springer handbook of robotics (pp. 207–223). Berlin/Heidelberg: Springer. Hoeller, A., & Toegl, R.(2018) Trusted platform modules in cyber-physical systems: On the interference between security and dependability. In 2018 IEEE European Symposium on Security and Privacy Workshops (EuroS&PW). IEEE, 2018, pp. 136–144. Hsu, C.-L., & Lin, J. C.-C. (2016). An empirical examination of consumer adoption of internet of things services: Network externalities and concern for information privacy perspectives. Computers in Human Behavior, 62, 516–527. IBM Watson. https://www.ibm.com/watson/. Accessed September 15, 2018. IoT devices used in DDoS attacks. https://www.ibm.com/blogs/internet-of-things/ddos-iot-platform-security/. Accessed September 20, 2018. Johnson, R. C. (1993). What is cognitive computing? Dr. Dobb’s Journal, 18(2), 18–24.

8  Security Challenges in the Industry 4.0 Era

135

Kagermann, H. (2015). Change through digitization—value creation in the age of industry 4.0. In Management of permanent change. Springer, 2015, pp. 23–45. Kelly, J. E. (2015). Computing, cognition and the future of knowing. Whitepaper, IBM Research, p. 2. Khan, A., & Turowski, K. (2016) A survey of current challenges in manufacturing industry and preparation for industry 4.0. In Proceedings of the First International Scientific Conference “Intelligent Information Technologies for Industry”(IITI’16). Springer, 2016, pp. 15–26. Khondoker, R., Larbig, P., Scheuermann, D., Weber, F., & Bayarou, K. (2017) Addressing industry 4.0 security by software-defined networking. In Guide to Security in SDN and NFV. Springer, 2017, pp. 229–251. Kopetz, H. (2011). Internet of things. In Real-time systems. Real-time systems series. (pp. 307–323). Boston: Springer. Lee, J., Kao, H.-A., & Yang, S. (2014). Service innovation and smart analytics for industry 4.0 and big data environment. Procedia CIRP, 16, 3–8. Leitão, P., Colombo, A.  W., & Karnouskos, S. (2016). Industrial automation based on cyber-­ physical systems technologies: Prototype implementations and challenges. Computers in Industry, 81, 11–25. Liddy, E. D. (2001). Natural language processing. In Encyclopedia of library and information science (2nd ed.). New York: Marcel Decker, Inc. Liu, Y., & Xu, X. (2017). Industry 4.0 and cloud manufacturing: A comparative analysis. Journal of Manufacturing Science and Engineering, 139(3), 034701. Liu, S., Liu, X.  P., & El Saddik, A. (2013) Denial-of-service (dos) attacks on load frequency control in smart grids. In Innovative Smart Grid Technologies (ISGT), 2013 IEEE PES. IEEE, 2013, pp. 1–6. Liu, Q., Li, P., Zhao, W., Cai, W., Yu, S., & Leung, V. C. (2018). A survey on security threats and defensive techniques of machine learning: A data driven view. IEEE Access, 6., pp. 12 103–12 117. Magruk, A. (2016). Uncertainty in the sphere of the industry 4.0–potential areas to research. Business, Management and Education, 14(2), 275–291. Malina, L., Dzurenda, P., Hajny, J., & Martinasek, Z. (2018) Assessment of cryptography support and security on programmable smart cards. In 2018 41st International Conference on Telecommunications and Signal Processing (TSP). IEEE, 2018, pp. 1–5. de Man, J. C., & Strandhagen, J. O. (2017). An industry 4.0 research agenda for sustainable business models. Procedia CIRP, 63, 721–726. Masoni, R., Ferrise, F., Bordegoni, M., Gattullo, M., Uva, A. E., Fiorentino, M., Carrabba, E., & Di Donato, M. (2017). Supporting remote maintenance in industry 4.0 through augmented reality. Procedia Manufacturing, 11, 1296–1302. Meany, T. (2017). Functional safety and industrie 4.0. In Signals and Systems Conference (ISSC), 2017 28th Irish. IEEE, 2017, pp. 1–7. Mell, P., & Grance, T. (2011). The NIST definition of cloud computing (draft) (p. 800). NIST Spec. Publ. 7, National Institute of Standards and Technology, Gaithersburg. Murphy, K. (2012). Machine learning  — A probabilistic perspective. The MIT Press, National Institute of Standards and Technology, Gaithersburg. Osorno, M., Millar, T., & Rager, D. (2011). Coordinated cybersecurity incident handling: Roles, processes, and coordination networks for crosscutting incidents. Johns Hopkins University Laurel MD Applied Physics Lab, Tech. Rep. Pereira, T., Barreto, L., & Amaral, A. (2017). Network and information security challenges within industry 4.0 paradigm. Procedia Manufacturing, 13, 1253–1260. Poindexter, D.  F. (2018). The Chinese information war: Espionage, cyberwar, communications control and related threats to United States interests. McFarland. Rennung, F., Luminosu, C. T., & Draghici, A. (2016). Service provision in the framework of industry 4.0. Procedia-Social and Behavioral Sciences, 221, 372–377.

136

M. M. Alani and M. Alloghani

Sadeghi, A.-R., Wachsmann, C., & Waidner, M. (2015). Security and privacy challenges in industrial internet of things. In Design Automation Conference (DAC), 2015 52nd ACM/EDAC/ IEEE. IEEE, 2015, pp. 1–6. Shahzad, A., Kim, Y.-G., & Elgamoudi, A. (2017). Secure IoT platform for industrial control systems. In Platform Technology and Service (PlatCon), 2017 International Conference on. IEEE, 2017, pp. 1–6. Smith-Atakan, S. (2006). Human-computer interaction (pp. 54–57). London: Thomson Learning. Stock, T., & Seliger, G. (2016). Opportunities of sustainable manufacturing in industry 4.0. Procedia CIRP, 40, 536–541. The only coke machine on the internet. https://www.cs.cmu.edu/~coke/history_long.txt. Accessed September 15, 2018. Tupa, J., Simota, J., & Steiner, F. (2017). Aspects of risk management implementation for industry 4.0. Procedia Manufacturing, 11, 1223–1230. Ustundag, A., & Cevikcan, E. (2018). Industry 4.0: Managing the digital transformation. Cham: Springer. Vaidyaa, S., Ambadb, P., & Bhoslec, S. (2018). Industry 4.0–a glimpse. Design Engineering, 2351, 9789. Waidner, M., & Kasper, M. (2016) Security in industrie 4.0-challenges and solutions for the fourth industrial revolution. In Design, Automation & Test in Europe Conference & Exhibition (DATE), 2016. IEEE, 2016, pp. 1303–1308. Wang, W., & Lu, Z. (2013). Cyber security in the smart grid: Survey and challenges. Computer Networks, 57(5), 1344–1371. Wang, S., Wan, J., Zhang, D., Li, D., & Zhang, C. (2016a). Towards smart factory for industry 4.0: A self- organized multi-agent system with big data based feedback and coordination. Computer Networks, 101, 158–168. Wang, S., Wan, J., Li, D., & Zhang, C. (2016b). Implementing smart factory of industrie 4.0: An outlook. International Journal of Distributed Sensor Networks, 12(1), 3159805. Waslo, R., Lewis, T., Hajj, R., & Carton, R. (2017). Industry 4.0 and cybersecurity: Managing risk in an age of connected production. Deloitte University Press, National Institute of Standards and Technology, Gaithersburg. Weyer, S., Schmitt, M., Ohmer, M., & Gorecky, D. (2015). Towards industry 4.0-­standardization as the crucial challenge for highly modular, multi-vendor production systems. IFAC-Papersonline, 48(3), 579–584. Wolf, W. H. (2009). Cyber-physical systems. IEEE Computer, 42(3), 88–89. Yan, J., Meng, Y., Lu, L., & Li, L. (2017). Industrial big data in an industry 4.0 environment: Challenges, schemes, and applications for predictive maintenance. IEEE Access, 5., pp.  23 484–23 491. Zawadzki, P., & Żywicki, K. (2016). Smart product design and production control for effective mass customization in the industry 4.0 concept. Management and Production Engineering Review, 7(3), 105–112. Zezulka, F., Marcon, P., Vesely, I., & Sajdl, O. (2016). Industry 4.0–an introduction in the phenomenon. IFAC-PapersOnLine, 49(25), 8–12. Zhou, L., Yeh, K.-H., Hancke, G., Liu, Z., & Su, C. (2018). Security and privacy for the industrial internet of things: An overview of approaches to safeguarding endpoints. IEEE Signal Processing Magazine, 35(5), 76–87. Zuehlke, D. (2010). Smartfactory—Towards a factory-of-things. Annual Reviews in Control, 34(1), 129–138.

Chapter 9

Meeting the Future Challenges in Cyber Security Ben Azvine and Andy Jones

It will then look at both the sociological and technological issues that the developing cyber environment will pose, covering issues such as artificial intelligence (AI), smart environments, the Internet of Things (IoT), quantum computing, and big data and will examine the complexity problems that will increase with their greater adoption. It will finish by examining a range of both technical and nontechnical potential solutions and mitigation measures that can be adopted to address the security problems of the future cyber environment.

The Emerging Threat Environment In the past few years, the concept of Cyber Warfare that has been theorized for nearly two decades has become a reality. Defining Cyber Warfare is difficult for a number of reasons. The first is that it is perceived to be different things depending on the region and nation. While the Western nations have largely agreed on the meaning of the term, Russia and China both have significantly different views on what the term (or its equivalent) means. To make it more difficult, the terms information warfare, Cyber Warfare, cyber war, cyberwarfare and cyber operations are very often used interchangeably. However, in reality, the term information warfare is commonly used to describe a wider environment than “cyber.” Dan Kuehl of the US National Defense University defined information warfare as the “conflict or struggle between two or more groups in the information environment.”

B. Azvine (*) British Telcome Labs, Ipswich, UK A. Jones Cyber Security Centre at the University of Hertfordshire, Hertfordshire, UK Edith Cowan University, Perth, Australia © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_9

137

138

B. Azvine and A. Jones

In Russia, the Chief of the General Staff of the Russian Federation’s Armed Forces, General Valery Gerasimov, detailed the “Gerasimov Doctrine” in an article that was published in 2013 in the weekly Russian trade paper, Military-Industrial Kurier. This article showed Russia’s modern strategy, which is a vision of total warfare that puts politics and war within the same spectrum of activities. The approach that is put forward is that operations are waged on all fronts and with a wide range of actors and tools. These can include hackers, the media, the use of leaks and fake news (disinformation), as well as conventional and asymmetric military means. The Gerasimov Doctrine develops a framework for these new tools and declares that nonmilitary tactics are not auxiliary to the use of force but the preferred way to win and that they are, in fact, the actual war. Gerasimov stated that the objective is to achieve an environment of permanent unrest and conflict within an enemy state. In a 2014 article, the Russian use of the term information warfare includes: • Psychological operations to influence the motivation of enemy soldiers • Disinformation – providing false information to the enemy about their capabilities and plans • Electronic warfare, the “blinding” the enemy of electronic intelligence systems • Physical destruction of elements of the information systems of the enemy • Information attack – the destruction or corruption of information without visible damage to the carrier systems • The protection of their own information The article goes on to show the Russian view that the aims of information warfare are: • To create an atmosphere of immorality and lack of spirituality, in order to create an atmosphere which is likely to cause conflict within a country and overthrow the authorities • The manipulation of public opinion and the political orientation of social groups to create a climate of political tension and chaos • Destabilization of the political relations to provoke conflict and incite an atmosphere of distrust and suspicion • To aggravate political struggle and to provoke repression against the opposition • To cause the outbreak of civil war in the society • To reduce the level of information support for governing bodies in order to impede the making of important decisions • To spread misinformation within the population about the state authorities to undermine their authority and discredit the government • To provoke social, political, ethnic, and religious conflicts • To cause mass protests, strikes, and riots • To undermine the international authority of the State • To cause damage to the vital interests of the state in the political, economic, defense, and other areas When compared to the Western definition of information warfare (IW), it can be seen that this is a much more encompassing use of the term.

9  Meeting the Future Challenges in Cyber Security

139

China has never formally published an information warfare strategy document or a computer network operations strategy, but there have been a number of high-level, long-term directives, known as the Military Strategic Guidelines published, that give the direction for defense policy and set out a long-term course of action for the modernization of the military. The Chinese have adopted the concept of “information confrontation,” which is aimed at integrating all elements of information warfare, both electronic and nonelectronic, offensive and defensive, under a single command authority. IW is not specifically addressed in any one issue area. The last time that the Central Military Commission (CMC) updated the Military Strategic Guidelines was in 1993. This update stated that the People’s Liberation Army (PLA) should prepare to “fight local wars under high-tech conditions.” This was then updated again in 2002 when Jiang Zemin, the then President of the People’s Republic of China, stated that the PLA must be able to fight and win “local wars under informationized conditions.” This was a minor change in terminology from “local wars under high-­ technology conditions” but revealed a change in the PLA’s strategic focus to informationization. This is the ability to link all services and units through a shared information infrastructure to enable joint operations. The Chinese strategy incorporates concepts such as that successful warfighting is predicated on the ability to exert control over an adversary’s information and information systems and that this may be achieved preemptively. China has adopted an approach where offensive and defensive IW missions are closely coordinated to ensure that these activities are mutually supporting and are closely integrated with their campaign objectives. There are an increasing number of nations that either possess or are developing Cyber Warfare capabilities, some of which are friendly western nations, others are not. Because of the nature of the capability and the problems with attribution for cyber activity, the scope of the capabilities of individual nations is very difficult to determine.

Cyber Warfare: The Big Picture The term Cyber Warfare (also seen as cyber war, cyberwarfare, and cyber operations) has been increasingly used in a wide range of contexts over the past few years. As is noted in the Tallinn manual “There are no common definitions for Cyber terms - they are understood to mean different things by different nations/organisations, despite prevalence in mainstream media and in national and international organisational statements.” The US Department of Defense Law Of War Manual that was released in June 2015 gives a description of cyber operations as follows: “Cyberspace operations may be understood to be those operations that involve ‘[t]he employment of cyberspace capabilities where the primary purpose is to achieve objectives in or through cyberspace.’ Cyber operations: (1) use cyber capabilities, such as computers,

140

B. Azvine and A. Jones

s­ oftware tools, or networks; and (2) have a primary purpose of achieving objectives or effects in or through cyberspace.” NATO, through the Cooperative Cyber Defence Centre of Excellence (CCDCOE), commissioned an Independent Group of Experts to produce “The Tallinn Manual on the International Law Applicable to Cyber Warfare” which was published in 2013 and subsequently updated in 2017. This document provides an in-depth look at the topic and the way in which International law relates to it, and, while it does not provide any definitions of its own, it offers a range of definitions from other sources. One of the most relevant definitions for Cyber Warfare is “cyber attacks that are authorized by state actors against cyber infrastructure in conjunction with government campaign.”

Actors Understanding the motivation of attackers plays an important part in choosing the correct defense strategy. In this section we’ll look at this issue in detail by presenting key classes of bad actors operating in Cyber space.

Nation State Actors Nation State Actors are known to have been actively carrying out cyber-attacks for a number of years, and the aims of these attacks have varied depending on the state(s) that has sponsored the attacks. Examples of the types of cyber-attacks that have been attributed to Nation State Actors include: • Stuxnet. This is a 500-kilobyte malware that targets Siemens industrial control systems (ICS) that are commonly used in infrastructure supporting facilities (power plants, water treatment facilities, gas lines, etc.). It was first identified in 2009 but was actually found to have attacked the infrastructure of Iran’s nuclear program as early as 2007. More than 15 Iranian facilities were reported to have been attacked by the Stuxnet worm. While it was initially found in Iran, it was later found in Indonesia and India, accounting for over 85% of all infections. The worm has now affected a large number of computers in many countries. Stuxnet is designed to affect the Programmable Logic Controllers (PLCs) used in those facilities. It was widely speculated that the USA and Israel were responsible for this attack. • Shamoon. In 2012, the Shamoon virus targeted Saudi Arabia’s Saudi Aramco oil company and destroyed the data from more than 35,000 PCs. A group called Cutting Sword of Justice claimed responsibility for the attack, which affected an estimated three quarters of the workstations used by the company. The motive was claimed to be in retaliation against the Al-Saud regime for the “crimes and

9  Meeting the Future Challenges in Cyber Security

141

atrocities taking place in various countries around the world, especially in the neighboring countries such as Syria, Bahrain, Yemen, Lebanon [and] Egypt”; however, there was suspicion that Iran was behind the attack.

Advanced Persistent Threats (APTs) An Advanced Persistent Threat (APT) is a stealthy cyber-attack in which an unauthorized person or group gains access to a network and maintains that undetected access over a long period of time. While the majority of APTs have been attributed to Nation States, other groups have also carried out this type of attack. The intention of an APT attack appears to most commonly be to steal data rather than to cause damage to the network. There are a number of APTs that have been identified with the majority of them being attributed to activity initiated or supported by either Russia or China and North Korea; however, more recently, there have been examples of non-nation state groups undertaking large-scale targeted attacks to achieve specific goals. Examples of APTs that have been identified include: • Titan Rain. This was the code name given by the US government to a series of attacks launched in 2003 on US defense contractors, including Lockheed Martin, Sandia National Laboratories, the Redstone Arsenal, and NASA.  The attacks were claimed to be of Chinese origin, although the Chinese government denied any involvement. The notable features of these attacks were the level of deception used and the use of multiple attack vectors (channels of attack), which combined well-researched social engineering attacks on specific, targeted individuals with stealthy Trojan horse attacks using malware with the intention of bypassing the standard security countermeasures. This APT went on to steal data from a wide spectrum of organizations from all major sectors of industry including aerospace, defense, energy, the financial services, manufacturing, pharmaceuticals, and technology. The motivation for these attacks appears to have been cyber espionage. • Pawn storm. Operation pawn storm is an active economic and political cyber espionage operation that has targeted a wide range of high-profile organizations, from government institutions to media personalities. Its activities were first seen as far back as 2004, and the motive of this APT appears to be cyber espionage. The targets that have been identified include NATO, Gov’t, Military, Russian Dissidents, Ukrainian organizations, and related interests. • Waterbug. This APT has been operating since 2005 and uses an attack network known as Venom that consists of around 84 compromised websites located in a number of different countries. The highest numbers of compromised websites are located in France (19%), Germany (17%), Romania (17%), and Spain (13%). The targets of this APT include government institutions, embassies, and educational and research facilities. The motive of this group appears to be cyber

142

B. Azvine and A. Jones

e­ spionage. It is thought that this group was also responsible for the 2008 compromise of US Central Command, which reportedly resulted in a clean-up operation that took more than 1 year. • Regin, which was first identified in 2008, is a highly complex threat that targets telecom operators, government institutions, multinational political bodies, financial and research institutions, and individuals involved in advanced mathematics and cryptography. The attackers appear to be primarily interested in gathering intelligence and facilitating other types of attacks. While much of the intelligence gathered includes spying on emails and documents, the attack group also relentlessly targets telecommunication companies, which is normal, and at least one GSM provider, which is less so. The development and operation of this threat would have required a significant investment of time and resources. • Dragonfly is a group of hackers that have probably been operating out of Eastern Europe since 2011. The activity has the characteristics of a state-sponsored operation, and an analysis of the time stamps for the compilation of the malware used by the attackers indicates that the group normally worked from Monday to Friday and the activity was mainly concentrated in a 9-hour period that corresponded to a 9 am to 6 pm working day in the UTC +4 time zone. The targets of this group include Defense, aviation in the USA and Canada, and more recently European energy firms, and the motive appears to be cyber espionage and sabotage. • Black Vine. This APT group was first identified in 2012, and it is believed that some actors of Black Vine may also be associated with an IT security organization called TopSec that is based in Beijing. Black Vine typically conducts watering-­hole attacks against websites that are relevant to its targets’ interests and uses zero-day exploits to compromise computers. This is the group that is thought to be responsible for the 80 million record breach at Anthem. The malware then opens a back door on the compromised machines and allows the attackers to steal valuable information. The main targets are aerospace, energy, and healthcare, and the motive appears to be cyber espionage.

Non-state Actors In addition to the nation state actors, there are a wide range of other parties that may engage in cyber operations that use the same tools and techniques as those identified for Cyber Warfare. These include non-state groups such as the Islamic State of Iraq and Syria (ISIS), also known as the “Islamic State of Iraq and the Levant” (ISIL) and Daesh (an acronym of the group’s full Arabic name, al-Dawla al-Islamiya fi al-­ Iraq wa al-Sham, which is translated as “Islamic State in Iraq and Syria (or the Levant).” This group displayed an understanding of the cyber environment and used it to a significant degree for propaganda, recruitment, and fund raising.

9  Meeting the Future Challenges in Cyber Security

143

Hacktivists Hacktivists, which in itself covers a wide range of individuals and groups, have demonstrated a good depth of knowledge on how to utilize the cyber environment to apply pressure to organizations in support of their interests. Perhaps the best known hacktivist group is Anonymous, which came into being in 2003. This group is an anarchic, decentralized collective of people with an agreed common goal. Starting in 2008 the group undertook a number of actions against the Church of Scientology and subsequently became increasingly associated with collaborative hacktivism on a number of issues internationally. Other targets of Anonymous included government agencies of the USA, Israel, Tunisia, Uganda, the Islamic State of Iraq and the Levant (ISIL), child pornography sites, copyright protection agencies, and corporations such as PayPal, MasterCard, Visa, and Sony. A number of the members of Anonymous have been arrested for involvement in Anonymous cyber-attacks, in countries including the USA, the UK, Australia, the Netherlands, Spain, India, and Turkey. Interestingly, in 2012, Time magazine called Anonymous one of the “100 most influential people” in the world.

Criminals Criminals and organized crime groups have increasingly made use of the cyber environment to achieve their ends. To quote the old adage, criminals will follow the money, and the money has gone online. The Internet has provided the environment where traditional crimes can now be carried out online. The benefits of this include that the crimes can be committed without physically visiting the scene and so can be carried out from anywhere in the world, there is a significantly lower likelihood of being caught, and, even if they are, because the law cannot be brought up-to-date fast enough to address the rapid changes in technology, the likelihood of a conviction is much lower than for conventional crimes. Examples of the types of conventional crimes that have successfully moved to the cyber environment include fraud, blackmail, and identity theft. New crimes that have arisen from the use of technology include hacking, phishing, denial of service, cyber espionage, and ransomware. The global cost of cybercrime in 2017, according to a report by McAfee and the Center for Strategic and International Studies (CSIS), was reported as being between $445 and $608 billion, which is higher than the gross domestic product of all except the top 23 countries in the world. An increasingly commonly seen attack type from criminal groups is the ransomware attack in which malicious software (or malware) is loaded onto a computer, through one of a number of vectors. One of the most common delivery methods is through phishing spam (attachments that are sent to the victim in an email, masquerading as a file they should trust). Once downloaded and opened, it takes over

144

B. Azvine and A. Jones

the victim’s computer. Some of the more aggressive forms of ransomware, such as NotPetya, exploit security vulnerabilities to infect computers without the need to trick users. In the last 2 years, there has been a dramatic increase in the number of ransomware attacks, with hospitals being particularly badly affected.

Individuals/Lone Hackers It may seem to be perverse to discuss lone individual hackers in the same context as nation states, but the cyber environment has provided a situation where the capabilities of an individual can have a disproportionate effect. There have been numerous examples where an individual has caused significant damage to the cyber environment. The first publicized example was the release of the “Morris worm” onto the early Internet in 1988 (1 year before the World Wide Web was invented and 3 years before it became available to the public). On November 2, 1988, the Morris worm was released by its author, Robert Tappan Morris, Jr., the son of Robert Morris, Sr., a well-known cryptographer and computer professional, and within 24 hours had caused significant damage around the world. The Morris worm infected about 10% of the computers connected to the Internet, the only malware in history that reached that level. However, it is worth noting that the number of computers connected to the Internet at that time were much lower, since around 60,000 computers were connected to the ARPANET and 6000 became infected by the malware. The majority of the compromised computers belonged to NASA, Berkley and Stanford universities, MIT, and the Pentagon, and it took around 72 hours to bring the infection under control. Another example is the ILOVEYOU worm, also known as the Love Bug or the Love Letter, that was released on May 4, 2000, and originated in Manila in the Philippines. The authors of this malicious software were Reonel Ramones and Onel de Guzman. This was a computer worm that attacked tens of millions of Windows personal computers. The worm overwrote random types of files; however, after overwriting MP3 files, the virus would hide the file and sent a copy of itself to all addresses in the Windows Address Book used by Microsoft Outlook. Interestingly, as there were no laws in the Philippines against the writing of malicious software at the time, both Ramones and de Guzman were eventually released with all charges dropped by the state prosecutors. In 2002, the ILOVEYOU virus gained a world record for being the most virulent computer malware at the time.

Current Issues In this section we’ll focus on a number of emerging issues that play an important role in the design and operation of future cyber defense systems.

9  Meeting the Future Challenges in Cyber Security

145

The Password “Ticking Bomb” User authentication is one of the fundamental building blocks of any security system. This is true in both physical and digital security. Most secure locations require a form of identification such as a pass card and/or a code. More sensitive locations use multiple mechanisms to identify legitimate users of the facilities. Current methods for identity and access management rely on what you know, i.e., a password; what you are, i.e., biometrics; or what you have, i.e., a token or a combination of these. However, most of today’s computer systems use some form of password. Given the increasing number of breaches, companies have policies that force users to remember longer and longer passwords that are not easy to remember for most people let alone those with memory limitations. This has led to people writing down their password or trying to bypass the guidelines on how to choose secure passwords. According to the Australian government “Stay Smart Online” website, “123456” has been the most commonly used password every year since 2013. Although resetting your password today is quick and cost-effective, it causes anxiety and delay, e.g., in the middle of a payment process. Many existing companies force their customers to choose more secure passwords that are harder and harder to remember. As a result, many people use the facility in their web browser to “remember” the password for them. The fundamental problem that future identity and access management systems need to solve is to increase the overall level of security without sacrificing cost or usability. In fact, this is the goal in all areas of security. So, one possible scenario is that the future of passwords is to have no passwords at all. One way to think about this is to look at how humans identify each other. We use lots of information very quickly to decide who might be talking to us. Face to face, we use images first, and then the way someone acts, talks, moves, what he/she says, etc. On the phone we have a subset of this information, so we may need more information and a slightly longer time to be sure who we’re talking to. It is more difficult to be sure of someone’s identity from an email or when using social media, hence the social engineering attacks we see today. What is being developed are systems that use artificial intelligence (AI) to identify people. AI does this by combining information that can be observed such as how you type, how you use your phone, and biometrics including face and voice recognition. Many of these techniques have been around and used individually for identity management; what is new is to combine them using AI in order to learn and improve the accuracy and speed of the system rather than hard-code the logic.

Technological Challenges It is hard to ignore the impact technology has had on people’s lives in the last 30 years. Children today cannot imagine a world without Wi-Fi, social media, and mobile phones. Technology has also transformed work, health, entertainment, and

146

B. Azvine and A. Jones

even money. What isn’t in doubt is that the pace with which we’ll embrace new technology in our lives will increase, and this will have profound consequences on the state of Cyber security. In this section, we’ll explore the impact of a few technologies that will have a significant effect on the future of Cyber security.

The Internet of Things (IoT) Every few years, a technology comes along that has the potential of transforming society. The Internet transformed how we shop, social media transformed how we interact with other people, and the Internet of Things (IoT) is going to transform every aspect of our lives, from health to education and transportation. It is hard to overestimate the impact IoT will have on our lives in the next 20 years. But for IoT to deliver this promise, we need to solve a few hard problems first. The most pressing is that of Security. Some estimates point to 50 Billion IoT devices being connected by 2020. These range from sensors in your waste  bins to fully connected smart hospitals. What most of these devices will have is memory and processing power so in essence there will be 50 Billion potential targets for hackers to attack. What is different about IoT systems is that they can interact with the physical world. A smart toaster can switch itself on and off, and a smart car can operate the brakes and 1 day, in the near future, drive itself. This means that an insecure IoT device can, potentially, cause damage that could lead to loss of life or disruption of vital services. It is true that security attacks today could have similar consequences, but with 50 billion devices, the attack surface will be much larger, and the risk will be higher. Both the government and industry are aware of the risks and opportunities. For example, The IoT Security foundation (IoTSF) was set up in September 2015 following a gathering of technology professionals and security experts at the iconic Bletchley Park (England) in May 2015 to explore IoT security issues. “IoTSF mission is to help secure the Internet of Things, in order to aid its adoption and maximise its benefits. To do this it will promote knowledge and clear best practice in appropriate security to those who specify, make and use IoT products and systems.”

Quantum Computing In 1982, the Nobel prize-winning physicist Richard Feynman proposed the idea of quantum computing, i.e., using quantum properties of subatomic particles such as superposition and entanglement to perform computation. Later, Peter Shor (Bell Laboratories) proposed an algorithm for a quantum computer to break current cryptography techniques in a matter of seconds. With the motivation provided by this algorithm, the topic of quantum computing has g­ athered

9  Meeting the Future Challenges in Cyber Security

147

momentum, and researchers around the world are racing to be the first to create a practical quantum computer. Conventional computers use bits to do their computation; bits can be either 0 or 1 and computer operator on 1 bit at a time. Quantum computers use qubits that can have values between 0 and 1 simultaneously. According to David Deutsch, widely known as one of the pioneers of quantum computing, quantum computers are millions of times faster than conventional computers. This is great news for medical research or building AI systems that learn instantly, but the downside is that such machines can break most of today’s encryption techniques within hours or days. Although there are no general-purpose quantum computers yet, there have been promising developments on building prototypes with sufficient number of qubits to worry governments and companies into action to develop quantum safe encryption. One of the approaches that have shown good results is known as quantum key distribution that uses quantum properties to establish a quantum safe encryption key on optical networks.

Artificial Intelligence (AI) It may surprise most people that artificial intelligence (AI) is older than the Internet, social media, and smartphones. In 1936, Alan Turing proposed the universal Turing machine that laid the foundation for today’s computers. Later in 1950 he published a paper called Computing Machinery and Intelligence. In the same year, Isaac Asimov published his three laws of robotics. The first use of the term artificial intelligence is attributed to John McCarthy at the famous Dartmouth Conference, the first conference devoted to AI. Most people’s view of AI today is formed by science fiction movies such as AI; I, Robot; Terminator; and Star Wars. In terms of the impact of AI in the real world, we’re beginning to see promising developments mainly due to the availability of large volumes of data and cheap processing power. AI systems require lots of training and powerful processors. Our smartphones today have the minimum processing and memory requirements for simple AI systems such as SIRI and Alexa. Moore’s law means that we’ll have more AI in our lives in the next 5 years not only at home but also at work. AI can already be used to diagnose disease more consistently that human experts and can solve optimization problems much better than humans. This will have significant benefits in achieving better Cyber security. Like any new technology, AI will have both positive and negative consequences on security. A lot of work is focussed on how AI can be used to detect abnormal patterns in data preceding a cyber-attack. This is called Cognitive Security, and most established security vendors either already offer AI in their products today or are working on it. What is less talked about but equally important is understanding how AI can be weaponized. For example, new forms of adaptive malware could learn and evade anti-malware systems. Another example is using machine learning (ML) to analyze vulnerabilities of a target network; the implication of this type of attack is that the “low-hanging fruits” will be discovered much

148

B. Azvine and A. Jones

more quickly and with lower cost by criminals necessitating much stronger security prevention and detection capabilities. Also AI systems have inherent vulnerabilities that can be exploited by bad actors, e.g., adversarial examples and poisoning machine learning attacks: introduction of small noise in training data to degrade or change learning accuracy. Another area of focus is around the impact of AI on privacy; the question is if AI systems are compromised and how personal or sensitive information that are vital to train them and keep them up-to-date be protected.

Big Data Privacy and Security Big data technologies have significantly reduced the cost of storage and processing of very large volumes of data. What was only possible for governments and large organizations can be achieved now with a cluster of low cost servers. Naturally big data plays an important role in the future of cyber security. Many experts believe that cyber security is a big data analytics problem. Most practical applications of security analytics today use big data technologies. Attackers are aware of this and are targeting big data repositories. This makes privacy and security of data assets more important than even. New forms of encryption are being developed that can operate at much faster speeds and on big data scale volumes to enable full exploitation of this powerful technology to analyze and detect patterns and anomalies on network data. Standard analytics have always relied on sampling of data, and although sampling is very effective when looking for repeatable patterns in data, in security one is looking for rare “Black Swan”-type anomalies, and sampling doesn’t always capture such events. Big data provides the capability to search the entire data space thereby eliminating the possibility of missing the weak signals. But big data also creates a big target for criminals. Many of the breaches we see today involving credit card information of tens of millions of people pale into insignificance compared to a breach involving a big data leak of medical records. So, going forward, we need to develop powerful encryption technologies to enable analysis of large volumes of encrypted data without the need to decrypt. Doing this requires significant advances in homomorphic encryption. Today there are promising approached such as differential privacy and searchable encryption that are beginning to offer practical solutions to this problem.

Complexity One way to look at the future of security is to treat cyber systems as Complex Adaptive Systems (CAS). In short, this means that it is impossible to predict the impact of a cyber-attack based on the behavior of the individual components in a cyber system. This includes technology, people, and processes. A lot of research has

9  Meeting the Future Challenges in Cyber Security

149

been done on how biological systems that are CAS respond to biological attacks. Some useful principles can be learned from the study of such system to build future cyber defense. For example, formal validation and testing and perimeter defense are two widely used approached in cyber defense. These approaches however rely on the assumption that an attack can be replicated and prevented by a predefined set of patterns, rather than treating an attack as a CAS with unpredictable consequences. Biological systems use diversity as a defense mechanism, e.g., different parts on an organism react differently to biological attacks so if a virus gets through the defenses, it has less chance of damaging all parts equally. Current critical IT systems today are built to be the same which makes them vulnerable to attacks. More research is needed to understand how CAS analysis techniques such as agent-based modelling can be used to understand vulnerabilities and defense strategies for future cyber defense.

The Solutions So far we have focussed mainly on the challenges we see in building and operating future cyber security systems and processes. In this section we’ll turn to what individuals and businesses should do to be better prepared for the future challenges.

Changing the Mind-Set Given the evolving threat environment and the technological challenges, we need to do more than strengthening our current defenses. The most current approach to cyber security follows the perimeter defense model, what we can call the Coconut approach. In this approach, a hard shell is built around the entire organization with the assumption that the thicker the shell, the better protection it provides. All the evidence points to the fact that this approach does not work. The insider threat is a by-product of how this approach would fail. A more effective approach is the avocado approach. In this approach critical assets that form a small proportion of the total assets are protected with multiple layers of protection thereby making them less attractive to attacks; this is the center of the avocado. Other assets are protected but not to the same level. For this approach to work, one needs to add detection capability to monitor and respond to early signs of attacks before they get to critical assets. This approach is more affordable and more realistic in the future. If we try to protect everything to the same level, we’d end up with security that is either too expensive or too restrictive and unusable. The avocado model fits what is known as the risk-based approach to security. It does require a comprehensive risk assessment leading to an ordered list of critical assets by priority. It is a model that has been adopted by companies and organizations who are ahead of the curve in cyber security, and others are trying to catch up.

150

B. Azvine and A. Jones

Thinking End-to-End One way of thinking about how we should implement security going forward is to think in terms of stages of a cyber-attack. Clearly the goal for everyone is to prevent attacks in the first place. This is achieved by designing security into products and systems not adding them afterward. This means introducing security at all the stages of the software and hardware design life cycle followed by extensive testing of the systems in as realistic an operating condition as possible. It is also important to design the systems to allow for patches and upgrades, preferably initiated remotely, to reduce operating cost and effort further down the line. Building a solid prevention foundation is critical but not enough. No matter how good the designs are, bad actors will find holes in them or at least that should be the assumption when we’re thinking about future security. We must be prepared for the defenses to be breached. We need another layer of defense for monitoring, detection, and prediction. As described in previous sections, technologies such as AI and big data analytics can be deployed today to look for anomalies in real time at relatively low cost. Future cyber defense strategies must include an element of monitoring and detection to complement secure by design strategies. Let’s assume prevention stops 99% of the attacks, for the remaining 1% we must rely on detection. This isn’t easy and there is a race among security vendors to come up with products that can find the needle in the hay stack. Academia has a significant role to play to undertake fundamental research on better algorithms that can be easily used for detection of extremely rare patterns in data in real time. We mustn’t stop there; detection is getting better, but it isn’t perfect. In the same way that prevention doesn’t stop all attacks, some very sophisticated attacks will not be detected in time, so we must add another layer of protection. This is response. Today on average it takes 31 days for companies to deal with a cyber-attack after it has been discovered. If we can reduce this to minutes or seconds, the ability of attackers to steal information and cause disruption or damage will be greatly reduced. To achieve this, we need to introduce much higher levels of autonomy in our Network and IT infrastructure. Recent moves toward virtualization at both Network and IT levels, e.g., Cloud, Network Function Virtualisation, and software-­ defined networking will facilitate near real-time response to future attacks.

Intelligence Augmentation In the AI section, we discussed the impact AI can have on the future of cyber security. While advanced AI systems are being researched to automatically detect and respond to cyber-attacks, there is a need to stop the growing number of breaches today. The current state of AI/ML systems does not provide a fully automated detection and response capability. Although many ML systems are statistically impressive, the small number of false positives they produce makes them unsuitable for

9  Meeting the Future Challenges in Cyber Security

151

security applications. Therefore, an engineering approach to combine the power of AI/ML systems with humans is to build systems that help humans detect and respond to attacks quickly and cost effectively. This is what is known as intelligence augmentation which is a subset of a broader field known as Cognitive Security. The question for researchers has been how to build systems that allow human experts to work with AI/ML. The field of visual analytics has been used extensively to create a natural, interactive interface for human experts to hunt for patterns. In such environments, each mouse click can be a feedback loop to the underlying AI/ML system, and the sequence of interactions can be used to determine the direction of training for the system. This approach has produced impressive results in terms of reducing the time to detect early signs of future attacks.

Artificial Immune Systems In the early 1990s, the field of artificial immune systems (AIS) was initiated that proposed the creation of networked code or agents that could operate in a continuous proactive fashion to defend against the insertion of malicious code into host computers. The research was fruitful in simulation and concepts but failed to scale into the real world. Standard security models based on firewalls, intrusion detection, and antiviral signatures became the norm. The utility of an immune system model is likely to be one way forward to defend against malicious code. The other issue is that state-of-the-art AI is also combining evolutionary algorithms with deep neural networks [e.g., PathNet], to create evolvable AI systems. Future cyber defense needs to adopt and further develop AIS principles. Although full autonomy may not be feasible or acceptable in the near future, it is possible to implement AIS at the edges of the networks to deal with high-­ volume low-impact attacks, complemented with a more centralized monitoring and response mechanism for very sophisticated and/or high-impact cyber-attacks.

Conclusions It is clear that the volume and scale of attacks will continue to increase as the attack surface continues to grow and become more complex. In the future we cannot continue to do what we have been doing and need to explore new approaches, tools, and techniques to, at the very least, ensure that the security levels we currently achieve are maintained and to move toward more effective security environments. The implementations of security measures will always be behind the techniques used by attackers as new vulnerabilities in the technologies and the users are discovered and exploited. In the future we need to improve the prevention and detection of attacks and reduce the time taken to respond to them. The use of machine learning/AI, deep

152

B. Azvine and A. Jones

learning, big data analytics all have the potential to provide capability that will assist in the protection of the electronic environment and the detection of attacks. If secure environments are to be developed and maintained, it may be necessary to rethink the way in which we currently operate and perhaps either develop secure operating systems and applications or, alternatively, remove our critical infrastructure from the Internet. In terms of staying ahead of the threat landscape, intensifying the focus on R&D in cyber security is essential. We must rethink the way we approach security and not be afraid to try radical new ways such as biologically inspired approaches. What is clear is that those intent on stealing information and disrupting our lives are adapting to our defenses, using new technology, and infiltrating our organizations. One of the most effective defenses against future adversaries is to keep innovating, invest in R&D in the technologies covered earlier in this book, and do it at a pace faster than the dark side.

Chapter 10

Ready for Industry 4.0? The Case of Central and Eastern Europe Wim Naudé, Aleksander Surdej, and Martin Cameron

Introduction Technological innovations such as digital platforms, artificial intelligence (based on big data) and automation, additive manufacturing (3D printing), and smart materials are in the process of disrupting the world economy. This disruption is expected to be keenly felt in the manufacturing sector. Here a “new” industrial revolution (Marsh 2012), a “second machine age” (Brynjolfsson and McAfee 2015), or a “4th industrial revolution” (Schwab 2016) has been diagnosed. The digitization and automation of manufacturing characterize what is known as industry 4.0, from Germany’s industrie 4.0.1 Industry 4.0 technology enables firms to improve operational efficiency, productivity, time to market, and customer satisfaction and to reduce carbon emissions, waste, costs, and downtime (see, e.g., McKinsey Digital 2015). Industry 4.0 (I4.0) will have significantly implications for the global distribution of manufacturing activities, the nature of manufacturing, and the contribution of manufacturing to employment and productivity growth (Naudé 2017). For instance, given the centrality of computers and data, locations with strong connectivity, ICT software and hardware, large availability of quality data, and availability of highly  The term “Industrie 4.0” is said to have been coined by Henning Kagermann from the German Academy of Science and Engineering (The Economist 2015). 1

W. Naudé (*) Maastricht University and MSM, Maastricht, The Netherlands RWTH Aachen University, Aachen, Germany e-mail: [email protected] A. Surdej Cracow University of Economics, Kraków, Poland e-mail: [email protected] M. Cameron Trade Advisory Research (Pty) Ltd, Pretoria, South Africa e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_10

153

154

W. Naudé et al.

skilled labor, with vibrant entrepreneurial ecosystems, will become even more desirable for manufacturing. In the era of I4.0, it is not low labor costs or the presence of large local markets that will attract and sustain manufacturing: it will be how amenable a location is for hosting manufacturing that can be automated and digitized. As such, countries and regions should ask themselves, how ready are they for industry 4.0? One of the regions in the world that face an imperative to be ready for I4.0 is the eight Central and Eastern European Countries (CEEC8)2 that have joined the EU in the early 2000s. By 2017 these eight countries were home to 98 million people with a combined GDP of US$ 2,7 trillion. When in the early 1990s these countries transitioned from socialism to the free market, they went through significant restructuring, including large numbers of jobs being shed from uncompetitive Soviet-era industries. Eventually all the CEEC8 achieved relatively high economic growth rates and experienced a gradual convergence in per capita incomes with those in Western Europe. Manufacturing in particular recovered, mostly due to an inflow of FDI, much of it in the automotive sector, which was attracted by lower labor costs, good skills, and improving local business conditions in the CEEC8. This offshoring, largely from West European countries, has been described as “invest east, export west.” By 2017 the average contribution of manufacturing to these eight countries’ GDP was at 20 percent higher than the EU average of 15 percent. If the region does not absorb and apply the technologies of I4.0, its international competitiveness may suffer. Its labor and local markets may not be attractive enough to attract or maintain further manufacturing investment. In addition, in some CEEC8 countries, such as the Czech Republic, a shortage of labor is already proving a constraint on manufacturing. Without adopting I4.0, the region could again experience deindustrialization as in the early 1990s. In this chapter we determine how ready the CEEC8 is for I4.0. What we mean by readiness in this context is how possible it is, or will be, for manufacturing firms to identify, absorb, and successfully apply the technologies and techniques that are characteristic of I4.0? We do so using measures reflecting three key dimensions of I4.0 readiness, technological, entrepreneurial, and governance competencies, and measure each using a broad array of variables and calculate a composite distance normalization index in order to rank the countries relatively in terms of these three competencies. To the best of our knowledge, this is the first time that such a fairly comprehensive approach has been used to determine the comparative readiness of a group of countries to I4.0. The rest of the chapter is structured as follows. In section “Industry 4.0 and Manufacturing Competitiveness” we discuss the nature of I4.0, from which we identify adequate measures for the preparedness of countries. In section “How Ready Are the CEEC8 for I4.0?” we apply these measures and rank the countries. We find that the Czech Republic, Lithuania, Hungary, and Slovenia are most I4.0 ready and that Bulgaria, Slovakia, Romania, and Poland are the least ready. In sec The eight who joined the EU were the Czech Republic, Hungary, Lithuania, Poland, the Slovak Republic and Slovenia (in 2004), and Bulgaria and Romania (in 2007). The Czech Republic, Hungary, Poland, and Slovakia are also known the Visegrad countries (V4); see http://www.visegradgroup.eu 2

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

155

tion “Industrial Policy Implications” we make a number of policy recommendations; in particular we call for the countries to do more to (i) promote entrepreneurship, to (ii) diversify and grow manufacturing export markets through focused trade facilitation and competitive exchange rates, and to (iii) cooperate regionally on industrial policy – through, for instance, establishing a regional CEEC I4.0 Platform. The chapter concludes with a summary.

Industry 4.0 and Manufacturing Competitiveness I4.0 is different from previous industrial revolutions in that its technologies are leading to an integration of the physical (material) and digital aspects of production and consumption. Key technologies are the Internet of Things (IoT), advanced materials, digital platforms, robotics, artificial intelligence, the Interface of Things, and big data analytics (Naudé 2017). Table 10.1 lists and explains these technologies. The integration of physical and digital worlds results in mass customization through 3D printing (additive manufacturing), production-as-a-service, and new business models such as the sharing and on-demand economies. Cost reductions in computing power, data storage, and bandwidth are facilitating this integration and the spread of the technologies (Deloitte 2018). The 4IR is also different than previous industrial “revolutions” because the speed of change is exponential, not linear (Deloitte 2018). One way to view I4.0 is as new business models made possible by a combination of digital technologies, new materials, and new processes, as discussed by the OECD (2016). The digital technologies and new materials are discussed in Table 10.1. As far as the processes are concerned, I4.0 is allowing new production processes. For instance, in manufacturing these processes can refer to the democratization and dematerialization of production (Diamandis and Kotler, 2012). These processes depend on entrepreneurial capabilities and governance capabilities in order to be harnessed. An example of the democratizing effect is reflected in the rise of the maker movement. The maker movement consists of small and microenterprises making use of 3D printing and online e-commerce platforms such as Etsy or Amazon Web Services (AWS) to design and deliver unique products to their customers – resulting in mass customization (Anderson 2012) also termed Localized-Additive-­ Manufacturing-on-Demand (LAMD) (Graham 2018). It is par excellence an entrepreneurial movement. The dematerialization of manufacturing is facilitated by the rise of digital manufacturing through use of artificial intelligence (AI) (e.g., in predictive maintenance) and advanced materials such as nanomaterials and carbon fiber composites. The dematerialization of production means that less stock needs to be kept, products use fewer physical inputs, lasts longer and allow shared use, and allows growth of the called shared and circular economies. It depends also on entrepreneurs to drive the spread and development of the technology, but more requires deep technology and governance capabilities, for instance, in regulation of digital business, to facilitate this.

156

W. Naudé et al.

Table 10.1  Key I4.0 technologies Technology (Industrial) Internet of Things

Digital platforms

Advanced materials

Robotics

Artificial intelligence

3D printing

Interface of Things

Description and role in manufacturing The Internet of Things refers to a system of devices, networks, software platforms, and applications that makes possible for “sensors on physical objects to gather and shares information on the objects and their environment” (ECLAC 2018: 25) Applications are in optimization of production, predictive maintenance, the “servicification” of manufacturing, tracking products, automated flows, customized production. Around 8,4 billion objects were connected to the IoT by 2017 (ECLAC 2018) A digital platform is “a technology-enabled business model that creates value by facilitating exchange between two or more independent groups…built on a shared and interoperable infrastructure, fuelled by data and characterized by multi-stakeholder interactions” (ECLAC 2018: 61) Applications are in online and digital trade, software-as-services, infrastructure-­ as-­services, the on-demand economy, collaborative manufacturing and manufacturing design, customization, recruitment, and financing. The five most valued global firms in terms of market capitalization in 2017 were all platform firms, namely, Apple, Amazon, Google, Microsoft, and Facebook (ECLAC 2018) “Chemicals and materials like lightweight, high-strength metals and high-­ performance alloys, advanced ceramics and composites, critical materials, bio-based polymers, and nanomaterials” (Deloitte 2018, p. 32) Applications are in automotive and aviation manufacturing, sporting goods, wind turbine generators and batteries, building materials (e.g., coatings), and displays “Machines or systems capable of accepting high-level mission-oriented commands and performing complex tasks in a semi-structured environment with minimal human intervention” (Deloitte 2018, p.34) Applications are in assembly and packaging of products, including welding, painting, and loading, and in manufacturing of drones “The theory and development of computer systems able to perform tasks that normally require human intelligence” (Deloitte 2018: p. 36) Applications are in predictive maintenance, computer vision (e.g., quality assurance of production), automated driving, and personalizing consumption “An additive process of building objects, layer upon layer, from 3D model data” (Deloitte 2018, p. 28) Applications are in automotive and aviation design, dental printing, and medical implants. By 2014 already more than 11 percent of US manufacturers had “switched to volume production of 3-D printed parts” (Tuuli and Batten 2015:3) The Interface of Things includes “virtual reality (VR) which creates a fully immersible digital environment that replaces the user’s real-world environment augmented reality (AR) which overlays digitally-created content into the user’s real-world environment; mixed reality (MR) which seamlessly blends the user’s real-world environment and digitally created context; wearables and gesture recognition technology that enables humans to communicate and interact with a machine” (Deloitte 2018, p. 50) Applications are in virtual assembly manuals for factories, virtually designing factories and products, quality checks, instruction and training for manufacturing, and remote assistance

Source: based on Naudé (2018)

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

157

For the CEEC8 the I4.0 offers the potential to remain competitive in the light of decreasing labor supply and rising wages. Labor supply is dropping due to low fertility but also due to large outflows of workers, through migration to other EU countries, especially after 2011. Poland, for instance, saw more than 2 million workers, mostly young persons and potential workers, leave after 2004 (Piatkowski 2013). Other CEEC countries also find a shortage of labor in manufacturing as a growing challenge, for instance, manufacturing firms in the Czech Republic, Hungary, and Slovakia that reported labor shortages as a factor limiting production increased from an average of around 5 percent in 2010 to over 50 percent in 2017. One response has been to hire more immigrant workers, in Poland, for example, the number of work permits issued to immigrant workers increased from around 25,000 in 2008 to over 200,000 in 2017.3 Another response is to automate production. Thus, as reported by Szakacs (2018) “…here automation is a godsend…companies across Eastern Europe are ramping up investment in automation to cope with labour shortage.”

How Ready Are the CEEC8 for I4.0? In this section we compare and rank the CEEC8s based on a large number of measures that capture how well they are faring in terms of the technology capabilities, entrepreneurship, and governance capabilities that I4.0 adoption requires. While there have been at least two I4.0 “readiness indices” for countries4 compiled in recent times as far as we are aware (by Compagnucci et al. 2017 and by Roland Berger, 2014), the approach in this chapter is both more comprehensive, by taking a more extensive approach toward the readiness of countries, and more specific, as the comparison is only relative to each other for the CEEC8.5

Approach The approach is summarized in Diagram 10.1, which indicates that there are three broad dimensions of readiness to I4.0: (i) technological competencies, (ii) entrepreneurial and innovative competencies, and (iii) governance competencies. This reflects, as per the discussion in section “Approach”, that digital and automation technologies are central in industry 4.0. For instance, countries that already have experience with industrial robots may be better suited to be able to leverage further automation.  See https://www.mpips.gov.pl/analizy-i-raporty/cudzoziemcy-pracujacy-w-polsce-statystyki/  Dachs et al. (2017) compile an I4.0 readiness index on a firm level using data from the European Manufacturing Survey 2015, measuring readiness by the extent to which firms are using (i) digital management systems, (ii) wireless human-machine communication, and (iii) cyber-physical systems. 5  The approach can however easily be extended to a broader group of comparator countries subject to data availability. 3 4

158

W. Naudé et al.

Diagram 10.1  Dimensions of I4.0 readiness. (Source: Authors)

Technological competencies

Governance competencies

Entrepreneurial competencies

The three broad dimensions in Diagram 10.1 also reflect that the ability to identify the opportunities in these technologies, such as in providing better customer service and benefit from the circular and shared economy business models that become possible, and adapt these technologies to local circumstances, will be necessary for absorption and adoption. Finally, technology adoption and entrepreneurship do not take place in a vacuum, but in a context wherein government policies and institutions can play a facilitating (or obstructive) role. If, as per the smart specialization strategies of the EU, which are also adopted in the CEECs, the triple helix model of universities, companies, and government needs to work on the local level to develop high-tech manufacturing, then the three broad dimensions that are captured in measuring the readiness of countries are appropriate. As Diagram 10.1 suggests, the three dimensions of I4.0 readiness are not independent or separate: better technological capabilities may improve government competencies, and vice versa; similarly, countries with better entrepreneurial competencies may fare better in terms of technological competencies.

Technological Competencies Given that I4.0 is driven by new technologies, as summarized in Table 10.1, the first dimension to be considered is the technological ability of countries. To measure a country’s technological ability for purposes of I4.0, one needs measures reflecting to what extent countries are already using these technologies, and in particular one needs to measure the extent to which the countries are digitizing.

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

159

In the first regard, a wide range of such measures have been reported for the CEECs and other EU countries by Compagnucci et al. (2017) who also derived an I-Com Industry 4.0 Index for the preparedness of EU countries. They used 13 indicator variables that reflect the extent to which countries are adopting key technologies of I4.0. These 13 indicator variables include the shares of manufacturing firms that use radio frequency identification technologies (RFID), enterprise resource management (ERM) and cloud-computing services, customer relations management (CRM) systems, big data analytics (BDA), and supply-chain management (SCM) processes. Their variables also include indicators of the physical and human capital to support manufacturing firms in the use of these technologies, such as the extent of 4G coverage, the share of STEM graduates, the share of ICT specialists in total employment, the extent to which firms provide ICT training to their staff, and the share of data workers in total employment. As Compagnucci et al. (2017) compile all of these variables into a single indicator summarizing a country’s position, one need not here report the situation for each individual country but can use the indicator or index score for each of the CEEC8s. These are reported in Table 10.2, column 2. For the EU 28 that average index score is 80 (the top scoring countries are Finland and the Netherlands). Only two CEEC8 countries achieve a higher than average score – Lithuania and Slovenia – two of the higher-income countries in the group. The other countries perform below-average, and one CEEC8 country, Romania, one of the lower-income countries in the group, has the lowest score in the EU, i.e., is the least prepared country for I4.0 according to the Compagnucci et al. (2017) index. For present purposes, we add to the Compagnucci et al. (2017) index variables reflecting the state of the digital economy, security in the digital economy, and the potential ease for digital manufacturing and working with robots. In this regard, we use the Digital Tax Index, the IMD’s Digital Competitiveness Ranking, the ITU’s Global Cybersecurity Index (GCI), and the International Federation for Robotics (IFR) data on robotic use. All the indicators are shown in Table 10.2. The Digital Tax Index6 ranks countries based on how attractive they are from a taxation point of view for locating digital businesses. The average tax rate on digital businesses, taken from this index, is reported in column 3 of Table 10.2. It can be seen that this tax rate on digital business in CEEC8 countries such as Hungary, Lithuania, Romania, the Czech Republic, Slovenia, and Bulgaria is lower than the average of 10,2 percent for the EU. Moreover, it is lower than the effective tax rate on traditional business (in 2017). More specifically, the effective tax rate on digital business is lowest in Hungary with −6,85 percent, which means that investments in digital businesses are basically subsidized (Compagnucci et al. 2017). The digital tax rates are highest in Slovakia.

 The effective average tax rate on digital business reflects the tax burden on digital business (Compagnucci et al. 2017). 6

46 29 36 47 50 34

−6,85 0,44 12,63 6,62 15,09 9,51

I-Com Industry 4.0 Index Score 2017 [Highest = best] 64 78

68 85 66 53 77 84

0,534 0,504 0,622 0,585 0,362 0,343

1,11 0,06 0,51 0,20 1,93 2,21

0.76 0.75 0.73 0.63 0.74 0.85

Technological competencies Global Cybersecurity Density of industrial Composite normalized score robots in 2015 Index Score, 2017 [Highest = best] [Highest = best] [Highest = best] 0,579 0,07 0.65 0,609 2,17 0.93

Source: Authors’ compilation based on data from PWC Digital Tax Index and IMD Digital Competitiveness Index; Compagnucci et al. (2017) and Filippetti and Peyrache (2013):1016; and International Federation of Robotics data

Country Bulgaria Czech Republic Hungary Lithuania Poland Romania Slovakia Slovenia

Ranking on Digital Competitiveness Ranking, 2018 (of 63) [Lowest = best] 43 33

Effective average tax rate on digital business, 2017 (%) [Lowest = best] 9,52 7,48

Table 10.2  I4.0 Readiness in the CEEC8: technological competencies

160 W. Naudé et al.

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

161

We use the IMD’s Digital Competitiveness Ranking7 which aims to “assess the extent to which a country adopts and explores digital technologies leading to transformation in government practices, business models and society in general.” In its 2018 ranking of 63 countries, the CEEC8 were all ranked in the bottom half, from 29th to 50th position. Their rankings are shown in column 4 of Table 10.2. Slovakia is the lowest ranked country and – also as was seen – the CEEC8 with the highest tax rate on digital business. Given that digitalization is central to I4.0, the degree to which engaging in the digital world is secured from theft, fraud, and corruption, in other words secure online property rights, is becoming of rising importance. This should also be seen against the rise in cybercrime in recent years – for instance, it is estimated that in 2017, cybercrime costs the global economy US$ 600 billion (0,8 percent of global GDP) (Lewis 2018). As such, countries aiming to make headway in I4.0 will need to make cybersecurity a priority. Cybersecurity measures need to go beyond the merely technical to include training, organization process changes, legal changes, and improved cooperation. It is, as the ITU (2017:17) point out, that “cybersecurity is an ecosystem where laws, organizations, skills, cooperation and technical implementation need to be in harmony to be most effective.” The ITU’s Global Cybersecurity Index (GCI) aims to measure a country’s state of cybersecurity across the technical spectrum using 25 different indicators covering legal, organizational, capacity building, and cooperation domains. We include the Global Cybersecurity Index scores for the CEEC8 to our analysis of their I4.0 readiness – these are contained in column 5 of Table 10.2. According to the ITU (2017), none of the CEEC8 are leading in terms of their commitment to cybersecurity, although they all are “maturing” in their commitments. As can be seen from Table 10.2, the country with the highest score in cybersecurity is Poland, followed by the Czech Republic and Romania. Slovenia and Slovakia do the least well in terms of this indicator. Globally, out of 164 countries, the CEEC8s fall in the midrange in terms of their ranking on the index, between 33rd (Poland) and 83rd (Slovenia) position. A fourth measure that we add to the I4.0 index of Compagnucci et al. (2017) is a measure of the extent to which manufacturing is already seeing automation and workers are getting used to working with robots. We use the density of industrial robots per 1000 of workers reported by the IFR, and this is shown in column 6 of Table  10.2. It can be seen that the country with the highest density of industrial robots in the CEEC8s is Slovenia, followed by the Czech Republic and Slovakia. The least use of industrial robots is in Lithuania and Bulgaria. Table 10.2 summarizes all our indicators on the technological competencies of the CEEC8.

 The EU publishes a related index, the Digital Economy and Society Index (DESI), for EU member states. There is a large overlap between components of these indices. We prefer the more globally oriented IMD index, as this perspective seems more relevant given that the digital economy is predominantly global in nature. 7

162

W. Naudé et al.

By normalizing the distance between scores for the five individual rankings and scores and taking an unweighted average of the results, the “Technological Competencies Composite Normalized Score” is calculated. Based on this approach as indicated in Table 10.2, the highest ranking CEEC8 countries (and hence those at more industry 4.0 ready based on this combination index from this data) are the Czech Republic, Slovenia, and Hungary, while the lowest ranking (the least industry 4.0 ready based on this approach) are Romania, Bulgaria, and Poland.

Entrepreneurial and Innovation Competencies Whether economies are I4.0 ready depends not only on technical competencies and industrial sophistication but also on how entrepreneurial and innovative the economic agents in a country are. This is because digital infrastructure, skills and know-how, and experience with manufacturing do not necessarily translate into new products or new firms or new processes being adopted and disseminated. Adopting and disseminating I4.0 technologies and approaches in the CEEC8 will also depend on innovative entrepreneurship. Thus, in addition to the indicators in Table 10.2, we consider the following indicators, contained in Table 10.3, as measures of entrepreneurial and innovative dynamism in the CEEC8 and as additional indicators of how I4.0 ready the countries are. For present purposes, we do not consider measures of self-employment or small business prevalence as good measures of entrepreneurship. Rather, in the context of a 4th industrial revolution and its potential for creative destruction, we have more in mind the kind of Schumpeterian entrepreneurship as discussed in Henrekson and Sanandaji (2017). Thus, we use five indicators to measure and rank the extent of innovative entrepreneurship in the CEEC8. These are contained in Table 10.3. The first indicator, opportunity entrepreneurship, which is a measure of the share of early-stage entrepreneurship in a country, refers to people that are actively pursuing an opportunity and thus excludes necessity or forced entrepreneurship. We use the total entrepreneurial activity (TEA) measure for opportunity from the Global Entrepreneurship Monitor (GEM). This measure is reported in column 2 of Table  10.3. It can be seen that Poland and the Czech Republic have the largest shares of opportunity entrepreneurs and Bulgaria and Romania the smallest. Second, we use the number of billionaires per million of the population as another measure of Schumpeterian entrepreneurship, following Henrekson and Sanandaji (2017). This measure is obtained from the Forbes List of Billionaires and shown in column 3 of Table 10.3. In 2018 there were according to Forbes six billionaires in the Czech Republic, six in Poland, and one each in Hungary, Romania, and Slovakia. In terms of population size, the Czech Republic stand out, with 0,57 billionaires per million – which is higher than the Western European average of 0,42 and also higher than Germany’s average of 0,52. It is followed by Slovakia and Poland.

0,10 – 0,16 0,05 0,18 –

2,2 2,4 3,7 1,2 1,3 2,4

0,054 0,079 0,036 0,037 0,014 0,006

Venture capital (% of GDP, 2017) [Highest = best] 0,037 0,006 3,8 5,8 4,9 4,4 4,8 2,8

Labor productivity growth (% p.a.) 1993–2007 [Highest = best] 3,4 3,1 0,359 0,332 0,270 0,157 0,323 0,465

0.58 0.67 0.63 0.40 0.47 0.44

Entrepreneurship and innovation Innovation Index Score Composite normalized score (2017) [Highest = best] [Highest = best] 0,229 0.45 0,415 0.65

Source: Authors’ compilation based on data from the EU Innovation Scoreboard, Global Entrepreneurship Monitor, World Bank Doing Business database, World Bank Development Indicators Online, Filippetti and Peyrache (2013); Forbes List of the World’s Billionaires, 2018

Country Bulgaria Czech Republic Hungary Lithuania Poland Romania Slovakia Slovenia

Billionaires per million people 2018 [Highest = best] – 0,57

Opportunity TEA (%, 2017) [Highest = best] 1,0 2,7

Table 10.3  I4.0 Readiness in the CEEC8: entrepreneurship competencies

10  Ready for Industry 4.0? The Case of Central and Eastern Europe 163

164

W. Naudé et al.

Third, we use the extent of the availability and use of venture capital (VC) in a country as percentage of GDP. Venture capital (VC) is seen as a good indicator for high-tech entrepreneurship, as this has become a primary means of funding high-­ tech start-ups (Bocken 2015). This indicator is shown in column 4 of Table 10.3. Lithuania and Hungary stand out, where the VC as GDP share is the highest, with Slovenia and the Czech Republic the lowest. Fourth, we use labor productivity growth, measured as growth in output per worker, as an indicator for the ability of entrepreneurs to commercialize technology. The more and better capital and technology a worker has access to, including “managerial” and other intangible technology (such as firm routines), the higher will labor productivity be. Hence, if lagging countries are closing the technological gap, through their entrepreneurs being able to turn ideas and inventions into innovations, this will reflect in growth in labor productivity. Filippetti and Peyrache (2013) calculate labor productivity for the CEEC8 for the period 1993–2007, and this is shown in column 5 of Table 10.3. This shows that all CEEC in our sample achieved relatively high labor productivity growth rates from 1993 to 2007, indicating that they were closing the technological gap. The average for the eight CEECs over this period was 4,4 percent, which exceeded by a fair margin the average labor productivity growth in the old EU member states, which was on average 1,8 percent over this period (although of course the old EU member states had almost double of the levels of labor productivity of the new EU member states by 2007). The countries with the highest labor productivity growth were among the higher- and middle-income countries in the group, i.e., catch-up, Lithuania, Poland, Slovakia, and Hungary. Finally, we report on the EU Innovation Scoreboard’s Innovation Index the scores for the CEEC8 – see column 6 of Table 10.3. The EU’s Innovation Index measures broadly the innovativeness of the economy using a wide variety of measures, including R&D expenditure, patent applications, exports of high-tech products, process innovations adopted by SMEs, and others. Slovenia and the Czech Republic are the leaders in terms of innovation in the CEEC8, and Romania and Bulgaria have the lowest scores. Again, we normalize the distance between scores for the five individual rankings and scores and take an unweighted average of the results, termed the “Entrepreneurship Competencies Composite Normalized Score.” Based on this approach as indicated in Table 10.3, the most I4.0-ready country measured on these dimensions is Lithuania, followed by the Czech Republic and Poland. The lowest relative scores are obtained by Romania, Slovenia, and Bulgaria.

Governance Competencies The governance competencies that are most relevant to support I4.0 would be those that support agile manufacturing, entrepreneurial start-ups, and process innovations. These types of competencies can be measured through five variables, as contained in Table 10.4. These are first, the country’s leaderships’ political management

0,51 0,98 0,63 −0,17 0,81 1,17

4,44 7,18 6,25 5,89 6,70 6,78

0,19 0,01 0,05 0,03 0,02 0,07

Public support for business R&D, % of GDP (20145) [Highest = best] 0,01 0,08 48 16 27 45 39 37

Rank on Doing Business Index (2018) [Lowest = best] 50 30

Source: Authors’ compilation based on data from Eurostat, The World Bank and Bertelsmann Stiftung

Country Bulgaria Czech Republic Hungary Lithuania Poland Romania Slovakia Slovenia

Government Effectiveness Score 2017 [Highest = best] 0,26 1,02

Bertelsmann Governance Index 2017 [Highest = best] 5,98 7,03

Table 10.4  I4.0 Readiness in the CEEC8: governance competencies

47 48 31 9 47 50

Citizens using public e-services (%, 2017) [Highest = best] 21 46

0.73 0.79 0.66 0.41 0.67 0.78

Governance competencies Composite normalized score [Highest = best] 0.45 0.80

10  Ready for Industry 4.0? The Case of Central and Eastern Europe 165

166

W. Naudé et al.

skills. The Bertelsmann Governance Index ranks the countries “according to their leadership’s political management performance between February 2015 and January 2017.” We show this index for the CEEC8 in Table 10.4, column 2. Lithuania and the Czech Republic score the highest in terms of this index and Hungary and Bulgaria the lowest. Second, we use the World Bank’s Government Effectiveness Index, which “captures perceptions of the quality of public services, the quality of the civil service and the degree of its independence from political pressures, the quality of policy formulation and implementation, and the credibility of the government's commitment to such policies” (World Bank Governance Indicators, online). Here countries can score between −2.5 (very poor) and 2,5 (very good). In Table 10.4, column 3, we show this indicator for the present countries. It can be seen that Slovenia and the Czech Republic are ranked the best in terms of government effectiveness and Romania and Bulgaria the worst. Third, we measure how the government supports innovative entrepreneurship. Here, we use two variables: the extent to which there is public support for business R&D (as percentage of GDP) and the ranking on the World Bank’s Doing Business Index. Table 10.4 shows in this regard that most public support for business R&D is in Hungary and the Czech Republic, with the least support in Lithuania and Bulgaria. In terms of the ease of doing business, the most highly ranked country is Lithuania (in 16th place in the World Bank’s Index) by quite margin over the other CEEC8. It is followed by Poland (27th rank) and the Czech Republic (30th rank). The lowest ranked in terms of doing business are Bulgaria (50th ranked) and Hungary (48th ranked). Finally, we measure how the government serves its customers (citizens) through offering digital services. Given the predominance of the digital economy in the I4.0, it is also imperative that government be able to act and interact in the digital domain. We use Eurostat’s measure from its community survey on the percentage of citizens using public e-services to interact with government. This is shown in column 6 of Table 10.4. Slovenia and Lithuania top the league, and Romania and Bulgaria lag behind, the former quite substantially, with only 9 percent of citizens using public e-services. Based on the same normalization approach, we find that in terms of the “Governance Competencies Composite Normalized Score,” the Czech Republic relatively performs best, followed by Lithuania and Slovenia. Romania, Bulgaria, and Poland are in the lowest positions.

Total I4.0 Readiness Ranking For each of the dimensions in Diagram 10.1 and the composite normalized score (between 0 and 1) based on the various indicators in Tables 10.2, 10.3 and 10.4 in the previous sections, we ranked the CEEC8 countries and took a simple average ranking to determine which of the CEEC8 are relatively to the others more or less

167

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

1

Czech Republic Overall Rank = 1

Romania Overall Rank = 8

Bulgaria Overall Rank = 7

Lithuania Overall Rank = 2

0

Hungary Overall Rank = 3

Slovenia Overall Rank = 4

Slovakia Overall Rank = 6 Poland Overall Rank = 5 Technological Competencies Governance Competencies

Entrepreneurship and Innovation Overall

Fig. 10.1  I4.0 Readiness in the CEEC8 per dimension. (Source: Authors’ compilation)

ready for I4.0. In Fig. 10.1 we depict the outcomes for the CEEC8 in terms of the three dimensions, using a radar chart. From Fig. 10.1 can be seen that the Czech Republic is ranked in first place, followed, respectively, by Lithuania, Hungary, Slovenia, Poland, Slovakia, Bulgaria, and Romania in the 2nd to 8th positions. The radar chart in Fig. 10.1 is useful to show how the various countries rank in terms of the three dimensions. Thus, for instance, in technological competencies, the Czech Republic leads, while in governance competencies, it is Lithuania that is ranked first. Similarly, all countries seem to be doing least well in terms of entrepreneurial competencies, especially Slovenia and Bulgaria. This indicates that there is no one recipe for all CEEC8 to improve their I4.0 readiness: all will have to focus on the three dimensions of I4.0 readiness, but different dimensions may have to be prioritized in different countries. Overall, the conclusion is that within the CEEC8 group, the Czech Republic, Lithuania, and Hungary are most I4.0 ready and that Bulgaria and Romania the least. This is good news for the Czech Republic, which also has the 2nd largest manufacturing sector (in terms of output) after Poland as well as the largest relative contribution of manufacturing to GDP in the region (see Fig. 10.2).

168

W. Naudé et al.

Fig. 10.2  Real manufacturing output in the CEEC8, 2008–2017 (billions of US$). (Source: Authors’ compilation based on World Bank Development Indicators Online)

It is not such good news for especially Romania, which after Poland and the Czech Republic has the most substantial manufacturing sector in the region. Poland, with the largest manufacturing sector, is overall ranked at position 5 only after much smaller manufacturing economies such as Lithuania and Hungary. Furthermore, a concern for Bulgaria is that it has already been noted by the EU as starting to fall behind in terms of digitizing its economy (as measured by the EU’s Digital Scoreboard 2016) and hence may find itself diverging from the other CEEC8  in terms of industrialization. The less-ready countries with substantial manufacturing have potentially much to lose through reshoring and offshoring and declining international demand for their manufacturing production if they are not able to provide a more competitive environment for local I4.0. In the next section, we discuss implications for industrial policies.

Industrial Policy Implications Many European countries have developed industrial policies and strategies to implement I4.0. In Western Europe the major initiatives include “Platform Industrie 4.0” (Germany), “Alliance pour L’Industrie du Future” (France), “Industria 4.0” (Italy),

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

169

“Produktion 2030” (Sweden), “Industria Conectada 4.0” (Spain), and “Catapult High Value Manufacturing” (UK) (see European Commission 2017). The CEEC8s have in recent years (mainly from around 2015 to 2016) followed suit with their own I4.0 initiatives. These are summarized in Table 10.5. From analyzing these initiatives, a number of comments are in order, which we will frame with respect to our approach (see Diagram 10.1) of the three dimensions of technology, entrepreneurship, and governance. First, in almost all countries, initiatives follow the German “platform” approach, wherein the I4.0 initiative aims to create a dialogue and cooperation between various stakeholders, such as government, industry, and the science and education sectors. In this respect, and also in efforts in all the countries to improve the e-Government, existing policies give attention to the governance dimension of I4.0 readiness (see Diagram 10.1). Second, in all of the region’s I4.0 initiatives, the focus is on digitization, technology diffusion into industry, and digital skills development. This is consistent with the technological capabilities dimension of the I4.0 readiness. As in the case of other EU countries’ initiatives (see European Commission 2017), the CEEC8 focus herein on expanding Internet infrastructure and the penetration of the Internet of Things for industrial use. In Romania in fact this is the essence of the country’s current strategic approach to deal with I4.0, with relatively little emphasis on the hardware and industrial aspects of I4.0 (see Table 10.5). Third (with the exception of Poland), the initiatives tend to neglect the entrepreneurial capacity dimension. This is also a feature of the I4.0 initiatives in other EU countries. It is not the case that these countries do not have initiatives to promote entrepreneurship or improve the business environment8; rather these are not integrated or coordinated with their I4.0 strategies and, moreover, tend to neglect vital aspects for technology entrepreneurship, such as venture capital provision and the promotion of entrepreneurship to commercialize inventions and to find new opportunities for exporting and new export markets. For instance, the Czech Republic, which is top ranked in terms of both technological capabilities and governance, falls to second place when it comes to entrepreneurial capabilities. Relatedly, R&D in the region is indeed comparatively low, and innovation activities in the CEECs differ significantly from that in the West. Around 55 percent of innovation expenditures in 2010–2012  in the CEECs were on the acquisition of machinery, equipment, and software and only 39 percent on R&D. In comparison, Western EU countries spend 19 percent of innovation expenditure on the acquisition of machinery, equipment, and software and 73 percent on R&D (Radosevic 2017). In the context of I4.0, it may be argued that more R&D, focusing on adapting and generating technologies for local specifications, and the commercialization thereof, through support for entrepreneurial start-up ecosystems, wherein entrepreneurial skills and entrepreneurial finance are key ingredients, need more attention. This  Indeed, reform of the business environment, for instance, to make it easier to do business, has been a common theme in the countries’ policies at the end of the socialist era (Stojcić and Aralica 2017). 8

Key elements Ambition the “Digital Transformation of Bulgarian Industry, 2017–2030” Focus: Innovation and technology diffusion. Strengthen relationships between science and industry. Skills and capacity building A platform for industry-government interaction Focus: Data and communication infrastructure, education and skills, flexible labor markets, global supply chains IPAR 4.0 National Technology Platform/ A platform for industry-government interaction Irinyi Plan Ambition to be one of the most industrial economies in Europe and raise share of industry in GDP to 30 percent by 2020 Focus: Digital technology, digital transformation of industry, skills, export growth Pramonė 4.0 (See http://www.industrie40.lt/ A platform for industry-government interaction platform/) Focus: Technology, infrastructure, and digital skills Future Industry Platform A platform for industry-government interaction Raising awareness and demonstrate I4.0 technologies Focus: Technology, digital transformation, and business development, including SMEs Focus: ICT infrastructure, digital skills, Internet penetration, e-Government services National Strategy for Romanian Digital Agenda 2020 (See http://gov.ro/en/ government/cabinet-meeting/nationalstrategy-on-the-digital-agenda-forromania-2020 ) Smart Industry Platform A platform for industry-government interaction Focus: Technology adoption, R&D, education and skills, awareness of smart manufacturing Slovenian Digital Coalition/Slovenian A platform for industry-government interaction Industrial Policy 2013 Focus: Digital skills, digitization of industry, and digital regulation

Strategy Kontseptsia Industria 4.0 (See https://www. mi.government.bg/files/useruploads/files/ip/ kontseptsia_industria_4.0.pdf) Průmysl 4.0

Source: Authors’ compilation based on European Commission (2017) and EC Digital Transformation Monitor Online (See https://ec.europa.eu/growth/toolsdatabases/dem/monitor/category/national-initiatives)

Slovenia

Slovakia

Romania

Poland

Lithuania

Hungary

Czech Republic

Country Bulgaria

Table 10.5  National industry 4.0 industrial strategies in CEEC8

170 W. Naudé et al.

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

171

needs to be stressed in the environment of I4.0 where countries may expect less FDI from the West and less manufacturing exports to the West, as these countries reshore their production. In the past, it was FDI and trade that brought in technology (Stojcić and Aralica 2017) – in the future the local entrepreneurship and innovation systems, focused moreover on export diversification, will need to play a greater role. With regard to incentives and the promotion of the diversification of manufacturing exports toward nontraditional (non-EU) markets and to keep manufacturing exports to EU markets more competitive, the CEEC8 need to consider exchange rate policy as a tool to promote I4.0. None of the current I4.0 strategies in the region considers this. This may be a significant lacuna, given that maintenance of a competitive exchange rate (i.e., undervalued) is an important industrial policy tool – it has played a critical role in the industrial development of China, and at present it is the Euro exchange rate which is conferring a significant competitive advantage for Germany, to maintain its strong exporting of manufacturing products. The CEEC8 need to grow their exports of manufacturing goods more aggressively, given that the current model of Western investment in CEEC8 manufacturing for exporting to the West may be coming to an end. For instance, Fig.  10.3 shows the sudden decline between 2008 and 2009  in exports from the CEEC8 during the global financial crisis. It also shows that export levels recovered by 2011, but from then on tended to be rather constant with no strong growth as in the previous period. High-skill manufacturing growth is largely stagnant and a small share of total manufactured exports. Export competitiveness is significantly influenced by a country’s exchange rate. Many of the newly industrialized Asian countries, as well as China in recent decades, had used maintenance of an undervalued exchange rate in order to promote manufactured exports. Also, at present Germany benefits from having the Euro, which provides it in essence with an undervalued currency. In the CEECs, three countries have adopted the Euro,9 in effect removing exchange rate devaluation as a policy tool. The problem for the CEECs in this respect however is that since 1994 the CEEC8 real effective exchange rates10 all appreciated – in effect reducing the international competitiveness of their manufactured exports. This is shown in Fig. 10.4. Moreover, Fig. 10.4 shows that in recent years, especially after the global financial crisis, a number of CEECs have seen their real broad effective exchange rates markedly appreciate against the Euro. By end-2017, the CEECs with their own currencies followed flexible exchange rate regimes and inflating targeting, the only exception being Bulgaria, which maintained a fixed exchange rate through operating a currency board.  Slovenia became the first CEEC8 to adopt the Euro, in 2007. It was followed by Slovakia in 2009 and Lithuania in 2015. 10  The real effective exchange rates (REERs) are based on weighted averages of bilateral exchange rates of major trading partners as well relative (trading partners’) consumer prices. This implies that while Slovenia, Slovakia, and Lithuania adopted the Euro, disparate relative price levels and relative different trade patterns with a variety of trading partners influences the outcomes for each country’s REER. Policy makers therefore also need to consider relative terms of trade with partners and comparative consumer prices. 9

172

W. Naudé et al.

Fig. 10.3  Aggregated CEEC8 exports by skill and technology intensity, 2008–2016. (Source: Authors’ compilation (The international trade data applied in this study are based on the Base Analytique du Commerce International (BACI) data set which is a reconciled version of the UN COMTRADE database provided by CEPII (Centre d’Études Prospectives et d’Informations Internationales), 2017 – HS 2007 revision, in constant 2000 US$ values based on the annual average consumer price index research series published by the US Census Bureau and skills and technology classifications based on the initial work by Basu (2011); Basu and Das (2011)))

Fig. 10.4  Real broad effective exchange rate for the CEEC8, 1994–2017. (Source: Authors’ compilation based on Bank for International Settlements, retrieved from FRED, Federal Reserve Bank of St. Louis)

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

173

A further focus (combined with exchange rates where feasible) should also be on the identification11 of alternative new markets and products for CEEC8 manufactured exports. A concerted effort to identify and develop opportunities in nontraditional markets (in addition to that of Western Europe) is strategically important. Finally, the current I4.0 initiatives in the region are, like those in other EU countries, largely funded by government, in programs that hope to become self-­ sustainable in the future by private sector funding, and, furthermore, are characterized by little coordination and cooperation between countries. Here, our recommendation is that the CEEC8 countries seek more structural and long-term funding for their I4.0 initiatives, including through supporting EU-level initiatives12 to tax large international digital platform-based firms (such as Amazon, Google, Facebook, etc.) based on the revenue that they generate from a particular country (the proposed “equalization tax”). In this, and more generally in addressing challenges to the implementation of I4.0 posed by labor market and skills shortages, greater cooperation between CEEC8 is recommended, perhaps through a regional “CEEC I4.0 Platform.”

Summary and Concluding Remarks What is the future of manufacturing in Central and Eastern Europe in light of industry 4.0 (I4.0)? Can they benefit from its technologies to combat the loss in competitiveness that labor market scarcity and rising labor costs are causing? Are these countries ready to implement these new technologies? To answer this question, we explained the nature of I4.0 so as to identify measures for the readiness of the countries. We emphasized the premium that I4.0 places on technologies and business models (entrepreneurship) requires agility, flexibility, and customer orientation and the implication that this has for the location decision of manufacturing firms and plants and the supporting ecosystem (governance). Using measures reflecting the three key dimensions of I4.0 readiness, namely, the mentioned technological, entrepreneurial, and governance competencies, we find that the Czech Republic, Lithuania, Hungary, and Slovenia are most I4.0 ready and that Poland, Slovakia, Bulgaria, and Romania are the least ready in relative CEEC8 terms. For the latter this is of a concern, especially given that Poland and Romania have among the most substantial manufacturing sectors in the region: hence stand much to lose if they are not ready for I4.0. Finally, we made a number of recommendations for these countries’ industrial policies. Although each country should craft its own strategic response in terms of where it stands with respect I4.0 readiness, we called for all the countries in the

 For example, one of the approaches to identify and investigate potential “unusual” suspects is the TRADE-DSM methodology. See, e.g., Cuyvers et al. (2012) and Cuyvers et al. (2017). 12  See, e.g., https://euobserver.com/economic/138954 11

174

W. Naudé et al.

region to do more to promote entrepreneurial skills and to diversify and grow both manufacturing products and export markets. In this regard, we commented on a number of perceived gaps in the countries’ I4.0 strategies and policies. One is that the initiatives tend to neglect the entrepreneurial capacity dimension, in particular vital aspects for technology entrepreneurship, such as venture capital provision and the promotion of entrepreneurship to commercialize inventions and to find new opportunities both in the production of products for exporting and new export markets. In the past, it was FDI and trade that brought in technology. In the future the local entrepreneurship and innovation systems, focused moreover on export diversification, will play a greater role. A related gap identified is in the linking of industrial policy for I4.0 with trade and exchange rate policies. To promote the diversification of manufacturing exports toward nontraditional (and non-EU) markets and to keep manufacturing exports to EU markets more competitive, the countries in the region outside the eurozone need to consider exchange rate policy. All countries certainly should focus on export diversification/promotion as tools to promote I4.0. It would seem that none of the current I4.0 strategies in the region explicitly consider these potential approaches. This may be a significant lacuna, given that maintenance of a competitive exchange rate and export promotion/diversification are important industrial policy tools. In all CEEC8 members developments exchange rate appreciation occurred after 1994 with the result that today many of CEEC8 countries have to face I4.0 without the competitiveness that exchange rate policy can bring. Finally, we concluded this chapter by recommending that the CEEC8 countries should seek more structural and long-term funding for their I4.0 initiatives, including through supporting EU-level initiatives to tax large international digital platform-­based firms (such as Amazon, Google, Facebook, etc.) based on the revenue that they generate from a particular country (the proposed “equalization tax”). In this, and more generally in addressing challenges to the implementation of I4.0 posed by labor market and skills shortages, greater cooperation between CEEC8 is recommended, perhaps through a regional “CEEC I4.0 Platform.”

References Anderson, C. (2012). Makers: The new industrial revolution. London: Random House. Basu, S. R. (2011). Retooling trade policy in developing countries: Does technology intensity of exports matter for GDP per capita. Policy issues in International Trade and Commodities, 56. Basu, S. R., & Das, M. (2011). Export structure and economic performance in developing countries: Evidence from nonparametric methodology (No. 48). United Nations Conference on Trade and Development. Bocken, N. M. T. (2015). Sustainable venture capital – catalyst for sustainable start-up success? Journal of Cleaner Production, 108, 647–658. Brynjolfsson, E. and McAfee, A. (2015). Will Humans Go the Way of Horses? Foreign Affairs, 94, 8–14.

10  Ready for Industry 4.0? The Case of Central and Eastern Europe

175

Compagnucci, S., Berni, G., Massaro, G., & Masulli, M. (2017). Thinking the future of European Industry: Digitalization, Industry 4.0 and the role of EU and national policies, EU Study 3.17. iCom, institute for competitiveness. Cuyvers, L., Steenkamp, E., & Viviers, W. (2012). The methodology of the decision support model. In L. Cuyvers & W. Viviers (Eds.), Export promotion – a decision support model approach (pp. 57–83). Stellenbosch: SUN Press. Cuyvers, L., Steenkamp, E., Viviers, W., Rossouw, R., & Cameron, M. (2017). Identifying Thailand's high-potential export opportunities in ASEAN+3 countries. Journal of International Trade Law and Policy, 16(1), 2–33. (Available from Emerald Insight). Dachs, B., Kinkel, S. and Jäger, A. (2017). Bringing it All Back Home? Backshoring of Manufacturing Activities and the Adoption of Industry 4.0 Technologies, MPRA Paper no. 83167. Deloitte. (2018). Global manufacturing competitiveness index. Diamandis, P.H. and Kotler, S. (2012). Abundance: The Future is Better than you Think. New York: Free Press. ECLAC. (2018). Data, algorithms and policies: Redefining the digital world. Santiago: Economic Commission for Latin America and the Caribbean. European Commission. (2017). Key Lessons from National Industry 4.0 Policy Initiatives in Europe. Digital Transformation Monitor. May. Filippetti, A., & Peyrache, A. (2013). Is the Convergence Party Over? Labour productivity and the technological gap in Europe. Journal of Common Market Studies, 51(6), 1006–1022. Graham, A. (2018). 8 crucial manufacturing predictions for 2018, Association for Manufacturing Excellence, 2 February. Henrekson, M., & Sanandaji, T. (2017). Schumpeterian Entrepreneurship in Europe Compared to Other Industrialized Regions, IFN Working Paper no. 1170. ITU. (2017). Global Cybersecurity Index 2017. Lewis, J. (2018). Economic impact of cybercrime – no slowing down. McAfee, Feb. Marsh, P. (2012). The new industrial revolution: Consumers, globalization and the end of mass production. New Haven: Yale University Press. McKinsey Digital. (2015). How to navigate digitization of the manufacturing sector. McKinsey and Company. At: https://www.mckinsey.com/business-functions/operations/our-insights/ industry-four-point-o-how-to-navigae-the-digitization-of-the-manufacturing-sector. Accessed 10-10-2018. Naudé, W. (2017). Entrepreneurship, education and the fourth industrial revolution in Africa, IZA Discussion Paper no. 10855, Bonn: IZA Institute of Labor Economics. Naudé, W. (2018). Brilliant technologies and brave entrepreneurs: A new narrative for African manufacturing. Journal of International Affairs, forthcoming. OECD. (2016). Enabling the next production revolution: The future of manufacturing and services – interim report. Paris: Organisation for Economic Cooperation and Development. Piatkowski, M. (2013). Poland’s new golden age: Shifting from Europe’s periphery to its center, Policy Research Working Paper no. 6639. Washington DC: The World Bank. Radosevic, S. (2017). Upgrading technology in Central and Eastern European Economies, IZA World of Labor, 338 (Feb). Roland Berger (2014). Industry 4.0: The new industrial revolution: How Europe will succeed. Online at : https://www.rolandberger.com/en/Publications/Industry-4.0-–-the-new-industrialrevolution.html Schwab, K. (2016). The fourth industrial revolution: What it means, how to respond, World Economic Forum, 14 January. Stojcić, N., & Aralica, Z. (2017). Choosing right from wrong: Industrial policy and (De) industrialization in Central and Eastern Europe, EIZ Working Paper EIZ-WP-1703. Szakacs, G. (2018). Enter the robots: Automation fills gaps in East Europe’s Factories, Thomson Reuters, 7 Feb. The Economist (2015). Does Deutschland do Digital? The Economist Magazine, 21 Nov. Tuuli, S., & Batten, S. (2015). Back to the future: Why we’re optimists in the Secular Stagnation Debate, Bank Underground Blog, 2 July.

Chapter 11

From Big to Small Data Peter Cochrane and Ahmed Elmagarmid

Preamble • With the aid of our machines, satellites, instruments, devices and networks, we are creating more data than ever before in our history. And it is no exaggeration to assert that the future of industry, commerce, countries, our species and the planet now hinges on data, its analysis, accurate interpretation and the knowledge revealed. Our big picture understanding of global warming, climate change, farming, vital mineral and biological resources, industry, markets and healthcare, etc. is entirely furnished by data, Big Data! • Year on year our technology creates exponentially more data, and it is easy to become overwhelmed by the volume, velocity, variety, variability and the sheer complexity. This is the Big Data conundrum! Data is now predominantly generated by our machines, communicated and distributed by our networks in volumes well beyond our abilities to analyse and, in many cases, apparently impossible to verify and understand. This has to a large extent driven the need and development of AI systems in the form of analysis engines and the amplifiers of human intellect. • It is important not to measure Big Data by how big it is, but by what it is hiding and what we can do with it. Analysing data to realise vital market information, mineral and gas deposits, important medical trends, new alloys, drug combinations, societal and political change and of course web posting and search preferences is but a short list of the positive application of Big Data.

P. Cochrane (*) · A. Elmagarmid Qatar Computing Research Institute, Doha, Qatar e-mail: [email protected] © Springer Nature Switzerland AG 2019 M. Dastbaz, P. Cochrane (eds.), Industry 4.0 and Engineering for a Sustainable Future, https://doi.org/10.1007/978-3-030-12953-8_11

177

178

P. Cochrane and A. Elmagarmid

• A major difficulty here is that it is impossible to assess the usefulness of (BoG) Big Data by just looking at it; moreover, data that turns out to be useless today may mutate into a ‘goldmine’ with the passage of time. And so there is a tendency to store and keep data because it might just be useful later. As Big Data and its analysis are really in a period of genesis, or at least infantile development, it is hoped that further experience and research will see automated data cleansing with the discarding of the useless and trivial, the preservation and storage of all that is valuable. Until that time we have little choice but to stumble forwards in something of a slightly out of control fashion measuring and recording everything just in case! • One inherent property of Big Data we have to recognise is its almost singular ability to provide macro information – a global synopsis if you will – the big picture. A good example is the Google search engine and the way it steers us to the most likely choices on the basis of the selections made by millions of users. Alternatively, an automotive manufacturer will analyse the service history of its millions of vehicles exported all over the world to identify weaknesses of design, manufacture, supply chain and repair shops. What they see is effectively a global average without regard to the nature and habits of individual drivers and a myriad of other conditions. This is also true for political voter analysis, farming data, banking, commerce and so on. Nevertheless, we cannot deny the powerful nature of Big Data and what we have learned and what it continues to reveal. 'There is nothing quite so powerful as a realy good approximation' With an emphasis on the really good!

• There now comes a second Data Genesis, a more recent development with the embedding of monitoring, recording and intelligence in all our appliances, office machinery, entertainment, wearable and mobile devices, to be complemented and greatly amplified the impending roll out of the IoT. This will be the ‘Small Data’ revolution with the recording of all the ‘small grain minutiae’ of our lives, our habits and that of all our technologies. • At this time the impact of a pending Small Data revolution is seeing a change in thinking with the potential of medicine for the individual rather than the average, search engines for me and not the whole of humanity and a market of one with product customisation, focussed marketing, online pre-selling, adaptable insurance, adaptive interfaces and more. • A big difference between Big Data and Small Data is the aggregation and concentrated storage and analysis of the former and the distributed, localised and networked nature of the latter. In most respects Small Data is the far bigger problem but rendered small by distribution and localisation. The IoT will undoubtedly be the epitome of this but most likely emergence of local centres of analysis around individual group types, campus, company operation, etc. This is already evident in some military mesh net realisations for troops, transport, weapons and systems. 'A reasonable bet might be that the IoT will eventually store more (small) data than the aggregate of all other systems on the planet'

11  From Big to Small Data

179

Perspective So far the mining and analysis of data have been dominated by ‘the big’ looking at ‘the small’ – businesses, institutions and governments looking at society with an eye to welfare, security and commercial opportunities on a macroscale (Han 2011). The impact of Big Data has been to expand this into the arena of networking and association with a focus on more relational services, advertising, ‘pre-selling’, healthcare, security and tax avoidance (Marr 2015). But this currently leaves the critical arena of Small Data to be addressed – the small looking at the small and individuals and things examining and exploiting their own data (Lindstrom 2016). Moreover, it is an essential component for personalised care, individual health and more generally the migration to markets able to focus on the precise needs of the individual instead of an average based on a global population. This wider scenario automatically challenges today’s networks and modes of operation. For example, will 1000 vehicles in a traffic jam seek to communicate via 3, 4, 5G networks, or will they talk to each other daisy chain fashion? Will all the white and brown goods in our homes, and health equipment in our bathrooms, seek out the Internet, or will they talk to each other? The reality is; the IoT operating over mobile networks will not scale, and most of these things predominately need to intra-communicate and not extra-communicate. This we see as a network and operational game changer beyond today’s Internet (Baesens 2014; Cochrane and Moschella 2015), automatically leading to new forms of analysis and sharing. In short a ‘society of the dumb’ or ‘network of very simple’ can rapidly assume levels of useful intelligence. This is borne out by biology (i.e. ants or the single cells building a human) and indeed swarm robotics.

Information Growth Our ability to collect, store, organise, analyse and rationalise data and make wise decisions is now challenged by the sheer volume and speed of change. Long-­ established paradigms of centralisation are breaking down, and in many instances, localised storage and processing are the only viable options (Shultz 2011). This is in line with biological precedents, the most likely path for a projected Internet of Things (IoT) spanning 50–250 Bn coexisting items. The likely depth and breadth of the IoT data to be generated and stored (Fig. 11.1 and Table 11.1) will present many new challenges and demand new modes of thinking, analysis, networking and operation (Evans 2011). Whilst much of the IoT will be autonomous and largely invisible to us, a small proportion will demand our attention to render more personal advantage along with that afforded to business, institutions and government (Purdy and Davarzani 2016). Not so obvious are the necessity of the IoT as a component in the Green Agenda and the realisation of ecologically sustainable societies (Boulos and Al-Shorbaji 2014). Living audit trails of material sourcing, transportation, processing, production,

180

P. Cochrane and A. Elmagarmid

Fig. 11.1  Projected IoT-embedded information storage growth

delivery, use and maintenance and eventual disposal through reusing, repurposing and recycling are the only way we can achieve a near-zero material and energy wastage future (Louchez and Thomas 2014). This also applies farming, food supplies, commodities and consumables. It has been estimated that logistics operating loss world-wide is ~$3Tn per year due to a lack of timely information. For example, shipping container sharing, tracking and inventory control is almost non-existent as the industry does not have the necessary technology. Customers ask: where are my goods? But no one is sure! Why? Ships and trucks are tracked, but containers, pallets and boxes are not (Dey et al. 2011). However, if containers, pallets, boxes and packages were electronically tagged and could talk to each other, the logistics world will be changed dramatically. Efficiency would increase, losses would reduce, and customer confidence and economics would improve (Colliers International 2015). Thought industry and commerce just-in-time is now the norm with customer demand and attitudes largely driving the competition. And so Big Data has been both a necessary and vital component and enabler as exemplified by the logistics case above. But in the next phase, it will be Small Data that will assume the primary component. Big Data now satisfies the macro production and delivery, whilst Small Data is expected to deliver new levels of customisation. I want what I want; I want it now; I want it in my style and colour; and this is what I will pay - or I will go get it elsewhere

Finally, we observe that communication and information availability always invoke change, but in the case of the IoT, that availability will be dominated by Small Data. This new dimension will promote a fundamental rethink of much of what we do at an individual, organisational and societal level (Fig. 11.2). To illustrate how these changes are likely to evolve, we examine a small number of exemplars offering big opportunities.

11  From Big to Small Data

181

Table 11.1  The Small Data springboard based on the devices and data available in 2016 ITEMS Smart watch Google glass Wearable health Smart Cannera Haelth scanner Blood pressure meter Glucose tester Thermometer Balhroom scales Mobile phone Tablet Laptop PC Radio TV HiFI Τoaster Coffee machine Kettle/water heater Hob Oven Microwave Washing machine Tumbler dryer Vacuum Central healing Security system Scooter Motorbike Car Total memory

Min Memory 0.25 12 1 4 1 0.1 0.01 0.01 0.01 8 16 256

Max Memory 8 16 8 32 64 0.25 0.1 0.1 0.1 32 256 1000

0.01 0.1 0.01 0 0 0 0 0 0 0 0 0 0 0 0 0 0.1 298.6

0.1 16 16 0.01 0.1 0.01 0.01 0.1 0.01 0.1 0.01 0.01 32 32 1 16 1000 2530.01

Small Data ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓

Big Data ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✓ ✓ ✓ ✓ ✗ ✓ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✗ ✓ ✓ ✓ ✓ ✓

High Growth ? ? ? ✓ ✓ ? ? ? ? ✓ ✓ ✓ ? ? ? ? ? ? ? ? ? ? ? ? ? ? ✓ ✓ ✓ ✓

Industry Today’s production processes and supply chains are a very polished and efficient version of their Victorian forebears: are automated, faster and slicker, have higher precision, better performance and reduced material and energy and are more reliable for sure, but fundamentally unable, to deliver sustainable futures (Panigrahy et al. 2011). That will take for more than a veneer of greater efficiency! The future demands reductions in material and energy use whilst eradicating much of the distribution, supply and support costs along with efficient recycling (Clancy 2015). But what is the purpose all these if old equipment consumes 10 W when the new requires less than 1 W? Ultimately the route to sustainability has to be driven by science and thermodynamics.

182

P. Cochrane and A. Elmagarmid

Fig. 11.2  Human progress with technology and data expansion and growth

So, we need the Big and Small Data to track materials from their point of extraction, through transport and refinement to processing and manufacture and to the distributor, seller, buyer/user, support and maintenance organisation, disposal, and finally, fully eco-sensible recycling. This is to be almost every mm, cm, m and km of the journey (Kanth 2011) with designers, producers and users engaged in a closed loop of creative thinking and continuous improvement (Fig. 11.3). More laterally, a full record of production and use protects against counterfeit products whilst facilitating spare part recovery and reuse in the face of other failures such as accidents, fire-damaged goods and misplaced, rare and hard-to-find items. The military already put this approach to good use in maximising the number of working vehicles, aircraft and weapon systems in real time and often under duress. (The) 'This is effectively high-tech cannibalisation'

The farming sector is well advanced in tracking, monitoring and analysing technologies for animal care (Floyd 2015) and the maximal use of large plant items with the sharing of combines and other high-cost plant items. They also employ sophisticated food mixing to adapt animal diets to weather conditions along with growth, pregnancy and milk production cycles. Such practices are equally relevant to industry where raw materials, machines and facilities can be shared right down to plastics, sheets of metal, specific 3D printers, millers or shapers. Less equipment and fewer people standing idle, greater output for less energy, the efficient reclaim, repurpose, reuse and recycling of everything, but best of all integrated learning and innovation when design and production data becomes a part of individual component records. Further advances also include security (UK Data Tag for Farm and Industrial Machinery n.d.etc) and protection against substandard repairs, replacement and installation errors, excursions beyond the operating specifications and electronic infections/malware/attack. 'In general distributed incursions demand distributed defence and action'

11  From Big to Small Data

183

Fig. 11.3  Sustainable futures are more sophisticated and complex than recycling alone

Appliances The average lifetime of consumer items (radio, TV, cookers, fridges, freezers, etc.) is ~7 years during which significant improvements are implemented in successive generations of product. These often see fewer parts, less and better materials, better performance, more built-in ‘intelligence’ and of course new time- and energy-­ saving production methods. Purchasing, usage and replacement habits vary by individual and home, region-to-region and country-to-country, but upgrades tend to occur within a repeatable window ~ 5 years. To date designers, producers, suppliers and maintainers know little or nothing about the user and what they do with their devices, and perhaps worse, recovery and recycling are often crude and a long way from sustainability. Communicative appliances within a contained Small Data IoT (Fig. 11.4) are set to change both sides of this equation (Jie 2013) and may also transform security (Palmer 2015). So, here is a seemingly insignificant example: Do you know how much coffee you drink and when? Do you know the cost and the potential benefit or, worse, risk to your health? Does the manufacturer know you, your needs and irritations with the design? Do they have any pertinent reliability and in the field performance data? The answer across the board has to be a resounding no! A coffee aficionado colleague recently had a top-end machine fail after only 1 year of operation. It was duly sent for repair but returned with a full usage and performance report. The number one shock was the fact that he and his wife had

184

P. Cochrane and A. Elmagarmid

Fig. 11.4  Distinct zones of usage, Small and Big Data applicable to both a domestic and office situation

apparently consumed over 8000 cups in 12 months. Given that they have to sleep, go out and about and travel for work and pleasure, that is an awful lot of coffee – close to one every hour on average! Even taking into account visitors, business meetings and social events at their home/office, there may be some serious and pending health issues here. Information veracity issues may also be lurking in the background, but it still raises further important questions: what has their fridge, freezer, microwave, cooker, dishwasher, health monitors and activity recorders got to contribute to this picture? With the ability to monitor comes an ability to detect likely tech failures ahead of time. Reporting back to the manufacturer and requesting maintenance action are obvious steps, but how about the suppliers of coffee beans, milk, wash powder and other sundries – will those supply chains become automated too? Hopefully, some choices will remain in our gift as we progress towards programmable homes (Shah et al. n.d.). 'Living a 21C life demands the support of 21C technologies'

11  From Big to Small Data

185

Instrumentation Every healthcare system on the planet is failing for the same reasons – the models are wrong! We are not living in the 1940s/1950s; we are in the twenty-first century, and death by a predictably small number of mechanisms has been supplanted by far longer life spans and a very broad spectrum of progressive organ failures (WHO n.d.). This often sees protracted physical and mental decline with a growing dependency incurring huge expense. So any new model has to start with individuals assuming responsibility for their health, and not the assumption that ‘doing what you like’ (in full knowledge or ignorance) is OK and the health system will come to the rescue and administer the necessary fixes at suitable points down the line. The social idealists healthcare vision of free to everyone at the point of need no matter what the cost is clearly unattainable. Countries in the First World have largely exhausted their talent pools of capable people and reached their limit to educate and train enough doctors, nurses and carers (Britland 2015). The emerging solutions appear to come in the guise of low-cost sensors, automated diagnostics and self-­ help solutions. In short, the instrumentation and laboratory capabilities of hospitals are slowly migrating to the doctors’ office and onto our bathroom cabinets. Weighing scales, glucose testers, blood pressure and oxidation testers, thermometers and blood flow and respiration devices (and more) are available at a modest cost (Elton et al. 2016). In addition, online AI diagnostics already surpass the capability of the average GP. So, we might just become a Star Trek Society for real with the handheld body scanners also under development and some already demonstrated. If we now add fully characterised and tagged food, wearable activity and vital signs monitors, along with smart toilets that measure/characterise human waste, then we have an almost complete picture. What did we consume? What activities did we participate in – how much energy and fat did we burn – and what is our body and its waste output telling us? For our AI diagnostics, human doctors and hospitals, such a day-by-day history will become a vital element in future care and treatment programmes (IBM Healthcare Industry Solutions Guide 2015). Early detection, with fast and accurate diagnosis followed by accurate treatment, will keep people out of the doctors’ office and hospital. In turn this reduces further damage and medical risk. Medical errors in the USA result in over 400,000 (digital) death per year. Diagnostic, treatment and procedural errors in drug and medicine prescription and dosage are the key culprits, and sadly, other countries and health systems fare no better! Accurate and timely data via the IoT at every level, from individual through doctors’ office and hospital, is a fundamental game changer. With every domestic and personal device communicating and contributing data in close proximity, the marshalling of data and affording it to some diagnostic engine will demand orchestration (Islam et al. 2015). Of course, this picture extends to the instrumentation during any period of hospitalisation and/or visit to the doctors’ office and totally changes the nature of patient records. We might thus anticipate the prime responsibility and ownership migrating to the individual and away from doctor and hospital.

186

P. Cochrane and A. Elmagarmid

Fig. 11.5  Health monitoring and local analysis with near-zero need for Internet connection

To illustrate this trend, consider how many health monitoring apps and test devices you own (Fig. 11.5). Most of us now have 5–10, spread across bathroom, bedroom, pocket and wrist. From FitBit and biochemistry monitors to solid and fluid waste analysers, availability is growing, prices are falling, and adoption rates are growing as they become irresistible components of our personal DIY healthcare and longevity assurance. And so to the machine support now available! IBM Watson is now superior to human doctors in diagnostics using verbal inputs based on Q and A (Friedman 2014). The US pharmacy chain CVS is deploying Watson terminals, so customers can ‘self-consult’ and get a diagnosis and a prescription in one visit! But this is only the start. Imagine such a session augmented by all lifestyle and bodily information. So the trick is going to be data aggregation and proximal delivery to a terminal at home, doctors’ surgery and hospital or Watson terminals in CVS (Jaspen 2015). The only thing we can be certain of is that the Small Data produced, along with the emergent properties, machines and the IoT, will take us by surprise, as will healthcare outcomes and revelations. 'We are creating new collaborative partnerships & environments of people and machines'

11  From Big to Small Data

187

Vehicles The automotive industry is progressively equipping vehicles with more intelligence and automated functionality (Faezipour et al. 2012). GPS, road and traffic conditions have been with us for decades along with cruise control, engine and climate monitoring, instrumentation and entertainment. Voice control and hands-free mobile phone connections are more recent along with cameras, radar, collision avoidance, automated braking, park assist, etc. This Small Data environment aligns with the coming IoT. In turn, a domination of short range (direct) car-to-car communication (Knight n.d.) appears to be evolving to create another sector unlikely to see 3, 4 and 5G mobile network domination (Fig. 11.6). We might therefore anticipate pulling in for gas and our vehicle OS and apps being automatically upgraded along with the latest maps and traffic information and, perhaps, entertainment content for the younger members of the family. There is of course an advertising opportunity with access to the main and peripheral seat back screens. The engine might wish to report performance data, along with location and journey information plus the updated driving history for the insurance company. But for the driver of the vehicle, the single biggest benefit might well be the latest road and freeway traffic data relayed directly car-to-car (daisy chain fashion) in real time giving incident data and alternate routing advisories (de Looper 2015). For many countries it would realise higher road use efficiencies with more closely packed vehicles per km that is fundamentally impossible with human drivers (Coelingh and Solyom 2012). 'Human and machine behaviours and idiosyncrasies will at one be integrated as one'

Fig. 11.6  Short-range vehicle-vehicle-garage/gas station dominates the Small Data sector

188

P. Cochrane and A. Elmagarmid

For owners and users of vehicles, the IoT and Small Data plus machine intelligence doesn’t just mean a safer and more economic mode of travel; it opens even more important doors  – two of which are the driverless car and new ownership models (Cairns and Harmer n.d.). How many of us buy oversized vehicles on the basis of occasional maximal loading and our longest anticipated journeys. The growth of leasing, renting, personal payment plans and community ownership are portents of a fastchanging models that provide the most appropriate vehicle for purpose when required. Greater road use, reduced energy costs, fewer vehicles and drastic reductions in accidents, injuries and deaths, along with shorter journey times and a better travel experience, may turn out to be but a short list of the pending benefits to be realised. Whilst it is easy to predict what the technology will and can do, the same is not true of people; they always innovate and do surprising things.

Integrated Outcomes Applying the same rationale to entertainment, work, leisure, home, office, hotel, airport, etc., it is easy to conjure a future of greater convenience where local machine and Small Data provide stewardship of our health, wealth, information and general wellbeing. This may soon become the norm. What we know for sure is today’s comms technologies cannot support >50Bn things online. We simply do not have sufficient energy, and the migration to very low power, short distance, localised networks and communications is assured. Today the mobile carriers support