In this book, we will work with several open source technologies and proprietary technologies. To simplify the tests that we will carry out, we will use Docker to deploy databases and frameworks.

Over the last few years, the IoT has become a viral topic in the digital world, one that is discussed, debated, and analyzed in many channels, forums, and contexts. This is common for all new or emerging software technologies. Developers, architects, and salespeople discuss the capabilities, impacts, and market penetration of the new technology in forums, blogs, and specialized social media. Think, for example, of Docker or Kubernetes, which are changing the ways in which software applications are designed, implemented, and deployed. They are having a tremendous impact on the digital world in terms of time to market, software life cycle, capabilities, and cost. Despite this, they remain primarily confined to their specialized fields. This is not the case for the IoT.

Over the last 10 years, the IoT has become a familiar topic in the mass media. This is because the IoT is more than just a new technology that impacts a restricted range of people or a specific market. It can be better understood as a set of technologies that impacts us all, and will change markets, even creating new ones. The IoT is changing our lives, feelings, and perceptions of the physical world daily, by modifying how we interact with it. The development of the IoT is a crucial moment in the history of humanity because it is changing our mindset, culture, and the way we live. Just like the internet age, we will have a pre-IoT phase and a post-IoT phase. The IoT era will be not an instantaneous transition, but a gradual and continuous shift during which the evolution never stops.

Currently, we are just at the beginning of this journey. Like the arrival of e-commerce or mobile applications, there is a certain time lag between when you hear about an upcoming technology and when it actually exists in the real world. But the change has started. We are moving toward a world in which we interact increasingly not with physical objects, but with their digital images that live in the cloud and communicate with other digital images. These images are integrated through a powerful injection of digital intelligence, which makes them capable of suggesting actions, making decisions autonomously, or providing new and innovative services. You might currently be able to regulate your heating system remotely, but if it lived in the cloud and received information from your car, your calendar, your geolocation, and the weather, then your heating system would be able to regulate itself. When an object lives in the cloud and interacts with other digital images in a web of artificial intelligence, that object becomes a smart object.

These developments might seem to be paving the way for a new and perfect world, but there is a dark side to the IoT as well. A lot of personal data and information is stored in the cloud, in order that artificial intelligence can extrapolate information about us and profile our behaviors and preferences. From a different perspective, therefore, the cloud could also be seen as a sort of Big Brother, as in George Orwell's novel 1984. There is the possibility that our data and profiles could be used not just to enhance our lifestyles, but also for more malicious purposes, such as political or economic influence on a large scale.

An example of this was the Cambridge Analytica scandal that occurred in March 2018. It was widely reported at the time that this company had acquired and used the personal data of Facebook users from an external researcher who had told Facebook he was collecting it for academic purposes. This researcher was the founder of the third-party app thisisyourdigitallife, which was given permission by its 270,000 users to access their data back in 2015. By providing this permission, however, these users also unknowingly gave the app permission to access to the information of their friends, which resulted in the data of about 87 million users being collected, the majority of whom had not explicitly given Cambridge Analytica permission to access their data. Shortly afterward, the media aired undercover investigative videos revealing that Cambridge Analytica was involved in Donald Trump's digital presidential campaign.

This book will not go into detail about the social, legal, political, or economic impacts of the IoT, but we wanted to highlight that it does have a dark side. More often than not, human history has demonstrated that a technology is not good or bad in itself, but instead becomes good or bad depending on how it is used by humans. This is true for the IoT. Its power is tremendous, and it only just starting to be understood. Nobody yet knows how the IoT will develop from now on, but we are all responsible for trying to control its path.

The IoT as a concept wasn't officially named until 1999. One of the first examples of the IoT was a Coca-Cola machine, located at the Carnegie Mellon University (CMU) in the early 1980s. Local programmers would connect through the internet to the refrigerated appliance checking to see if there was a drink available and whether it was cold before making a trip to it.

Kevin Ashton, the Executive Director of Auto-ID Labs at MIT, was the first to describe the IoT in a presentation for Procter and Gamble. During his 1999 speech, Mr. Ashton stated as follows:

Kevin Ashton believed that radio-frequency identification (RFID) was a prerequisite for the IoT, and that if all devices were tagged, computers could manage, track, and inventory them.

In the first decade of the 21st century, several projects were developed to try to implement and translate into the real world the IoT philosophy and Ashton's innovative approach. These first attempts, however, were not so successful. One of the most famous and emblematic cases was the Walmart mandate (2003). By placing RFID tags with embedded circuits and radio antennas on pallets, cases, and even individual packages, Walmart was supposed to be able to reduce inefficiencies in its massive logistics operations and slash out-of-stock incidents, thus boosting same-store sales.

In 2003, Walmart started this pilot project to put RFID tags, carrying electronic product codes, on all pallets and cases involving all of its suppliers. In 2009, Procter and Gamble, one of the main suppliers involved in the project, stated that it would exit from the pilot project after validating and checking the benefits of RFID in merchandising and promotional displays.

The unsuccessful story of the Walmart RFID project was caused by various factors:

• Most of the technologies used were in their initial stages of development and their performance was poor. They had sensors with little information, and Wi-Fi or LAN connectivity with high power and bandwidth usage.
• The sensors and connectivity devices were expensive due to the small market size.
• There were no common standards for emerging technologies, and there was a lack of interoperability between legacy systems.
• Business cases were not very accurate.
• The technology infrastructure and architecture was organized in vertical silos with legacy hardware and middleware, and a lack of interactions between each silo.
• Technology infrastructure and software architecture was based on a client-server model that still belonged to the so-called second digital platform.

From 2008, several changes were introduced to deal with the preceding issues, which were led mainly by the mobile market. These included the following:

• New higher-performing processors were produced on a large scale at lower cost. These processors supported commercial and/or open operating systems.
• New sensors, which were much more developed, with computation capabilities and high performance embedded at a low cost.
• New network and wireless connectivity, which allowed the user to interconnect the devices with each other and to the internet by optimizing bandwidth, power consumption, latency, and range.
• Sensors and devices using commercial off-the-shell (COTS) components.
• The third, cloud-based, digital platform.

Due to these changes, the IoT evolved into a system that used multiple technologies. These included the internet, wireless communication, micro-electromechanical systems, and embedded systems such as the automation of public buildings, homes, factories, wireless sensor networks, GPS, control systems, and so on.

The IoT consists of any device with an on/off switch that is connected to the internet. If it has an on/off switch, then it can, theoretically, be part of a system. This includes almost anything you can think of, from cell phones, to building maintenance, to the jet engine of an airplane. Medical devices, such as a heart monitor implant or a bio-chip transponder in a farm animal, are also part of the IoT because they can transfer data over a network. The IoT is a large and digital network of things and devices connected through the internet. It can also be thought of as a horizontal technology stack, linking the physical world to the digital world. By creating a digital twin of the physical object in the cloud, the IoT makes the object more intelligent thanks to the interaction of the digital twin with the other digital images living in the cloud.

What has changed from the Walmart scenario to make companies support the advent of the IoT? There is no one specific new technology or invention, but rather a set of already existing technologies that have been developed. These have created an ecosystem and a technology environment which makes the connection of different things possible, efficient, and easy from a technical perspective, profitable from a market perspective, and attractive from a production-cost perspective.

The technologies that have led to the evolution of IoT ecosystems are as follows:

• New sensors that are much more mature, have more capabilities, and offer high performance at a lower cost. These smart sensors are natively designed to hide the complexity of the signal processing, and they interact easily through a digital interface. The smart sensor is a system itself, with a dedicated chip for signal processing. The hardware for signal processing is embedded in each sensor and miniaturized to the point that it is part of the sensor package. Smart sensors are defined by the IEEE 1451 standard as sensors with a small memory and standardized physical connections to enable communication with the processor and the data network. As well as this, smart sensors are the combination of a normal sensor with signal conditioning, embedded algorithms, and a digital interface. The principal catalyst for the growth of smart-sensing technology has been the development of microelectronics at reduced cost. Many silicon manufacturing techniques are now being used to make not only sensor elements, but also multilayered sensors and sensor arrays that are able to provide internal compensation and increase reliability. The global smart sensor market was evaluated at between $22 and $25.96 billion in 2017. It is forecast to reach between $60 and $80 billion by the end of 2022.
• New networks and wireless connectivity, such as personal area networks (PANs) or low power networks (LPNs), interconnect sensors and devices in order to optimize their bandwidth, power consumption, latency, and range. In PANs, a number of small devices connect directly or through a main device to a LAN, which has access to the internet. Low-Power Wide-Area Networks (LPWANs) are wireless networks designed to allow long-range communications at a low bit rate among battery-operated devices. Their low power, low bit rate, and their intended use distinguish these types of network from the already existing wireless WAN, which is designed to connect users and businesses and carry more data, using more power. (More information can be found on WANs at https://en.wikipedia.org/wiki/Wireless_WAN.)
• New processors and microprocessors coming from the world of mobile devices. These are very powerful and very cheap. They have produced a new generation of sensors and devices based on standardized and cheap hardware that is driven by open and generic operating systems. These use common software frameworks as an interface, allowing you to transition from a legacy solution, with strictly coupled hardware and software, to a platform built on the COTS component and the adoption of an open software framework.
• The battle of the real-time operating system (RTOS) to gain a larger slice of new markets between the big market players. This places more sophisticated and powerful integrated development platforms at the maker's disposal.
• Virtualization technology, which divides naturally into the data center, big data and the cloud. This leads to the following features:
  • CPUs, memory, storage, infrastructures, platforms, and software frameworks available as services on demand, with flexible and tailored sizing. These are cheap and available without capital expenditure (CAPEX) investment.
  • Elastic repositories for storing and analyzing the onslaught of data.
  • The profitable and flexible operational expenditure (OPEX) model per CPU, memory, storage, and IT maintenance services. This creates a business case for migrating the legacy data, infrastructure, and applications to the cloud, and making the collection of big data and subsequent analytics possible.
• The convergence of IT and operational technology (OT). This has led to the increasing adoption of COTS components in sectors in which hardware was traditionally developed with specific requirements such as is the case in industrial plants.
• The diffusion of mobile devices and social networks has created a culture and a generic mindset with an embedded expectation for the market consumers to encounter the world through an app which shares related information.

The preceding factors are making it possible to transition from a vertical, legacy platform with an application organized hierarchically, with the data confined in silos, to a horizontal, modular, and cloud-based platform. This new platform uses a standardized API layer that provides high interoperability capabilities and the ability to share data and information between applications.

Let's consider what might happen if the Walmart project was carried out now. In 2003, the only RFID technology that existed was active RFID systems. Active RFID systems use battery-powered RFID tags that continuously broadcast their own signals. They provide a long-read range, but they are also expensive and consume a lot of power. Passive RFID systems, on the other hand, use tags with no internal power source, and are instead powered by the electromagnetic energy transmitted from an RFID reader. They have a shorter read range, but they are unembeddable, printable, and much cheaper, which makes them a better choice for many industries. Also, at the time of the Walmart project, there were no PANs or LPNs to capture and transmit the label data, meaning the developers had to adopt an expensive, wired connection to transfer the information. The data was then stored in a legacy database and processed by a custom application. If the Walmart project were to be carried out now, instead of in 2003, the tracking information could be carried out by passive RFIDs. The data could be captured by a PAN and transmitted via the cloud to be processed by an application built on top of a common API and framework. This means that all data and information could be easily shared between the project partners. According to Forbes and Gartner, the IoT market and connected devices is expected to grow strongly in the next year, as shown by the following statistics:

• The IoT market will grow from$900 million in 2015 to $3.7 billion in 2020 (McKInsey). Source: https://www.forbes.com/sites/louiscolumbus/2016/11/27/roundup-of-internet-of-things-forecasts-and-market-estimates-2016/#10e08343292d.
• There will be 21 billion IoT devices by 2021 (Gartner). Source: https://www.informationweek.com/mobile/mobile-devices/gartner-21-billion-iot-devices-to-invade-by-2020/d/d-id/1323081.

As previously discussed, the IoT is not just a specific technological innovation, but a radical change that will impact the whole of human society. This means that the IoT will affect nearly every aspect of our personal and professional lives and any sector of the economy, including the following:

• Industrial and manufacturing
• Supply chain
• Retail
• Financial and marketing
• Healthcare
• Transportation and logistics
• Agricultural and environmental
• Energy
• Smart cities
• Smart Homes and Buildings
• Government and military
• Security forces
• Education
• Sports and fitness

All of these are already involved in the digital transformation that has been caused by the IoT, and are likely to play a greater role in this in the future.

Across all uses of the IoT, the common feature is the smart object. From a qualitative perspective, a smart object is a multidisciplinary object which includes the following elements:

• The physical product.
• Sensors, microprocessors, data storage, controls, and software, managed by an embedded operating system.
• Wired or wireless connectivity, including interfaces and protocols. This is used to connect the product to its user, all instances of the product to its vendor, or the product to other types of products and external data sources.

In another definition, the article How Smart, Connected Products Are Transforming Competition, written by Michael E. Porter and James E. Heppelmann, details four increasing levels that classify the smartness of an object or product:

• Monitoring: Monitoring of product conditions, external operation, and usage. This enables alerts and notifications of changes.
• Control: Software embedded in the product or in the cloud enables control of product functions, and/or personalization of the user experience.
• Optimization: The previous capabilities are used to create algorithms to optimize the product's operation and use. They enhance product performance, and/or allow features such as predictive diagnostics, service, repair, and so on.
• Autonomy: The combination of the previous capabilities produces a product with autonomous control for self-coordination with other systems and/or self-diagnosis or services.

The IoT is sometimes used as a synonym of big data, is sometimes confused with the cloud, and is sometimes linked to machine learning and artificial intelligence. All of these things are partially true:

• IoT uses big data technology to store data
• IoT is normally deployed on the cloud to improve scalability
• IoT uses advanced analytics to process data

However, on the flip side, there's this to consider:

• IoT is focused on a data stream, rather than having huge amounts (petabytes) of data storage
• IoT can use on-premises solutions through virtualization technology
• Machine learning on IoT is not as productive as simple threshold rules or physics-based analytics

These concepts have been highlighted by the The Eclipse Foundation's 2018 IoT survey. The following diagram shows the adoption of the cloud technologies by companies:

The following shows IoT technology adoption from a storage point of view:

In this book, we will explore the most common IoT cloud solutions, such as AWS, GCP, and Azure, and the most common OEM I-IoT platforms, such as Bosch IoT, Predix, and MindSphere, to provide state-of-the-art IoT technology. We will also look at other common open source technologies, including those in the following table:

These technologies can be used to build an IoT platform from scratch or to integrate with an existing one. We will also consider other commonly used commercial software in the industrial environment.

We will discover the new generation of edge computing and the edge gateway, and, finally, we will deal with machine learning and artificial intelligence. This journey is also the journey of the IoT from the cloud to the big revolution expected around 2020:

After the advent of the steam engine in 1760, steam was used to power everything from agriculture to textile manufacturing. This caused the First Industrial Revolution and the age of mechanical production. At the end of the 19th century came the arrival of electricity, new modes of labor organization, and mass production, which started the Second Industrial Revolution. In the second half of the 20th century, the development of semiconductors and the introduction of electronic controllers produced the beginning of the automation era and the Third Industrial Revolution. In the Hannover exhibition of 2011, Henning Kagermann, Wolf-Dieter Lukas, and Wolfgang Wahlster coined the term Industry 4.0 for the project of renewing the German manufacturing system using the capabilities of the latest digital technology:

Industry 4.0 is expected to be able to do the following:

• Connect or merge production with information and communication technology
• Merge customer data with machine data
• Harness the capability of machines communicating with machines
• Manage production autonomously in a flexible, efficient, and resource-saving manner

The IoT is almost, by definition, the key for further development of the manufacturing industry by including technologies such as big data analytics, the cloud, robotics, and most importantly, the integration and convergence between IT and OT.

Generally speaking, the term I-IoT refers to the industrial subset of the IoT. The I-IoT, like the IoT, is not just a specific new technology, but instead refers to the whole chain of value of a product. Similarly, the I-IoT impacts all sectors of the industrial world by significantly modifying the processes at each stage of the life cycle of a product, including how it is designed, made, delivered, sold, and maintained. Like the IoT, we are just at the beginning of the I-IoT journey.

The I-IoT is expected to generate so much business value and have such a deep impact on human society that it is leading the Fourth Industrial Revolution.

This is according to Forbes:

• The global IoT market will grow from $157 billion in 2016 to $457 billion by 2020, attaining a compound annual growth rate (CAGR) of 28.5%
• Discrete manufacturing, transportation and logistics, and utilities will lead all industries in IoT spending by 2020, averaging $40 billion each

In few other industries are there so many opportunities to use the I-IoT than in manufacturing. In this field, it can be used to connect the physical and the digital, and build assets such as machines or production and non-production objects. It can also be used to create a variety of product and manufacturing process parameters as part of a vast information network. With manufacturing, we typically tend to think about goods and products, but the bigger opportunity for manufacturers lies in cyber-physical systems (https://www.i-scoop.eu/industry-4-0/#The_building_blocks_of_Industry_40_cyber-physical_systems), a service economy model, and the opportunities that are presented through exploring data. It is estimated that in the future, successful companies will be able to increase their revenue through the I-IoT by creating new business models, improving their productivity, exploiting analytics for innovation, and transforming the workforce.

The following is a list of several I-IoT use cases in manufacturing and their benefits:

• Manufacturing operations: This includes all operations typically performed by the manufacturing execution system (MES) that can take advantage of end-to-end visibility, such as planning, production optimization, and supplier management.
• Asset management: This includes production-asset monitoring, and tracking and monitoring parameters areas, such as quality, performance, potential damage or breakdowns, bottlenecks, and so on.
• Field service organizations: These are an important driver of growth, and, obviously, of margin. It's clear that having a hyper-connected, hyper-aware, digitized and IoT-enabled manufacturing ecosystem marks a company out.
• Remote monitoring and operation: This optimizes flow, eliminates waste, and avoids unnecessary work in the process inventory to save energy and costs.
• Condition-based maintenance: This is important to optimize machine availability, minimize interruption, and increase throughput.
• Big data: Big data can be used to monitor the quality and the makeup of services and enhance the outcome of this aggregated data.
  

Ultimately, all of these use cases highlights that data plays a key role. In the next few chapters, we will see how the data that comes from sensors and other industrial equipment is gathered and how big that data can be. Manufacturers who use this data can bridge the gaps between the planning, the design, the supply chain, and the customer of a particular product. In addition, thanks to this strong integration, shared data and information islands of automation can be easily linked together.

There are many similarities between the IoT and the I-IoT. The I-IoT, however, is strictly related to industry and so it has some specific features, as highlighted in the following list:

• Cyber security is a critical topic for any digital solution, but its implementation in the industrial world requires special attention. This is because the OT systems and devices in industry have a much longer life cycle and are often based on legacy chips, processors, and operating systems that are not designed to be connected over the internet. This means they live in an isolated LAN, protected by a firewall from the external world.
• It is critical to ensure that industrial digital devices stay running; any temporary disruption can imply a large economic loss.
• I-IoT solutions must co-exist in an environment with a significant amount of legacy operation technologies. They must also co-exist with different devices acting as data sources, including SCADA, PLCs, DCS, various protocols and datasets, and the back-office enterprise resource planning (ERP) systems as well.
• Industrial networks are specialized and deterministic networks, supporting tens of thousands of controllers, robots, and machinery. I-IoT solutions deployed into these networks must, therefore, scale tens of thousands of sensors, devices, and controllers seamlessly.
• Physical objects in the industrial world are more complex and have a wider range of typologies when compared to the consumer world.
• In the industrial world, robustness, resilience, and availability are key factors. Usability and user experience, however, are not as relevant as they are in the consumer world.
• Industrial and OT systems, from programmable logic controllers to machining equipment, are frequently reprogrammed and reconfigured to support new processes. I-IoT solutions must support and provide the same flexibility and adaptability to support operations.
• Intellectual property is a sensitive and important topic in the industrial world. Consider, for example, the design of a new machine, an engine, or a new food or drink recipe. The IP is often what differentiates a company in the market, and this cannot be lost or violated, since it is often managed by the company as a trade secret rather than covered through a patent.

With 50 billion industrial IoT devices expected to be deployed by 2020, the volume of data generated is likely to reach 600 zettabytes per year. A single jet engine produces about a terabyte of data in five hours. Given these assumptions, we need a fast and efficient way to analyze data through data analytics. In the last five years, big data technologies have been improved to scale computational capabilities. Big data analytics is about collecting and analyzing large datasets in order to discover value and hidden data, and gain valuable information. The applications of these analytics are as follows:

• Diagnostic: Understanding the cause of a fault or issue
• Maintenance: Predicting and adjusting maintenance intervals to optimize scheduling
• Efficiency: Improving the performance of the production or the utilization of resources
• Prognostic: Providing insight to avoid faults or to maintain efficiency
• Optimization: Optimizing resource consumption or compliance with local government regulation
• Logistic and supply chain: Monitoring and optimizing delivery

In the IoT, from the technical point of view, we can identify two broad categories of analytics:

• Physics-based: Based on mathematical formulas or knowledge expertise
• Data-driven: The model is built using past data

Physics-based and data-driven analytics can be combined to build a reliable hybrid model.

Recently, the introduction of deep learning (a branch of machine learning) in the contexts of image and audio processing has brought a lot of attention to data-driven technologies.

We are now aiming to expand the application of deep learning to the I-IoT to improve speed and accuracy in data analysis. In addition to audio and image data, IoT data can be processed with deep learning based on learning, inference, and actions.

However, there are two drawbacks:

• The abundance of false positives that are produced by these techniques
• The fact that companies do not always understand the outcomes of these techniques

Resolving both of these issues will ensure that an abundance of caution is built into machine learning models used in industrial applications. We need to not only create better algorithms, but also make sure that people with domain expertise understand machine learning suggestions. We also need to build systems that take in feedback, and are aware of the end user and the effects of a good or bad response.

From an infrastructure point of view, we need to shift from on-premises to cloud computing, and to provide a platform for data analytics in the cloud. This is known as Data as a Service (DaaS).

The industrial world is a very large category. It includes manufacturing, but also many other sectors, such as power and energy, renewable energies, health care, and so on. Inside manufacturing itself, there are a large variety of sectors, including the automotive industry, chemicals, food and drink, and pharmaceuticals. Production is also only one phase of the product life cycle. Besides this, we have design, provision, delivery (with its own supply chain), and the aftermarket phase.

In this book, we will focus on the manufacturing environment by considering factory processes and strictly tailoring our analysis on the data. We will look at how data is produced, stored, processed, enriched, and exchanged between different OT systems inside industrial plants, and also at how it can be gathered, transferred, stored, and processed in the cloud. We will consider a scenario in which we have a specialized device, the edge device, which is responsible for collecting the data from the OT systems of the factory and transferring it to the cloud on a very large scale. We will also cover scenarios in which each edge device gathers and manages thousands of signals coming from sensors with a sampling rate starting from 1 Hz. The analysis and the proposed solutions of these scenarios are also applicable to less complex cases in which there is no factory and/or with fewer signals to manage. For example, consider a wind turbine, where you need to monitor a piece of industrial equipment.

In this book, we will not cover scenarios in which there are too few signals to be collected for each data source to justify the need of an edge device.

In this chapter, we have analyzed the origin of the IoT and looked at how it came about through a combined set of technologies. We then learned about the key technologies that underlie the IoT, by going into its use cases and business models. We defined the IoT as a technological layer that creates a digital twin of a physical object in the cloud, making the object more intelligent due to the interaction of its digital twin with other digital images living in the cloud. We also identified four levels to define the smartness of a product or object.

We then looked at how the IoT can be applied to the industrial world, thereby beginning the Fourth Industrial Revolution and Industry 4.0. We looked at the key transformation elements that mark out the I-IoT. We also highlighted some of the main use cases of the I-IoT and the main differences between the IoT and the I-IoT. We then listed and understood the different types of analytics that apply to industrial data. Finally, we clarified and defined the industrial scenarios that will be covered in the rest of the book.

In the following chapter, we are going to look at how a factory is structured and organized from an OT perspective. We will consider who produces, processes, and enriches the data. We will also explore some key concepts, including deterministic, real-time, closed loop, sensor, fieldbus, PLCs, CNC, RTU, SCADA, HISTORIANS, MES, and ERP.

1. Why did the Walmart mandate fail?
2. What are the main enabling factors for the IoT?
3. Which are the main technologies underlying the IoT?
4. What is a smart object?
5. What is the main scope of the I-IoT?

6. What are the main differences between the IoT and the I-IoT?
7. What are the two main categories of analytics in the I-IoT?

Read the following articles for more information:

• Cambridge Analytica data scandal: https://www.bbc.com/news/topics/c81zyn0888lt/facebook-cambridge-analytica-data-scandal; https://www.nytimes.com/2018/04/04/us/politics/cambridge-analytica-scandal-fallout.html
• RFID info: http://www.rfidjournal.com/get-started
• The third platform – what it is, how we got there and why it matters: https://www.i-scoop.eu/digital-transformation/the-third-platform/
• How Smart, Connected Products Are Transforming Competition: http://www.gospi.fr/IMG/pdf/porter-2014-hbr_how-smart-connected-products-are-transforming-competitionhbr-2014.pdf
• Digital Transformation Monitor; Germany: Industrie 4.0: https://ec.europa.eu/growth/tools-databases/dem/monitor/sites/default/files/DTM_Industrie%204.0.pdf
• Acatech National Academy of Science and Engineering: Securing the future of German manufacturing industry: https://www.acatech.de/Publikation/securing-the-future-of-german-manufacturing-industry-recommendations-for-implementing-the-strategic-initiative-industrie-4-0/
• Competitive Advantage (1985): Michael E. Porter. Competitive Advantage: Creating and Sustaining Superior Performance, Free Press, 1985; 1998