Three-T integration
The development of the Industrial Internet is interwoven with three main threads: IT (Information Technology), OT (Operational Technology), and CT (Communication Technology). One thread is the development of equipment, automation , and industrial protocols represented by OT, including PLCs, Ethernet, and sensors. Another thread is represented by IT, encompassing software, the internet, and extending to cloud computing, big data, and artificial intelligence. The third thread is communication technology and chip technology, represented by CT. This represents a 3T convergence for CIOs.
The birth of the Programmable Logic Controller (PLC) was a pivotal moment in industrial development. The world's first PLC, the Modicon 084, was introduced in 1969, ushering in the PLC era for industrial control. This event was so profound that it is considered the beginning of the Third Industrial Revolution in the evolution of Germany's Industry 4.0. The PLC is a perfect combination of software programming, chip technology, and automation technology. In the world of automation, no single invention has had such an impact on manufacturing.
It has completely changed the course of mechanical and electrical automation .
In terms of connecting and controlling things, PLC can be seen as a revolutionary origin. Upon closer examination, this can also be seen as the beginning of the Internet of Things (IoT). PLC can be considered the earliest and most successful example of the integration of 3T (computers, communications, and communications).
This was followed by a series of intertwined developments in the fields of automation , IT, and CT, with ICT being the first to achieve convergence and become a unified term. With GE's proposal of the Industrial Internet , 2012 can be considered the inaugural year of the Industrial Internet. The legend of 3T convergence will be rewritten once again. It can be said that the Industrial Internet is inherently an intertwined result. The development of any one link missing would be unimaginable.
Image | CIO3T Convergence Enables Industrial Internet (Self-drawn)
The events listed in the diagram above include both technological breakthroughs and inspiring ideas and actions by companies that impact the industry (such as mergers and acquisitions). This also illustrates that the Industrial Internet cannot be understood solely from a technological perspective; conversely, fully realizing the value of the Industrial Internet cannot be achieved simply by upgrading technology.
The convergence of CT, IT, and OT (3T) has finally given rise to a new era: the Industrial Internet . Whether it's helping companies optimize operations, track and analyze equipment, perform predictive maintenance, or improve real-time decision-making, the interconnection of massive amounts of equipment has gained real meaning, industrial data has begun to shine, and the Industrial Internet has finally become a reality.
A revolution that goes all the way
2017 was a year full of innovation for the traditional automation market. The emergence of digital technology has rendered electrification and automation obsolete, and automation is now closely embracing software, moving towards a digital world. While the convergence of IT and OT has garnered significant attention, a landmark event occurred: in 2017, for the first time in automation history, the market share of Industrial Ethernet surpassed that of Fieldbus, becoming the most important factory-level communication system.
Even faster growth occurred as early as 2016. According to HMS's Industrial Networking Market Share Report released in February 2017, Industrial Ethernet grew faster than in previous years, with a growth rate of 22%. Industrial Ethernet now accounts for 46% of the global market, compared to 38% last year. In specific communications, EtherNet/IP and PROFINET have the largest share. Meanwhile, fieldbus saw a significant decrease, from 58% to 48%.
Figure | The two significantly narrowed the gap in 2017.
Data from 2018 shows that Industrial Ethernet surpassed Fieldbus for the first time, accounting for 52% of newly installed nodes (compared to 46% last year), while Fieldbus accounted for 42% (compared to 48% last year). Industrial Wireless, however, saw only modest progress.
Figure | Changes in Industrial Communications, 2015-2017
(Data source: HMS Annual Report)
For years, the automation market has been talking about a "revolution of seamless connectivity," which involves directly connecting high-end, mature commercial Ethernet systems to the underlying layers of factory equipment. With Industrial Ethernet becoming the most important industrial communication system, this long-awaited revolution has finally achieved a breakthrough in the era of the Industrial Internet .
Human Internet, Internet of Things, and Non-Internet of Things
In terms of the boundaries of the Internet of Things (IoT), it can be roughly divided into three types: the Internet of People, the Internet of Things, and the Internet of Things that are not connected to the Internet. The Internet of People, also known as the consumer internet, includes various devices connected to the internet, from PCs to mobile phones.
The Internet of Things (IoT) consists of various ubiquitous smart terminals with sensing and communication capabilities, but these terminals do not directly interact with people. They can sense certain parameters and then process or analyze them. In the industrial field, this process started with RFID attempting to establish connections between objects, then moved to machine-to-machine (M2M) communication, and finally to the IoT. Various relationships between objects, from identification and communication to interaction methods, are constantly being explored.
Image | The Boundaries of the Early Internet of Things
There are also a large number of "dark devices," which are connection devices without IP addresses (such as RFID and barcodes). They have existed for a long time and are not originally part of the Internet of Things (IoT). From the perspective of the boundaries of the IoT, they can be called "Not Internet of Things (NIoT)."
Significantly, replacing or activating these devices, bringing these "non-Internet of Things" into a brighter world, has become an important task of the Internet of Things.
Industrial Internet Disruption
With GE's Industrial Internet and the rapid rise of Industry 4.0, smart manufacturing has sparked widespread discussion and detailed application scenarios. Coupled with the rapid development of digitalization, IoT, and ICT technologies, the IoT camp has begun to differentiate, leading to a rapid evolution of IoT in the industrial sector from a practical perspective.
Firstly, to facilitate the distinction between consumer-facing (2C) and industrial-facing (2B) manufacturing, the categories of Consumer Internet of Things (CIoT) and Industrial Internet of Things (IIoT) emerged. Companies like Haier Appliances and Sany Heavy Industry's RootCloud are respectively geared towards people and construction machinery.
In other words, the Internet of Things (IoT) can be divided into two parts: the Industrial Internet of Things (IIoT) and the Consumer Internet of Things (CIoT).
It is worth noting that, in the Chinese context, it would be a stretch to include areas such as smart retail, healthcare, smart buildings, smart cities, and smart agriculture under the umbrella of "industrial Internet of Things" (IoT). They can be considered as other IoT fields outside of consumer IoT and industrial IoT (there is an impulse in China to call these "industrial internet," but this term has not seemed to gain mainstream acceptance).
Secondly, devices that were not originally connected to the Internet of Things were widely activated through various protocols, gateways, and data acquisition systems, unexpectedly creating a vibrant and promising landscape. Numerous dormant devices, terminals, and appliances—"orphan products"—were activated and entered network mode. Many terminals without web addresses also became part of the data resources through other means.
More importantly, with the rapid rise of smartphones, people, as highly profiled users, have presented enormous commercial value.
Both public opinion analysis and behavioral description have become the focus of consumer IoT and industrial IoT.
At this point, the Industrial Internet, as a holistic resource, has re-emerged, encompassing human elements, consumer products, connected devices, and non-connected devices—this is the meaning the Chinese have given to the "Industrial Internet." In the current world of the Industrial Internet, there are parts of the human-centered network, the Industrial Internet of Things (IIoT), and simple device connections.
Image | The Yellow Map of the Industrial Internet (Source: Self-drawn)
The terms "Industrial Internet" and "Industrial Internet of Things" are often used interchangeably. However, they are actually different. For example, the Industrial Internet of Things cannot use the familiar HTTP protocol. HTTP, a cornerstone protocol of the internet, is simply too slow for industrial-grade requirements. In other words, HTTP is suitable for the Industrial Internet, but not so much for the Industrial Internet of Things. Therefore, the boundary between the Industrial Internet of Things and the Internet of Humans (IIoT) is marked by numerous instances of HTTP usage, which can only be explained by the specific characteristics of the Industrial Internet.
To simplify, the Industrial Internet can be simply described as a component of the Human Internet, the Consumer Internet of Things (IoT), the Industrial Internet of Things (IIoT), and a large number of IP-free connected devices. What does this "industry" encompass? It can include "heavy" industries such as machinery manufacturing, oil and gas, and transportation; it can also include lighter applications such as smart cities and agriculture. This is because the latter also requires many industrial-grade applications.
I also need to briefly discuss its relationship with Industry 4.0. All IoT-related applications in Industry 4.0 are related to the Industrial Internet; however, not all Industrial Internet applications can be categorized under Industry 4.0.
Typical applications of the Industrial Internet are not all on machines, but include lighting, intelligent transportation, intelligent machine applications, factory control, factory applications, condition monitoring, and other applications on agricultural and power equipment.
Inspiration from boundaries
The distinction between these concepts is not always clear-cut. For example, many Internet of Things (IoT) devices are already integrated with 3C electronic products, such as wearable devices, mobile phones, and smart homes. In fact, smart cars, smart homes, and smart health devices may belong to the consumer IoT and the industrial IoT respectively, but both are closely related to the human IoT.
However, the intersection of boundaries is a very meaningful source of inspiration. Especially for corporate strategy, it's crucial to identify and find the most suitable boundaries. Many important things often happen at the intersection of fields. This is why suppliers from different sectors frequently engage in various cross-industry collaborations. ICT vendors, IT vendors, and automation vendors—previously completely different fields—are now discussing digital manufacturing under the same roof, something never before seen in industrial history.
We are still in an era where manufacturers are carefully defining their boundaries. Companies need to clearly define their strategic positions. In this regard, Haier's path selection seems rather contradictory. CosmoPlat, which has been advocating for large-scale personalized customization, seems to have chosen the wrong side. It appears to be building a consumer IoT, but it is actively linking itself to production, closely binding itself to its "connected factories," and pursuing the path of industrial IoT. This is currently the only attempt we see to actively connect consumer IoT and industrial IoT directly. Bosch, Europe's largest home appliance group, is quite different from its attempts at industrial IoT in its consumer IoT endeavors. Similarly, automotive giants are carefully differentiating their attempts at connected vehicles from the industrial IoT of their factories.
Industrial data is a dynamic and inquisitive youth.
For years, vast amounts of data from equipment have been neglected. They were simply left lying around in the workshop or allowed to dissipate freely near engines. The development of the Industrial Internet has brought renewed attention to the value of data. According to Gartner's 2016 forecast, in the following years, 40% of data would come from sensors: mobile phones, connected vehicles, home appliances, as well as machines and large equipment such as power grids, aircraft, and oil and gas equipment. This led to industrial big data analytics being considered a powerful tool for solving industrial problems.
However, industrial data is a problematic and often chaotic subject. This manifests in six main symptoms: the data is dirty (requiring extensive algorithmic cleaning to produce usable data), the frequency varies greatly (the frequency of triggering events varies significantly), the volume is massive, the size is inconsistent (data volumes vary), the types are numerous (various heterogeneous data sources), and the relationships are complex due to interdisciplinary connections (data mechanisms originate from different disciplines such as mechanics, optics, optics, thermodynamics, and magnetism).
Image | Industrial data is a problem youth
From this perspective, what value does industrial data storage hold? It's like an unfathomable underwater oil field. There might be treasures inside, but you definitely wouldn't want to go in and explore.
Industrial software is the key.
Since only a portion of the data is useful, and the majority is just unscrupulous noise, this raises a crucial question: how can we extract industrial data hidden deep beneath the seabed and turn it into oil?
The answer is that industry knowledge is what makes a professional drilling team.
This brings us back to GE's original concept of the Industrial Internet, which strongly emphasizes "advanced analytics." Advanced analytics is not simply data analysis; it's based on domain knowledge and integrates chip computing power, industrial software, automation, and material properties.
Diagram | The Essence of the Industrial Internet (Modified from GE Diagram)
In essence, the most important aspect of the Industrial Internet is the integration of "machines, computers, and people" to perform data analysis and thereby transform business output. Without knowledge, there is no data. Without a solid industrial foundation, industrial big data and artificial intelligence are merely theoretical concepts and castles in the air.
However, the knowledge-based transformation of industrial technology (or the "software-based transformation of industrial technology") is much more difficult. This is a goal that knowledge engineering and knowledge management have been striving for for decades, but the results seem to be very poor. The most difficult aspect for manufacturing companies is the transformation of tacit knowledge. Even in the PC era, this problem was not well solved—in fact, it was almost never successfully addressed. Now, in the mobile internet era, the dilemma of "knowledge-based transformation" remains difficult to overcome.
There seem to be more reasons to be optimistic now, as BAT giants and numerous small and medium-sized startups are entering this situation with the nuclear-level power of artificial intelligence.
However, industrial technology will not lose its inherent industrial complexity due to the influx of algorithms. The expression of industrial technology still follows its own inherent rules. The accumulated industrial technologies of companies like GE and Siemens remain a formidable barrier. Even on so-called open-source platforms, there are still vast knowledge "black boxes" that are difficult to explore. The disruption caused by BAT (Baidu, Alibaba, Tencent) has only one result: poaching a large number of IT talents already scarce in the manufacturing industry to work on easier modeling and algorithm development. For industrial software, this is more of a devastating blow.
Deeply mining complex industrial data is a nut to crack for Chinese industry. The roles of both the client and the contractor (developer) in Chinese manufacturing must be more strategically aligned and deeply collaborative. The combination of data analysis methods and industrial mechanistic knowledge is the result of a collaborative effort between the two parties: the client needs to become the "client among contractors," undertaking "knowledge-based poverty alleviation" work. This is a tremendous challenge.
Integration of Industrial Internet Platforms
The Industrial Internet is the most important component of the Internet of Things (IoT), with many application scenarios far exceeding consumer applications. According to IDC, nearly 60% of the IoT budget in 2016, reaching $102 billion, was spent in manufacturing, twice the budget of the second-ranked transportation industry. These expenditures primarily focused on three areas: production operations, factory asset management and maintenance, and equipment services. All of these costs aim to ultimately transform data into actionable intelligence.
This promising future of the Industrial Internet, however, is a long and arduous journey. It requires resolving numerous different device interfaces and handling diverse software environments to ensure compatibility with various IoT devices and software applications. The Industrial Internet platform is precisely such an integrated platform.
According to the white paper from the China Academy of Information and Communications Technology (CAICT), the Industrial Internet will be divided into four layers: edge computing, infrastructure as a service (IaaS), platform as a service (PaaS), and application as a service (SaaS). From a more specific perspective, the Industrial Internet can be divided into five levels: device-side, connectivity, software middleware, applications, and scenario-based services.
Figure | Five Levels of the Industrial Internet
(Source: machnation, 2016)
Because each level has its own solutions and numerous scenarios, the term "Industrial Internet Platform" can appear ambiguous. However, this doesn't deter ambitious attempts by enterprises, as the Industrial Internet Platform represents a massive enterprise-level market. Above the platform, numerous scenario-oriented application developments are underway. According to IDC data from November 2017, by the end of 2020, 50% of IoT applications will be enterprise-level. These applications, integrated on the Industrial Internet Platform, leverage complex analytics capabilities to provide greater value to industry.
With such a promising future, it's no wonder that hundreds of suppliers, large and small, have gathered here. It's quite a crowded place.
However, as mentioned earlier, the Industrial Internet platform remains a very broad concept. If we break it down further, at least two platforms stand out. One is the Application-Enabled Platform (AEP), which enables industrial internet applications. It unifies different software functional modules onto a single platform, easily handling compilation, packaging, and distribution. Platforms like PTC's ThingWorx and Tianjin Yike's industrial app development tools serve this role. The other is the Device Management Platform (DMP), a traditional battleground where automation vendors used various protocols to create barriers and isolate each other. However, equipment manufacturers and owners are now taking action themselves, connecting different devices. Companies like Rootcloud and Shijiazhuang Tianyuan are following this approach. Further subdivision reveals the Edge Platform, where ICT vendors are currently focusing their efforts, aiming to find breakthroughs in storage and computing at the connectivity end, such as Huawei's OceanConnect.
A close look at the current construction of industrial internet platforms reveals that many of the initial stages are deeply rooted in these areas.
According to predictions from MachNation in the US, the Industrial Internet platform will be a double-digit growth area until 2025. The author's comprehensive statistical analysis estimates the Industrial Internet platform market at approximately $1 billion, representing 1% of the entire Industrial Internet sector. While the market is small, it possesses immense appeal because the value of data is far from being fully realized. All the competition is for the next stage of deployment, and better, faster, and cheaper deployment and development will be the fundamental test for Industrial Internet platforms. Many large manufacturing enterprises have also begun to shift their overall focus to Industrial Internet platform strategies. Japan has also found the optimal time to integrate OT, IT, and CT. For example, in September 2017, Hitachi announced the establishment of Hitachi Vantara as a new business entity to provide data-driven solutions for industrial and commercial enterprises. This new company packaged Hitachi Data Systems data centers, Hitachi Insight Group IoT, and Pentaho business intelligence businesses into a single integrated business, Hitachi Vantara, leveraging Hitachi's capabilities in operational technology (OT) and information technology (IT). Lumada version 2.0 (Hitachi's IoT platform, a standalone commercial software product) was also released at the same time, featuring a comprehensive update with enhanced artificial intelligence (AI), machine learning, and advanced analytics capabilities.
Against this backdrop, it's not surprising to see Foxconn's lightning IPO in February this year and the launch of the BEACON platform.
Notes
The greatest value of the Industrial Internet lies in its comprehensive connection of data and physical entities, and based on this connection, it enables computation and control, making real-time decision-making possible. This was unimaginable ten years ago.
The concept of IoT is about to celebrate its 20th anniversary, and the current enthusiastic embrace suggests that its journey has only just begun. For businesses, IoT can no longer be viewed merely as a technology, but as a key element of their business activities—or rather, a strategic consideration.
However, to immediately see the immense value of the Industrial Internet, many significant investment pitfalls still need to be addressed. For many SMEs, the disproportionate ROI remains a challenge. Currently, the Industrial Internet lacks unified standards and perspectives; each company often speaks its own mind, presenting its own narrative. The market remains murky and far from clear. Moreover, it seems more like a market prepared for giants. While the big players boast and strategize, smaller players find a few opportunities amidst the chaos.
