Author: Peng Zhao (Founder of IoT Think Tank & Partner at Yunhe Capital)
IoT Think Tank Original
Please cite the source and origin
Introduction
As 2020 comes to a close, the industrial IoT has been climbing against the trend, crossing mountains and rivers, and making significant progress. As we extend into 2021, there are paths and forks; how do we navigate confusion and establish goals? This week, I had the privilege of consulting with Chief Engineer Yu Xiaohui, Deputy Director of the China Academy of Information and Communications Technology and Secretary General of the Industrial Internet Industry Alliance, on questions of common concern. He provided many different dimensions and perspectives, allowing me to validate my thoughts and correct my course in a timely manner.
Total word count: 4000 words, reading time: 10 minutes
IoT Queen: Short-term focus on performance, mid-term focus on business, long-term focus on format
Hello, this is my 206th article written in the [IoT Queen’s Heart] column.
In previous articles, we focused on the latest developments in 5G; this time, let’s “change the flavor” and talk about industrial IoT.
-
Is industrial IoT mature?
-
How should each enterprise measure its progress and pace?
-
What common problems and challenges do everyone face? How have others answered the same questions?
-
How can enterprises in the industrial IoT ecosystem achieve win-win benefits?
The closer we get to industrial IoT, the more we feel reverence and understand its difficulties.
Industrial Internet and industrial IoT are often mentioned together; they share a significant overlap. In a previous article titled “The ‘Industrial Internet’ Highlighted in the Central Political Bureau Meeting is Becoming the Core of the ‘Digital Infrastructure’ Era,” I also explained the differences between the two.
Here, I don’t want to get bogged down in concepts. Those of us in the industrial IoT world are quite clear about how complex and tangled each industrial IoT project can be; concepts need not be elaborated upon, as those who understand will naturally know (hearing the applause).
As 2020 comes to a close, the industrial IoT has been climbing against the trend, crossing mountains and rivers, and making significant progress.
As we extend into 2021, there are paths and forks; how do we navigate confusion and establish goals?
This week, I had the privilege of consulting with Chief Engineer Yu Xiaohui, Deputy Director of the China Academy of Information and Communications Technology and Secretary General of the Industrial Internet Industry Alliance, on the questions of common concern mentioned above.
He provided many different dimensions and perspectives, allowing me to validate my thoughts and correct my course in a timely manner.
This article shares some of my insights and feelings after communicating with Chief Engineer Yu.
01
What stage is industrial IoT in?
The development of industrial IoT can be divided into three major stages:
-
Stage 1: Building a digital foundation.
-
Stage 2: Optimizing specific links and fields.
-
Stage 3: Overall intelligent upgrading.
Stage 1 involves numerous challenges; every enterprise has likely experienced its darkest moments in this stage.
Hardly any concept starts with everything in place; this is the most arduous and foundational stage. The industrial system has developed to the point of redundancy and complexity, with fragmentation and islands. Various generations of industrial equipment, diverse communication protocols, and complex signals and systems—starting from the dirty and difficult work is the only way to enjoy the spotlight.
Stage 2, based on the digital foundation of interconnectivity, gradually realizes optimization of specific links and fields.
The local is the foundation for the whole. No matter how large the system, it must start from the simplest modules.
Quality optimization, process optimization, product optimization, operational management optimization… The initial effects of industrial IoT gradually become apparent. If local issues are addressed properly, the benefits will be tangible.
Stage 3, the overall intelligent upgrading, is the goal of industrial IoT.
If data can be exchanged and flow across the boundaries of the industrial system, it will represent a more significant and higher-level upgrade. Overall interconnectivity and optimization will bring about fundamental value.
Industrial IoT has the capability to connect and optimize real-time data from the production line, data from automation/information systems, data from upstream and downstream of the industrial chain, and data from various stages of the value chain, driving a significant increase in total factor productivity.
From the distribution of stages, according to the data statistics from the Academy of Information and Communications Technology, nearly 45% of enterprises are currently in Stage 1, over 30% have explored value in Stage 2, and a few pioneers are advancing towards Stage 3.
Understanding the three stages and the distribution of enterprises may raise more questions for you:
What if we started late? In the real industrial IoT track, there is no finish line; there are only potholes and bumps along the way. In such conditions, an F1 car may not necessarily outrun a tractor. Early movers need not be overly pleased with their leading position, as later entrants also have many opportunities.
How long until we see benefits? With persistence over time, significant progress can be observed. If expecting the entire industry to present a new look, be prepared for strategic patience of 5 to 10 years.
Can we skip steps from Stage 1 to Stage 3? Some enterprises indeed start from Stage 3: overall optimization of the industrial chain, establishing partnerships across industries, such as finance and logistics, and achieving results. Others advance two steps simultaneously, layering one stage on top of another, also with good outcomes.
Pain points and goals differ; thus, the stages and path choices of enterprises also vary. Their commonality lies in continuously experimenting and learning through action.
Because the successful practice of industrial IoT is not planned but evolved. As Mr. K, the author of the public account “Technical Leadership,” once joked, this process is like raising a fish; it unexpectedly dies, and instead of burying it, you want to cremate it. Who knew that during the cremation, it would smell so good, leading you to think of Bear Grylls and learn wilderness survival skills… Many things cannot be predicted at the outset, making planning impossible.
02
There is no single industrial IoT market
China’s industry includes over 40 major categories, with hundreds of subcategories, boasting the world’s most complete industrial system.
Different enterprises encounter significant variations in scenarios and applications while attempting industrial IoT.
Broadly, the classifications include:
Value extraction from equipment: For example, manufacturing enterprises in engineering machinery, machine tools, and gas turbines focus on equipment, using industrial IoT to track and optimize the entire product lifecycle around the equipment, establishing new value-added service models.
In industrial IoT applications, this type accounts for the largest share, close to 50%.
Control of production processes: Whether in discrete manufacturing or process industries, many enterprises are attempting to combine industrial AI with traditional processes and model mechanisms, using data to sense and analyze existing physical processes and personnel behaviors, advancing intelligence to a new paradigm.
Cross-chain collaboration: Some enterprises with factories in different regions are utilizing industrial IoT for cross-border collaboration, linking different plants and promoting collaboration across R&D, design, production, and sales, upgrading from value chains to value networks.
Continuous optimization of inventory: Small and medium-sized enterprises tend to enhance inventory efficiency through lightweight approaches, utilizing cloud platforms and existing tools.
Innovation in value-added services: Some enterprises are using data to explore and innovate service models, shifting from selling products to selling services, or connecting with consumer internet to pragmatically promote C2M (customer-to-manufacturing).
Shared and co-built capacity: Given that China has the world’s most complete industrial system and the largest industrial clusters, it has a unique advantage to develop a sharing economy for socialized production and manufacturing capabilities.
Enterprises across various industries are attempting to implement a “Didi model” in manufacturing, promoting information matching and sharing of capacity, molds, and machinery.
All types continuously remind us that industrial IoT is not a single market.
Enterprises in various vertical industries and different links within each industry have distinctly different internal and external environments, all requiring solutions that can digest and accommodate these differences.
Although we collectively refer to it as industrial IoT, each link, each scenario, and each application may correspond to a vast market.
Every enterprise possesses unique characteristics.
The greatest trump card in the hands of each enterprise is the enterprise itself.
There is no need to overestimate or underestimate; every person in the field stands on the solid ground of industrial IoT, rolling up their sleeves to get to work.
03
How can enterprises in the ecosystem achieve win-win benefits?
Industrial IoT is a field with high barriers and deep moats.
Today is an era of technological integration, where not a single technology is at work, but rather 5G, IoT, AI, cloud computing, edge computing, and big data converge in the industrial field.
Several technologies are particularly noteworthy.
First and foremost is 5G.
Currently, China has 700,000 5G base stations. Communication technology evolves approximately every ten years. No generation of technology has ever had such a profound impact on the real economy as 5G, nor has any generation of technology been so eagerly adopted while standardization is still in progress.
The reasons are twofold: on one hand, 5G has considered connectivity issues that meet industrial needs in its scenario design, such as large connections and low latency; on the other hand, 5G can form a more optimized combination with other technologies, such as the integration of 5G with edge computing and AI, which aligns very well with the future development of industrial IoT.
In terms of specific proportions, the analysis data from the Zhanhua Cup 5G Application Competition is somewhat representative; currently, industrial IoT accounts for about 30% of the overall applications of 5G.
A necessary result of promoting 5G is the integration of cloud and network, which is cloud-network integration.
Without edge and edge cloud, the characteristics of 5G, such as low latency, high bandwidth, and large capacity, would be difficult to fully utilize.
The application of 5G will inevitably be bound to edge computing, becoming a type of on-demand deployed private network service.
The second noteworthy technology is edge computing (MEC), which has become one of the essential technologies for enterprises to embark on their digital journey.
Edge computing technology is a product of ICT integration; it is not isolated from 5G and AI but is laid out simultaneously.
In 2016, ETSI (European Telecommunications Standards Institute) expanded the MEC concept to multi-access edge computing, extending edge computing capabilities from telecommunications cellular networks to other wireless access networks.
From an application perspective, the actual deployment location of the edge is very flexible; the closest edge to the user is the terminal, but not every terminal can be considered as edge.
In the past, the intelligence in industrial systems did not allocate much computing power to the “edge” level.
Although there are many sensors, variable frequency drives, and controllers on industrial sites with certain intelligent algorithms and expert rules, most remain in static local tuning, with no elastic processing capability.
Currently, the application of edge computing in the industrial field has not been standardized; edge and cloud computing complement each other, better supporting real-time intelligent processing and execution of local business.
The third noteworthy technology is industrial AI, which can be said to solve the “last mile” problem in industrial implementation of intelligence.
In the past, when many enterprises encountered bottlenecks in cost and efficiency, they relied on experience to explore. Now, through industrial AI technology, the value of massive data can be extracted, resolving many past issues.
In scenarios of quality optimization and machine vision analysis, the application of industrial AI has shown good results.
Of course, the implementation of these technologies in the industrial field is not easy.
For example, in the case of AI, many links in industry face limited data volume, irregular signal collection, insufficient negative sample sizes, and variable production conditions, all of which may lead to AI “failure.”
Due to the high difficulty level, the ecosystem of industrial IoT must be a complex, multi-party participating ecosystem.
In the past, industrial enterprises rarely built ecosystems; rather, it was the internet and telecommunications sectors that had richer experiences in ecosystem construction.
Industrial enterprises previously seldom mentioned ecosystems, not because it was unfashionable, but because it was unnecessary.
However, in the era of industrial IoT, industrial enterprises need to unite partners from different industries and roles to jointly build and maintain a brand new ecosystem.
First, benefit others, then benefit oneself.
To leverage the collective wisdom of ecosystem partners and create a system greater than the sum of its parts, it is essential to achieve healthy business models for all parties, fulfilling their respective demands and values.
From the supply and demand sides of industrial IoT, the demand side is leading in exploring business models, whether from products to services, integrating with finance, or extending to consumers, with many good practical cases. The supply side’s new business models are still in exploration, in the painful early stage of germination.
From a closed-loop logic perspective, the gradual formation of supply-side business models will also promote the iteration of demand-side business models.
—-In Closing—-
If applied correctly, industrial IoT will open up a brand new development landscape for industrial enterprises.
After my discussion with Chief Engineer Yu, if I were to distill and summarize my gains, it could be condensed into the following five understandings:
-
Understand the mission:
-
Understand the environment:
-
Understand the market:
-
Understand the technology:
-
Understand the ecosystem:
Our dialogue will be edited into a short video, to be aired at the “Advantech Industrial IoT Partner Summit” on December 5. Please scan to register.
And I salute all those who are writing the future of industrial IoT through their personal experiences.
