The industry is changing, digitalizing, let’s call it the industry of the future or industry 4.0, it is increasingly integrating data into decision-making. If data has been driving the industry for a long time, we are now talking about another scale, that of massive data, on which it is possible to make faster, more efficient, more relevant decisions with impactful effects.
The quality and performance of the decision is directly related to the diversity and quantity of data processed for decision making. The industry already produces large volumes of data from production processes which already allow a strong creation of value using industry 4.0 solutions. The capture of industrial data then develops in its diversity component to pass a new stage, bringing a look upstream and ahead of the production chain.
This data collection is made possible through new communication technologies such as 5G, but not only. After having experienced the media campaigns on the impact of 5G in the industry, and elsewhere, at the beginning of 2021, I will try to present where to position this technology by its singularities.
Understand the industrial production cycle
Historically, understanding production cycle, first of all by human experience, then supplemented by computer systems, has made it possible to optimize industrial production. The challenge was mainly economic, but it is also becoming ecological.
Understanding the production cycle now means collecting data from raw materials, their own production, transport and storage. It also means recording each configuration, each state of the production system and, at the end of the chain, measuring the quality of production. It is also in the future to follow the product in its use until its recycling.
All the data resulting from each of these phases can be used to stabilize production and guarantee product quality over time. This results in better productivity (less breakage, less loss), fewer defects and better product longevity (less waste).
The relevance, and therefore the performance of economic and ecological gains, is directly linked to the quantity and diversity of the data acquired by our system. In this sense, the multiplication of data sources in the industrial environment is a key to industrial transformation 4.0.
Industrial production is in fact in the process of mutating, initially towards better use of existing data, from industrial automations, but will quickly be hungry for new, more diverse data, from the environment, people and machinery.
The challenge is to enable data acquisition at scale. That is to say to allow the collection of data, both in volume and diversity but also by limiting the constraints, among others in the deployment phases. Apart from the security issues associated with industry constraints are major obstacles to this acceleration of data collection.
In-depth modification of production tools is costly, impacts productivity and is often complex on old machines.
The renewal of machines, which are more communicating and can be operated remotely, is done over long periods. The integration of smarter machines into the information system will be a topic.
The addition of multiple and diverse environmental sensors comes up against the difficulty of mixing products whose safety is quickly questioned on the same network.
The use of solutions external to production machines, autonomous regarding industrial networks, was an initial response to these problems. Many integrated solution providers favor and will favor this approach. Many manufacturers retain them and will continue to favor them because they bring great speed during deployments.
5G for quality of service
In the private wireless broadband segment, private 5G and WiFi-6 compete. Price and availability give WiFi-6 a big lead. However, the proliferation of WiFi networks in the company, whether set up by the manufacturer or inherited during the deployment of third-party solutions, leads to a difficult control of the quality of service, which is not guaranteed by technology. For critical industrial applications where latency is key, collisions, saturations are not acceptable. Private 5G networks could be favored in applications such as control of production tools, automatons, AGVs. However, this deployment will probably take place later, while the density of WiFi-6 equipment will reveal its limits. Initially, it is conceivable that the manufacturers of these critical solutions will prefer to integrate WiFi-6 solutions on these same uses for reasons of cost, universality and simplicity of deployment, both technical and regulatory.
5G for isolation from the industrial network
Using a 5G solution to isolate yourself from the constraints of the industrial network is an accelerator for deployments. Indeed, a manufacturer must be particularly attentive to the security of a solution communicating on its production network. This translates into long studies, audits, pentesting which impact deployment times and costs. In addition, this requires the integrated service provider to reveal many manufacturing “secrets” on its solution.
At a time when the value added by the Industry 4.0 transformation remains to be demonstrated from the point of view of manufacturers – or at least remains to be supported by small-scale success stories – the speed of execution is as critical for manufacturers as for integrated service providers that go from PoC to PoC without concluding significant volumes.
The use of isolated networks and hosted services (SaaS) are the keys to accelerating projects, reducing costs and demonstrating the value of moving from PoC to deployment at scale. Even if technologies evolve over the rise in maturity.
In this sense, 5G has the advantage of wide and international distribution allowing integrated service providers to reduce the risk of unavailability during on-site deployments. This technology supports public use, without CAPEX, very affordable and reliable, particularly suitable for PoC contexts and deployment of small fleets on multiple sites. It also allows private use allowing an additional contribution in terms of controlling its SLAs and securing its data. These two criteria are particularly sensitive in the industrial environment.
5G outside the factory walls
Data collection, to ensure production optimization issues, will extend upstream and downstream of the production site. In addition, production is increasingly distributed, between specialized sites, throughout the manufacturing process. The collection of data that will be done in multiple sites and between sites will become an issue. 5G as a global communication network meets these needs, through low and high speed.
With the increase in remote working which meets both the expectations of employees but also presents itself as a necessity in the face of environmental issues, 5G will be the simplest solution to implement to ensure high-speed and low-latency connectivity for employees. Of the industry. It provides better security for the company by isolating the employee from his personal network and facilitates the equipment of the latter at his home or another place of his choice (third places, secondary residence, travel). The use cases around remote maintenance, remote intervention, support are made possible.
In this context, it is also necessary to take into consideration industrial needs where the production space is not interior and walled. Mining, agricultural, maritime industries where data collection needs are as wide as elsewhere and cannot rely on WiFi-6 type technologies, which are too short range, too energy consuming. These industries will have no choice but long-distance radio connectivity, including 5G.
5G for out-of-band and protection of industrial secrecy
Industry 4.0 comes not only from the manufacturer but also from the suppliers of the means of production of these manufacturers. Production machines, be it a tractor or an automaton, will no longer just collect raw data for the manufacturer. These machines will come with their own production optimization algorithms. The value added by these treatments will become the key element of differentiation between the players. In a way, it will dispossess the manufacturer of part of its production data, but will allow greater value to be added by incorporating data from multiple manufacturers in the same vertical into the production improvement algorithms.
Ambitions in terms of production optimization and environmental issues go through the creation of a collective intelligence that requires the globalization of production data from the same industrial sector. There are of course consequences on industrial secrets, of “anonymization” which are short-term brakes but which will be overcome over a period of ten years. First, the use cases associated with predictive maintenance will be deployed.
As a result, production machines, like our vehicles, will send more and more data back to their manufacturers. These manufacturers who will also want to push settings and updates to the machines.
The use of independent and easily accessible connectivity such as 5G meets the expectations of this use case; in addition, it makes it possible to hide from the manufacturer the data exchanges thus carried out and to protect them against filtering or espionage by the manufacturer. The issue concerns the intellectual property of production equipment manufacturers in terms of industrial maintenance and performance optimization.
This is potentially only a step towards controlling industrial machines from the cloud. Indeed, the choice to deport the production chain management intelligence in the cloud could provide a better protection against counterfeiting, against the seizure of industrial means and know-how in a less stable geopolitics context than the one we know. It also opens up possibilities in the protection of strategic production such as the military field. The principle is to place the intelligence of control, necessary to animate the production machine, not any more in the factory. Not either necessarily in the hands of the manufacturer who produces, but rather in control of the owner of the product and its industrialization process. This therefore includes many productions under license to widen the scope. Apart from this type of approach – we are talking about PLC in the cloud – requires minimum latency between the tool and its control, maximum service guarantee and high speeds. Different elements that are brought by 5G.
Partly private 5G for industry
This is probably the most important point that differentiates 5G from previous 3GPP technologies: a design that is really intended for private use. 5G is a telecom technology. It stands out from other technologies in the IT world. Telecoms have always supported quality of service issues in the face of user density and controlled latency. IT networks have favored the ever-increasing throughput approach to counteract the effects of saturation but do not guarantee congestion, collision processing and therefore latency.
However, in the control of an industrial process, there is no question of giving confidence to an unpredictable system. 5G, in the industrial world, used as a local communication network, in the same context as WiFi, therefore has enormous potential. This is reinforced by the strategic positioning of Amazon, Microsoft and Google, which now offer turnkey solutions for the deployment of this technology. It must be understood that 5G is a software solution where the part of the software is very important.
5G can be both deployed by the manufacturer with the aim of creating a local quality of service network infrastructure, but it can also be deployed by a service provider as a means of integrating these elements. The universality of the technology (even if to date there is still a diversity of sometimes incompatible frequencies) allows, as for WiFi, an ease of deployment whatever the geography.
Private 5G is also a response to the search for long horizons, up to 30 years, classic in the industry, but which oppose the horizons of the telecom sector in which services and offers are linked and replaced on periods of 5 to 10 years. Faced with an OPEX that would not be controllable, the private CAPEX approach provides better guarantees.
The main obstacle today lies in local regulations which can make deployments complex, costly and which restrict access to the spectrum differently from one country to another.
5G is complementary to a wider range of radio technologies.
The world of industry, wired networks users at first, then radio (WiFi) is faced with a dilemma, both to minimize the risks linked to cyberattacks and cyber spying on its production network and to both own and protect its data, while by multiplying the data collectors, in number and diversity. The multiplication of networks and physical/logical isolation is an issue in this evolution of the industry. The need for a physical organization of the plant, more dynamic, more adaptable, makes the use of radio networks more weighty. But it is also and above all the capture of data on the move in the factory, from personnel or vehicles, which initiated the switch to this type of network.
Radio networks will prevail in industrial production because, once the infrastructure is in place, they greatly accelerate all future deployments.
In this context, various technologies will be deployed with distinct uses:
- WiFi-6 should prevail for core uses. By this we mean the critical processes which must be mastered by the manufacturer and which cannot tolerate failure or data leakage. WiFi-6 can however be restricted to uses where the notion of real time is not critical, being unable to guarantee latency or totally quality of service. Solutions connected by WiFi-6 will be after an in-depth study of cyber risks and detailed consideration of their impact on the availability of the network for other uses. It can coexist with a private 5G network for the uses specified below.
- Wirepas should find an important place in low-speed, very low-energy consumption in mainly indoor applications in industrial buildings. It is a complement to WiFi-6 for 100% autonomous objects over long periods. Among other things on inventory applications, traceability and capture of environmental and safety data.
- LoRaWan should prevail for low-speed data collection. This includes the capture of low frequency environmental data, site security, anomaly detection, inventory, geo-location… The network offers the security of a private network, at low cost, for coverage of a site and beyond. It is also a technology approved by integrated solution providers who find it a cost-effective way to deploy their solution while minimizing the impact on the company’s network ecosystem, resulting in faster validation and deployment processes. . In addition, LoRaWan offers the possibility of autonomous data capture: without external energy supply. This point is key in the speed of implementation and the cost of deployment.
- Sigfox finds its place in particular in traceability use cases where economic performance takes precedence. Use cases in the inventory are also particularly promising. The technology will be distinguished by an additional cost to the object < $1 opening up possibilities that no other technology can. It will also stand out on ultra low power, that is to say the ability to communicate without internal energy but by exclusive external input. It seems that the use cases are broader concerning the life cycle of the object produced than in its production cycle however.
- Low-speed 5G / Nb-IoT should prevail in high-frequency data capture use cases. It benefits from the main advantages of LoRaWan but also offers a higher throughput and a quality of service that can be guaranteed. Of a public nature, the deployment of this technology may be limited by an Opex-type offer, whereas the other technologies mentioned above are of the Capex type. This solution also offers a use outside the walls which gives it a certain advantage for the uses of logistics and traceability which are to date the most widespread use cases in IoT. It is the technical limits, such as roaming, which would most limit its use with respect to global and innovative LoRaWan networks under construction such as Helium.
- High-speed 5G will find its place for component video, real-time and remote maintenance use cases. Its positioning should be where it is difficult or impossible to use WiFi-6.
- When you want a guarantee of quality of service, a real-time need, guaranteed latency as in everything related to the control of the production process or the use of mobile robotics.
- When you want to isolate yourself from the industrial network to speed up deployment by reducing security constraints.
- When you have to operate remotely outside the walls (teleworking, remote maintenance).
- When a device needs to be operated out-of-the-band (updating, maintenance by the manufacturer, collection of raw data for centralized refining).
A major line of thought for manufacturers will focus on the sustainability of these solutions. If WiFi solutions have been mostly backwards compatible since their creation, if Sigfox, LoRa, Wirepas solutions have been built with an industrial vocation, the image sent back by 5G is a technological leap forward that ignores the lifespan equipment that the mainly b2c market has nothing to do with. In other words, the industry is experiencing the end of 2G and then 3G, which is leading to major replacements of equipment. France is somewhat spared due to long-term contracts which extend these technologies but cannot escape them. Even under these specific conditions where 2G has been supported for more than 20 years, it is not certain that 6/7/8 G will not erase 5G over a horizon of 5 to 10 years, which would be far from complete. Industry builds solution for next 30 years. The commitment of solution and service providers must, it seems to me, also cover this time horizon to reassure manufacturers about their investments, in this direction, over the long term.