The telecoms industry is at odds with the enterprise sector it is seeking to serve with ultra-reliable, low-latency (URLLC) 5G communications. Industrialists want both the control of private networks and the convenience of managed network services, and network operators want to bunch digital enterprises together, and treat them as they always have.
This was the final position of the two sectors at the end of the first day of inter-market discussions at URLLC 2018 in London, exploring the feasibility of 5G for ultra-reliable low-latency use cases. Brendan Lynch, board member of the Worcestershire 5G Consortium, one of six 5G testbeds in the UK, explained the disconnect.
“The industry wants the control of a private network and the convenience of something that is managed as a service. But I don’t think there’s the understanding [within industry] of what it takes to manage a sophisticated network. We haven’t bottomed that out yet,” he said.
“We need to promote a discussion to bring the ecosystem together – to help the manufacturing side understand the challenge of managing a network, and the telco side understand what it means to put a bespoke network into a factory, with its different requirements.”
The industrial sector has no room for blind faith in the promise of 5G systems. German manufacturer Bosch made clear its position on wireless services on its shop floor is non-negotiable. “We would never rely on public networks for critical services,” said Andreas Mueller, the company’s project manager for corporate research.
Debate around network management and spectrum allocations for vertical markets are developing. Mueller alluded to regulatory discussions in Germany about whether to carve out spectrum for industrial usage.
Japanese operator NTT DOCOMO, in conversation with Mueller on stage at URLLC 2018, acknowledged operators must be flexible and collaborative in their approach to the industrial space. “Maybe we need to develop a new business model,” said Takehiro Nakamura, senior vice president and general manager for the company’s 5G Labs.
“Each party has a strong interest in the cost of it. Even for us, of course, we want money from new services. But we need to collaborate with industrial enterprises, and with factory owners perhaps. We need to have that discussion – about how to share the money, how to manage the networks, and which spectrum bands to use.”
Nakamura suggested a role for operators “like a system engineer” in industrial plants. “We have excellent skills in radio systems,” he noted.
In an adjacent dialogue on industrial transformation, Johannes Springer, project lead for Deutsche Telekom’s 5G automotive programme, said a single network can support multiple use cases, reliably and cost effectively. “These industries can be operated as logical layers of one physical network,” he said.
UK operators BT and O2 are scoping out the new industrial landscape as the telecoms representatives in the Worcestershire 5G Consortium, working alongside system integrator AWTG, as well as key regional enterprises, in the shape of boiler maker Worcester Bosch, machine tool manufacturer Mazak, and defence technology firm Qinetiq.
“We are getting the telco folk with telco KPIs and the manufacturing folk with manufacturing KPIs to talk the same language,” explained Lynch.
“There is a paradigm shift from previous-generation of networks, where customers essentially got what they were given, to next-generation networks, where industry specifies its requirements and engages operators to figure out a trade-off between latency and throughput and so on.”
And yet a sub-text at the URLLC 2018 event was the operator community’s rather late interest, and lack of progress, in the industrial sector. Derek Long, head of telecoms and mobile at Cambridge Consultants, said the unique character of industrial sites provides good case studies for the more bespoke aspects of 5G.
“Manufacturing is geographically limited and under the jurisdiction of a single organisation. These parameters have attracted a lot of interest from lot of industry – perhaps less so from the telecoms industry so far, but certainly from IT and OT vendors,” he explained, pointing to the high number of industrial IoT platforms in the market.
Mueller at Bosch suggested the industrial sector is wary of the hype around 5G, and will not easily allow wireless networks into its plants. Mobile operators may even struggle to gain a firm foothold on factory floors until 2030, he suggested, when next-generation 6G networks will be lighting up.
“5G makes lots of promises about URLLC, but it has a lot of work to do to prove its availability and reliability. It promises everything, but so did 3G – and it took 4G to realise much of that vision. It could be similar with 5G – that it takes another 10 years to deliver what it originally promised,”
DOCOMO restated the Japanese roadmap for 5G, with services scheduled to go live in time for the 2020 Summer Olympics in Tokyo. But the first focus for NTT DOCOMO, like other operators, is enhanced mobile broadband (eMBB), as an extension of its traditional subscriber business.
Nakamura presented outdoor URLLC trials with Huawei and Ericsson, and slides on 5G demos with Komatsu and NS Solutions, respectively, for the remote control of construction equipment, and a concept for a 5G enabled humanoid robot.
But DOCOMO’s trials have invariably focused on data rates and capacity, and have mostly taken place outdoors. It has hardly touched industrial plants and factories. “URLLC will come later,” said Nakamura.
“There has been no decision on when, but we know they offer important and interesting new services for the 5G era. We want to provide those services as soon as possible, but we need to improve many things – the radio network, the core network. It takes time.”
But for industrial operatives like Bosch, a trailblazer for Industry 4.0, with 300 manufacturing plants worldwide, reliability and security are everything. They are more important than speed and throughput, in the end. It comes back to the earlier comment about the work operators must do to prove their technology in new market segments.
The competency of 5G networks to deliver such guarantees are unclear, said Mueller. One mistake, and the whole house comes down. “If a network is disturbed with a jammer outside a factory building, and it brings down the production line, then there is huge damage. And it happens once, and everyone switches off their 5G systems.”
The Worcester 5G Consortium, funded as part of the UK government’s £1.1 billion ‘digital infrastructure investment fund’, has been running since April. Its remit is to deliver productivity gains of between 1.5 and three per cent in the Midlands by smartening-up local manufacturing, which contributes up to 25 per cent of UK manufacturing, and about 13 per cent of total GDP.
The project has only just started building a network of smart manufacturing sites, connecting back to the University of Surrey’s 5G Innovation Centre. It focuses on three uses cases: data analytics and predictive maintenance, augmented reality and assisted maintenance, and security-by-design. The first results of its trials will be published in early 2019.
“It is early days in the UK. But what is manifestly clear is this is a great opportunity – and perhaps a once-in-a-lifetime opportunity to harness the potential of URLLC communications, and put the UK on the front foot,” said Lynch.
At the same time, Lynch, like other speakers, returned again to the challenge for 5G to prove its ultra reliability. “Reliability is key; it’s the first thing. The manufacturing industry expects 99.999 per cent to mean just that – when it asks for ‘5x9s’ reliability, that is what it wants. If 5G networks are not reliable, then that ecosystem won’t come together as harmoniously as it needs to,” he said.
In general, the event confirmed the urgent requirement for deeper collaboration, and urged patience to unravel the operational minutiae of 5G implementation. “There’s a great onus on industry to figure out its requirements and have a greater input into design,” commented Lynch.
Springer at Deutsche Telekom commented: “When it comes to the vision for the next 10-20 years, we have to understand quality-of-service, and this sector has to understand the industrial requirements, whether from the automotive, automation, or aviation market.”
Mueller said the challenge of plant optimisation is more profound than simply replacing cable systems with 5G radio units. “There needs to be a complete system re-design to achieve best performance, by joining up the communications infrastructure and production environment,” he said.
In the end, as always, the challenge is cultural, too. “It takes more than a standard to change how people think. It will take industrial control applications that are more tolerant, and resistant to jitter. But this will take 10 years. The industrial domain is very conservative, especially compared with the ICT industry,” said Mueller.
