‘Scale is the thing’ – Schneider Electric sets out strategy as private 5G user and reseller
Note, for more on this topic, and more on Schneider Electric’s private 5G story, catch Zach Nimboorkar on the Enterprise IoT Insights webinar on April 28 on Industrial 5G (From Setting Standard to Becoming Standard). He is joined by speakers from ABI Research, MFA (MulteFire Alliance)/Nokia, and Vodafone.
Schneider Electric, through a new trial partnership with NTT, finds itself in the auspicious position – as much for us – of being both an enterprise user and a channel reseller of private 5G. Its dual role is a little different to Bosch and Siemens, with similar profiles as factory mega-users, but with alternate visions of the reseller game. The former is seizing on 5G to equip and sell industrial devices, and the latter is seeking to develop and sell its own 5G systems.
Schneider Electric, by contrast, is happy to take 5G from someone else, ostensibly with NTT (which is itself leveraging US-based core network provider Celona and various RAN-others), and to integrate and collaborate on a shared sales agenda. But it has a problem. Like the rest of the Industry 4.0 market, it likes that private 5G, attached to dedicated spectrum, brings OT separation from general-purpose IT networks; at the same time, it wants to be reliant on the public cloud, in order to hand remote control to OT specialists on factory floors.
In other words, a private 5G network in a ring-fenced on-site edge compute architecture does not work. It wants backhaul access – low-latency, high-reliability, sliced or separated – to AWS, Azure, and all the rest. In terms of compute hardware and analytics software, it wants the heavy lifting to happen in the cloud, and not at the edge – and not even perhaps at the network (MEC) edge. And for this, it is unwilling to traverse the standard IT network.
Zach Nimboorkar is senior vice president for global IT infrastructure and operations at Schneider Electric, which means he is in charge of 5G in the French firm’s entire factory footprint, which spreads to 200-odd sites (and 800 locations, all-in) across the globe. It is a big undertaking, and describes, in microcosm – for a single enterprise user, with centralised budgetary control and well-defined use cases, but with its own industrial hangups and predilections, plus the headache of conflicting spectrum policy – the challenge to make industrial 5G scale.
Nimboorkar explains: “We are in exploration-mode with 5G. Private 5G is still maturing; it still leverages a lot of the traditional transport-to-backhaul traffic from the manufacturing plant to the cloud provider, to Azure or AWS, say. We are still using the same corporate software-defined transport network to carry data from the factory 5G network into the corporate IT network, which then takes it to its destination in the cloud.
“And ultimately, we want it offloaded directly to the cloud. Because OT should not have to rely on the corporate IT network. Period. It requires an edge solution so the traffic goes directly from private 5G to carrier-grade 5G, and out to its destination – leaving the corporate network completely out of the equation. That is what we want to see. That is the star everybody is watching for.” This is what it is developing with NTT, as a first proof for their new co-creation work.
Nimboorkar explains: “You have so many OT-connected sensors in the manufacturing plant; as soon as you start backhauling them over a corporate network, the whole thing loses its OT flavour. All these IoT sensors need to go directly to the cloud, as opposed to relying on traditional IT. We want to offload OT traffic directly via 5G. Because what is the point of 5G [for OT] if all the data goes over the IT network? You might as well use Wi-Fi.”
The NTT project centres, so far, on proving a machine-vision case for “wear-and-tear” on a factory conveyor at its self-styled ‘smart factory’ in Lexington, in Kentucky; it is actually a fairly standard proof point for industrial 5G, as described by many others. The trial uses Schneider Electric’s EcoStruxure data centre infrastructure management (DCIM) platform, coupled with industrial software from AVEVA, in which Schneider Electric also has a stake.
This complicates the picture; the AVEVA software, as per the IT trend with industrial OT and IoT software, resides in the cloud, and is sold as-a-service. Schneider Electric does not want, at this stage, to move it to the enterprise edge. Nimboorkar says: “The current AVEVA [model] is cloud-hosted for a reason… because we want to sell as-a-service to customers. So the customer does not have to invest in locally-hosted hardware in their manufacturing plants.”
Until now, images from the conveyor have travelled to and from AVEVA in the cloud over a Wi-Fi access network on the factory floor, and a corporate IT network in the backend. The NTT trial is considering, firstly, whether the existing IT network will hold up for the backhaul, if the data is switched instead from a private 5G setup. “We know we can get the packet from point A to point B; that is not a problem. The problem is whether the data coming back from AVEVA is real-time [enough], and whether the operator can act on it before an outage occurs.”
He says: “That is what we are trying to prove out – [while still] traversing over the corporate network. Because the direct sensor-to-cloud 5G access is not there yet; we’re still leveraging the corporate network for that.” But the NTT experiment has another dimension, too; to enable direct-to-cloud access by moving the cloud-compute service nearer to the industrial edge – by creating (and selling) new regional edge data-centre infrastructure.
This is the extended opportunity, and the reseller spin, for Schneider Electric; to double-up with sales of its ‘prefabricated’ data centres – easy drop-shipped computing containers, designed to go into war zones and music festivals, and all quick-fix compute points in between. The idea is these can be bundled with its EcoStruxure DCIM platform, and whatever compute and analytics gadgetry from NTT – and taken to the wider Industry 4.0 market.
“We want to provide… our prefab data centres [into the bargain] – so NTT handles the 5G network at the site, and populates our data centre [somewhere near] with edge components and a 5G antenna. So the sensor is connected to the 5G network, and the 5G network is connected to the edge data centre. Which provides a direct OT-to-cloud 5G access connection, as opposed to today, where you are going through the corporate side of things.”
He explains the model: “The container [plus management software] is from Schneider; all the services are from NTT. The 5G edge network [and edge extension] could be operated by NTT or by the customer. There are two ways to do it: as a fully managed service from NTT, or a hybrid model where the customer maintains everything at the site, and NTT maintains everything at that edge. NTT could have these POPs globally; there are no limits to it.”
And then he switches focus to the scale question. “Once we have identified the value [with the conveyor case], we have hundreds of plants with conveyor belts – and we can multiply the downtime saving across all of them. That is the way this will scale. And this data centre model with NTT will benefit other customers as well, which are looking for that direct-to-cloud connectivity out of their OT, wherever their data is hosted. That is the next evolution.”
It is an evolution that has been gathering pace, separately, in the network edge, of course, with tier-one operators agreeing to retrofit edge instances of hyperscaler cloud-platforms into their network infrastructure (in local data centres; points-of-presence / POPs). Nimboorkar implies NTT, offering edge networking and compute solutions as a managed service in any market, will work its carrier relationships, as required.
“NTT could also leverage services from last-mile providers, like AT&T or Vodafone. But the enterprise relationship will be with NTT, whatever its sub-contracts and sub-relationships,” he says. But Schneider Electric’s prefab alternative will work for plenty of Industry 4.0 setups, the message goes, with or without the carriers. “NTT is interested in as-a-service solutions, so the customer says, ‘Hey, I need a container’, and it is populated and shipped to them.”
And slicing? Does a slice of a public 5G network hold any allure for the industrial set? “Absolutely,” responds Nimboorkar. Slicing will be used on a variety of infrastructure, he implies; NTT, at the controls, will slice and dice, in concert with public operators, as necessary. “We will provide our requirements (for our factories), but the slicing happens with NTT, and at carrier level.”
He adds: “Our goal with this is to prove-out the transport from the OT to the cloud. If we see there are latency issues, because we are backhauling over a corporate network, then we have this compute edge proposition with NTT – so we could quickly pivot. It is a hybrid cloud service, which brings the cloud to the OT environment.” It is curious, as an aside, that Nimboorkar, in charge of IT for Schneider Electric, is tasked with keeping OT separate from it.
He explains. “OT increasingly needed connectivity back to the cloud, and relied on IT to facilitate it. Which brought some network segmentation, but also a lot of cross-collaboration with the OT innovation team. But private 5G, in its truest form, like we’ve discussed, brings this new capability, to take traffic straight to the cloud – which means, potentially, OT doesn’t need IT anymore. But right now the way it works, OT still retains a pretty heavy reliance on IT.
Interesting; but what about the big question, about how 5G trials scale from one factory to many, and from one use case to many? “Yes, I mean, if we identify the value for OT, then it scales, very quickly to 200-300 other plants. This case, with the conveyor in Lexington, could be replicated. But we have looked at various others – VR glasses, say, for customer tours, maintenance, things like that – and they have not been so [clear].
He goes on: “Use-case identification is critical. You know, 5G will be native in new machinery, and will be there in new factories. But we are talking mostly about backwards compatibility. And it wasn’t easy, actually, working with the OT team, to see what 5G could be applied to. The conveyor is an obvious pain point. We have a bunch of other stuff with NTT. But, again, they are proofs; if we can’t find the value, we will move on. If we can, we will scale. Because scale is the most important thing. Yes, I could have one plant using private 5G, but, then, so what?”
For more on this topic, and more on Schneider Electric’s private 5G story, catch Zach Nimboorkar on the Enterprise IoT Insights webinar on April 28 on Industrial 5G (From Setting Standard to Becoming Standard). He is joined by speakers from ABI Research, MFA (MulteFire Alliance)/Nokia, and Vodafone.