‘The objective is to minimise recovery time’ – the three-tier model for private 5G support
Spain-based Telefónica, as we have heard, is to establish regional network operations centres (NOCs) in Spain, Brazil, the UK, and Germany in order to manage private 5G networks on behalf of key industrial enclaves. This confirms what Enterprise IoT Insights heard previously, from Vodafone, that the operator community will build dedicated local resources, into their MEC infrastructure, to manage critical new 5G networks for enterprises.
It also puts the cat among the pigeons for a section of the commentariat, which doubts traditional telcos will have much of a role at all when the smoke clears over the new enterprise 5G landscape. A new report from Enterprise IoT Insights, about the role of operators in the ‘run’ phase of private 5G, suggests the many-headed enterprise sector will converge variously, and meaningfully, in the end around a hybrid public-private (PNI-NPN) model for enterprise 5G.
Whether all of the market shows such patience, as discussed recently by ABI Research, remains to be seen; at the same time, there is a tendency to conflate the enterprise market with early-adopter agitators among tier-one industrial types, and to forget most of the rank-and-file are following behind in Industry 4,0, more concerned in the early (?) post-pandemic era just to keep the lights on. What is clear is everything is to play, and anything goes.
And where Telefónica implies it is actively developing edge-based NOCs to provide specialist support for private users in manufacturing, alongside other key sectors, it also acknowledges its rollout strategy will develop as demand rises. Vodafone and Orange are more conservative about their private 5G NOC strategies, but they respond in the same way to the need for support at the edge of their national networks, and in the heart of enterprise.
“When you start to think about sub-networks, you start to get into a very different [NOC] model. You have a public network, but you also have these almost-dark areas inside where magic happens,” reflects Phil Skipper, group head of IoT strategy and business development at Vodafone Business. This new industrial magic casts different spells, he explains. It is concocted for “operational control of real-time processes”.
Vodafone proposes a “slice of the central NOC” for enterprises, staffed by specialists with specialist tools and specialist alarms, geared towards bespoke SLAs around network monitoring and performance. Skipper makes clear Vodafone is not ruling out the idea of dedicated regional NOCs to serve industrial strongholds, but that present demand, and business logic, dictates existing resources should be sweated as possible.
He comments: “The thing is the underlying technology doesn’t change; what changes is the way you manage the service. The secret is to use the same infrastructure to deploy a different service.” He turns to the question about how operators will provide urgent support in the field. “The bit people miss is there are two parts to this: how to manage things when the network is running fine, and what you do when things go wrong,” he says.
“The first part, about preemptive maintenance, is suited to a [national] NOC. Which is what we provide as a managed service with private networks – so the network is configured and performant. The second is about intervention services, and where to position them. For this, we see three options: these services are owned by the customer, co-located between the supplier and customer; or provided on-call, like if your boiler breaks.”
He adds: “In the end, the choice depends on the customer’s view of criticality.”
Arguably, Vodafone’s three-tier model (see below) is ordered funny. The co-location service (number two) puts a Vodafone engineer on-site, as per the Telefónica example, and is geared towards ‘mission critical’ infrastructure – for “making sure systems have the ultimate responsibility, which is more expensive”. The customer-owned support function (one) is for urgent box-replacement work – “a guy in a maintenance shed who can swap a board out if it goes wrong”.
In both these scenarios, the use case dictates network downtime and network fixes are time critical. “What you can’t do is have an engineer turn up, whistle through his teeth, and tell you he hasn’t got it on the van. ‘It’s the Pomegranate 900; I’m going to have to call my mate’. The objective is to minimise the recovery time, and the quickest way to do that is to replace the part. That is the philosophy – a one-hit fix, every time,” says Skipper.
Where the co-location solution (one) doubles and deepens the expertise on site, the on-call setup (three) outsources the work to local third-party technicians, and works where network uptime is measured in miles and not minutes. These second-party (enterprise-owned; number two) and third-party(one and three) services are supported remotely from a central NOC. Vodafone is issuing augmented and virtual reality (AR and VR) gear to walk-through fixes.
Orange, another operator widely credited for its approach to private 5G, is following the same recipe, combining (augmented) central NOC functions with local support services in each market. It describes its approach as “global-local”, with five “major support centres around the world” and “local support where needed”. Support comes from “field services organization and capabilities”, it says.
The firm is targeting medium-to-large multinationals in manufacturing, logistics, and the energy and resources sectors. A spokesperson for Orange Business Services (OBS), its enterprise division, says: “We are adding mobile private network NOC capabilities first in Europe and then in a phased approach to other regions based on demand… In general, low-latency networks can be monitored real-time from a central location. However, based on the situation, it may warrant a local on-site presence to complement the central NOC.”
Specifically, OBS cites computer vision use cases overseeing critical equipment and manufacturing processes. Which seems clear, then: centralised network monitoring and control, supported by localised engineering services distributed throughout the edge network. It is not, of course. You only have to look through the various commentary across the rest of this magazine to know the market sees elephant traps for operators everywhere.
But we should reintroduce another voice at this point, one that is playing the market from all angles. Nokia, despite the noise around its straight-to-enterprise strategy, will work with whoever gets the job done, it says; the governing factors are only about architectural demands, regional availability, and good old customer relationships. “We are completely agnostic,” remarks Stephane Daeuble, head of enterprise solutions marketing at Nokia.
Interestingly, he suggests mobile operators with SI-style divisions tend to have grasped the private 5G nettle better. “They have a stronger play to the enterprise, because they are known to them, and they speak the enterprise language. And they can bring real value.” He mentions Deutsche Telekom with T-Systems, American duo AT&T and Verizon; in particular he tells of work Nokia has completed for an unnamed chemical plant with OBS in France.
“OBS really impressed us. Because it knew the business and talked the enterprise language – instead of just trying to sell a private network. It started with the use cases and applications, and performance and latency requirements. We were like, ‘Wow, these guys…’. And it was probably the highest level of reliability and availability we have ever had to build into a network.”
He cannot say much; the official announcement is pending. But he sketches an outline. The pair designed an industrial-grade network with three levels of core network redundancy, to get to 5x9s (99.999 percent) availability, says Daeuble. This involved two standalone private cores on site, “so if the left side burns the right side functions”, and Orange’s public network geared as a last resort to backhaul radio transmissions via a local “co-facility” in case the whole plant goes up.
Meanwhile, the radio setup operates in (at least) two different spectrum ranges: the private 2.6 GHz TDD band, reserved by regulator Arcep for enterprises, plus a tranche of public spectrum leased by Orange to the customer locally. This, says Daeuble, created “double-layer coverage” with small cells affixed to poles at different angles, a couple of metres apart, in order to get behind each fixture and fitting and “make the system available”.
It is a good story. But network design, and indeed most network performance metrics, are mapped out in the early ‘sell’ and ‘build’ phases, as the Nokia story relates, and not in the network management ‘run’. What about ongoing management of this double-coverage, triple-redundancy setup? Daeuble says OBS is “babysitting” the network for the client. “Because it is a highly critical network, and if something goes wrong…”
The subtext here, with a chemical plant, is the risk to human life and the natural environment, as well as to business itself. There is not much more to say on network management at the plant, as yet; there is no word on how OBS is handling it, for instance. But the story makes clear that operators are, indeed, engaged in critical-grade private 5G projects from the start, and also trusted with their in-life care, at least in the short term.
The jeopardy for operators with private 5G is overplayed – of course / perhaps.