Good 5G kit is bad 5G kit if the network design fails – especially in private networks

You can have the best 5G network in the world, and make it into the worst without much effort; or, at least, you can render it as completely ordinary, and even maybe totally useless. This is the warning from Nazim Choudhury, director of market development at Canada-based iBwave Solutions. His point, as one might expect, given his company’s interest in network planning, is that good infrastructure design is half the battle, especially when deploying so-called carrier-grade or critical-grade cellular networks. “It is not like stringing up a bunch of Wi-Fi access points,” he says.

Which is how many customers and suppliers have approached the discipline in the enterprise space, where cellular has found a new footing lately with a flurry of mid-band spectrum liberalisation for enterprises in certain geographies and a bunch of in-band tech features for enterprises in new 5G releases. “The way with Wi-Fi is to survey and plan and design, all at once. Whereas cellular requires pre-planning, and then surveying, and then optimisation. There are more variables; it takes much more coordination. If you take the Wi-Fi approach, you’re going to run into issues.”

There were major issues a couple of years back, if we recall, with private 4G/5G test networks in Industry 4.0 venues. There was a lot of messing around, right – to get a steady LTE signal inside an assembly on a manufacturing floor, for example, or to stabilise the uplink channel on a video feed?  “Yes, exactly,” responds Choudhury. “And we were part of some of these trials – except they brought us in post-deployment, after things had gone wrong. And, you know, we showed them that, with proper planning, the outcomes can be much better.”

He adds: “Some of these private 4G/5G trials have failed because of improper planning – because they have taken a more Wi-Fi-centric approach, thinking they can just throw up these access points almost anywhere.” Indeed, he suggests the sort of easy shop-window sales – which certain hyperscalers, notably, sought to pioneer from disparate core and radio parts a few years ago; which have gone quiet in the meantime – effectively imposed a do-it-your-self Wi-Fi design ethos on the more rarefied task of cellular radio engineering.   

“Some offered very minimal planning – like you could just order based on square-footage. And that approach hasn’t been successful because the market understands now that planning is crucial,” he says. “You can have the best small cell, with the best specs and the best ratings, and put it in the wrong place. You won’t get the best from it unless it is planned properly. Poor radio placement – because it is enclosed, or because of the setting or signal interference – means a poor network. It means you don’t benefit from the technology, even if it is best-in-class.”

Large-scale outdoor cellular macro networks are one thing; smaller sized micro-campus installations, indoors and outdoors in complex industrial venues are another. The sale of private 5G networks in the broad Industry 4.0 sector, which actually perform as they are supposed to, is a gnarlier business. For starters, there are a bunch of other access technologies already in play, including Wi-Fi and Bluetooth (invariably), and also the likes of LoRaWAN and RFID (very often). The bands are different, but switching and segmentation of devices is a planning-phase consideration. 

But mostly, the clank and boom of Industry 4.0 just messes with cellular propagation. “Think of all the metal surfaces and heavy machinery that will impact the signal,” says Choudhury. His firm started in the in-building DAS market in 2003, with a plan, hatched by its founders, veterans of Telus and Bell, to standardise network design tools, as used variously by network operators, vendors, and integrators. “From there, and as we were adopted by the Canadian and American carriers, it just started to evolve – from a documentation tool to a design tool, to a network simulation tool.” 

Choudhury continues: “And with modelling, you have to understand all the different materials that impact signal propagation and transmission losses – in all different frequencies in all different environments. That’s how it grew; and the tier-one carriers wanted live data points, and so we partnered with different collection tools. Which all go into help with network planning – to define the network, manage the costs, and show the benefits of pre-planning versus doing it on the fly, almost, like in the Wi-Fi scenario.” 

Again, the kind of pre-planning afforded by iBwave’s software is critical for complex 4G/5G installations in enterprise venues, argues Choudhury. Anecdotally, he tells about recent experiences with 5G deployments for “very big events and big facilities” in the US, whether as private instances using CBRS spectrum or as public/private projects using carrier airwaves, or some combination of both. “You really need to pre-plan these 50,000 square-foot deployments. You can’t just plan-as-you-go like with a Wi-Fi network.”

The latest version of the iBwave network-planner scanner integrates 5G with LTE (4G) and Project 25 (P25) network planning capabilities. The Montreal-based firm launched separate private-networks software in 2022 to help US enterprises dovetail their local-area private Wi-Fi and LTE/5G designs in CBRS spectrum in a single platform. The new version incorporates this, too. The app works with a handheld PRiSM Scanner device from Epiq Solutions, which attaches to a regular smartphone. It is pitching to field technicians and IT teams in enterprise companies. 

It means enterprises and public-safety organisations, plus network operators and system integrators, can use the firm’s mobile app for faster and cheaper network surveys, it reckons. Choudhury says: “The key thing is that we are taking a technology-agnostic approach. It doesn’t matter if the network is Wi-Fi 6 or 7, or 4G or 5G; it’s about what’s best for the enterprise use cases – especially with private networks. Often, mostly, it’s a hybrid solution – mmWave in the warehouse, and Wi-Fi in the conference rooms. All these technologies are just part of the toolbox.”

He adds: “Our software provides a roadmap, of sorts, so the network can be simulated and advanced with all of these technologies as the use cases develop. It doesn’t matter if it’s cellular, Bluetooth, RFID even; what matters is that the right technology gets deployed for the right use case, and the whole system is future-proofed. It creates a digital twin, which evolves with the network. So you know what’s going on inside the facility all the time. And so a year down the line, when you have a use case for AI cameras in the halls, you have a live replica to know where to put them.”

But there are interesting dynamics in play right now in the private 5G market, he suggests, borne from escalating competition between vendors, which are – ironically, to an extent – looking to ape the Wi-Fi kit sales with smaller off-the-shelf systems. It means the more careful engineering of traditional cellular systems, perhaps made to appear like over-engineering in the box-shifting enterprise space, has been lost, he argues. “The problem is it’s a very competitive market, all these players are bringing out quick and easy solutions.”

He explains: “That’s the biggest issue in the private networks space right now. They’re taking this Wi-Fi-centric approach. It’s plug-and-play, so very little design [is supposed to be needed]; you just order them like access points. And it means the whole engineered design process gets pushed to the side – because a competitor is coming out with a rival product in two weeks. It’s gone from a couple of weeks of design and months of optimisation to something that can be designed in two days. So it’s about finding a fine balance between accurate enough and fast enough.”

This hyper-release schedule is damaging vendors’ brands, he says. “Their reputations have been impacted by improper design – because their products weren’t designed properly. These best-in-class products haven’t been performing. That’s what we hear from the OEMs. But at the same time, they’re not willing to wait two weeks to design this. So they’re looking for us to support them to get the designs done in two days instead of two weeks.” The pressure is coming from the operator community, as well, he says.

“The carriers want to know how to make it faster, too. ‘We’re trying to sell private networks in a competitive environment now,’ they say.” In particular, they are pushing quick-sale private 5G to the retail sector, he suggests. “All these Walmarts and retail stores, and such.” The challenge is for iBwave to solve, as well. The supply-market is rushing because the demand-market is rising, the message goes. But it is also because the demand-market says so. Enterprises want simpler cellular networks, which are not hard to build and manage. 

“It’s what we’ve been trying to address, also through different software applications. So, how do we become accurate enough and fast enough?” The implication is that enterprises should be trained, to some extent, along with integrators, and even operators and vendors; but the network planning tools, as provided by the likes of iBwave, should also be intuitive, too. At the end of last year, the company introduced a ‘stencil library’ of Industry 4.0 computer design (CAD) elements, to drop into network schematics.  

It describes them as “user-defined templates of frequently used markups, images, 3D materials, and objects” – things that directly require or impact the cellular signal in the enterprise venue. Instead of modelling elements every time, designers can simply drag the objects onto a floor plan, and store them in a library for easy access. “That’s an example of how we’re trying to innovate. You have all these Industry 4.0 stencils – of materials, machinery, things like AMRs and AGVs; and that could save two hours just there,” says Choudhury. 

He adds: “We’re launching different products to help speed-up deployments. We’re continuously evolving because it’s a problem for us to solve: how to go from two weeks to two days? So that’s the push in our roadmap and technology.” Does AI hold the answer? Is AI something that, from an engineering point of view, iBwave expects to utilise in serious fashion? Definitely, he says. “AI/ML has a pivotal role. There are so many functionalities that could support us, especially in these different Industry 4.0 environments – to get to two days, or whatever.”

As well, the company is looking to make more its network twin service – so the original simulation is a living and breathing reflection of the physical enterprise network, which can be tweaked to get sight of future infrastructure scenarios, to support future applications and workloads. Choudhury doesn’t want it to be called a service. He says: “It is more of an outcome, from the design [service]. You use the design tool and you get a digital twin. That’s the first part; the second is that it plays in a larger ecosystem of digital twins.” 

The point is to connect the network twin to the whole enterprise twin, and have them play off each other. He goes on, and finishes up: “For it to be a true digital twin, it needs to be interoperable with different systems. So we developed an API to pull impactful information into other platforms – so it helps to support that real live replica of the whole venue. And there is certainly a way for AI to help there – with automatic scaling, anomaly detection, predictive maintenance. That offers huge value to enterprises. It is something we are looking at.”