While many countries have re-used existing spectrum for 5G today, factors such as the projected growth in data usage alongside 5G adoption, demand for faster home broadband, and, in time, the use of higher frequencies are all encouraging network operators into building out fibre assets and towers in greater numbers. Indeed, GM Insights predicts that annual spending on network infrastructure will grow 50% between 2020 and 2027.

While no doubt welcome for network equipment vendors, it will create a significant extra ongoing burden for upgrades and maintenance. There are a variety of approaches being taken to solve this in the near to longer term.

“If we talk about bringing coverage to rural areas, it’s very expensive to go and install it,” Eugina Jordan, CMO at Parallel Wireless, pointed out. “It’s also very expensive to go and optimise it. One tower climb is $5,000 – just for one tower climb.”

There is a further, surprising element that can add cost and delay to maintenance – not all operators are very clear on where exactly their physical resources are. 5G location company Polte has developed an appreciation of this, as they depend on accurate knowledge of towers to locate other devices.

“Some are better than others, I will say that. What we found is Deutsche Telekom, SoftBank, and Verizon have the best tower databases. Apart from those three – everybody else, not so much,” said CEO Ed Chao. As a result, Polte developed their own directory of tower locations worldwide, which operators themselves frequently use.

With that obstacle overcome, Jordan discussed the opportunities of disaggregated hardware and software to reduce costs. With software enabling configuration and optimisation remotely, she argued, “You don’t have to send a very expensive engineer. You can send a TV installer. And then once it’s installed, you configure from the data center.”

Equally, she pointed out, radio hardware can be made less complex and costly.

“You need the radio; you cannot turn the radio to software. What you can do is create software like DU [Distributed Unit] or CU [Central Unit] software and real time RIC [RAN Intelligent Controller] to enable that radio to be smarter. Then you can take any dumb radio and make it smarter.”

When it comes to maintenance, Jordan distinguished two areas that are becoming less costly through disaggregation.

“The first area is software upgrades – new features, bug fixes. In mobile operators the IT organisation moved to a CI/CD [continuous integration, continuous development] approach years ago, but the wireless or the telecom part is still in the dark ages.

“Open RAN is now moving onto COTS servers and enabling CI/CD. Actually, with IPT Peru we’ve been doing remote upgrades through a CI/CD framework for years. They have thousands of sites there in Peru.”

The other part consists of physical optimisation of sites.

“Right now it’s a big revenue generator for legacy vendors,” Jordan noted. Indeed, Ericsson’s Managed Services arm typically takes revenues of over half a billion dollars per quarter. However, this too is an area where automation is taking a role – back to the idea of “taking a dumb antenna and making it smart”. Again, by creating automated systems for optimization of services, such as Open RAN’s RIC, it reduces the requirement for sending engineers on site.

“Optimisation happens via AI/ ML. The only ones that lose money are the legacy vendors, because they’re not innovating.”

While this activity reflects a current and near future development, BT’s Adastral Park research facility is looking further ahead to simplify work for a field force that will likely be increasingly stretched.

In November BT Group announced their plans to begin research into robots for fibre installation and maintenance. Managing Director of Research Tim Whitley was able to spell out the mission in more detail.

“Are there areas of technology that can help with the sort of tasks that we know we’re going to be doing for many years?” he asked. “It’s very practical things like when you’re deploying fibre cables, you encounter blocked ducts under the ground, that are the very base of this infrastructure. They get silted up, they collapse, tree roots grow in them.”

Practical problems such as these can add hours to a field engineer’s task or halt it altogether. While BT’s CEO has committed to extensive fibre rollouts over the coming decade, “These sorts of things just get in the way of building like fury.”

As a result, Adastral Park is working on a variety of robotic companions to fulfil particular functions, such as drilling new underground ducts, unblocking existing ones or accessing the tops of telegraph poles. Whitley, who began his career at BT as a field engineer himself, is focused on how to support such a field force.

“This is very much about providing a suite of tools which really help the field engineers who are contemplating, ‘Gosh, are we going to have to close this road?’ Can they deploy a little robot that maybe reuses the existing copper wire overhead to carry the optical fiber cable across the road, allowing them to complete the task quickly, or simply safely?”

The area of robotics has been fast evolving to support a wide variety of industries, from logistics through to aviation and manufacturing. As a result there is a good deal of work that has been done already on many of the technologies BT might employ – such as in drills for Martian probes.

“We’ve got a hole-drilling issue a bit closer to home. Mars exploration is quite different to delivering your fibre broadband, so there’s a technology translation thing we’re doing,” Whitley explained.

At Adastral Park the research organisation aims to bring together experts from many different backgrounds and specialisations from outside BT. This is necessary to assemble the right solutions for locomotion, power management, control, actuators and much more.

While Whitley is highly aware of the difficulties facing the project – not least managing the complexity of developing these robotic tools – there is one real measure of success over the next few years for the programme.

“The real challenge is – does it really work in the real world? I hate that phrase in a way, but the result you can get in a lab is one thing, making it work in a real operational environment is unusually challenging.”

More than that, bringing innovations to commercial use – to a point where there are devices that can be used not only by BT but potentially by utilities and telecoms providers globally – adds an extra dimension to the challenge.

To a degree, robotic solutions will be competing with alternative methods of infrastructure delivery which aim to reduce cost and accelerate delivery by adopting different models. Open RAN disaggregation and simplification of infrastructure is one approach, but a number of countries are also exploring different methods to reduce costs.

Within the UK, examples like B4RN in Yorkshire brought together a collective of landowners and farmers to dig fibre trenches across their land to villages and towns otherwise bypassed by fibre builds, in exchange for access to fast broadband and a small rent.

Such “DIY networks” can be built at a fraction of the cost that the necessary permitting, wayleaves and more would cost a major corporation, not least because much of the work is being done by volunteers. Robotic tools need to provide a very tangible benefit for a company like BT, as Whitley is keenly aware.

“The challenges, it seems to me, are often less technical and more to do with the value chains of who does what, for whom, to what end,” Whitley explained. “Of course, that’s a really hard question because you’re often talking about creating a new value ecosystem and that’s really difficult.

“We’re certainly not going to build massive teams and just say ‘Do this for the next two years and see what happens’. We’re going to experiment quickly change, review, experiment, and develop that work.”

Ultimately the success of robotic tools will depend on being able to bring together viewpoints not just from academia and startups but from the field.

“We’ve got go after the use cases, and I’ve already started working with our field engineering colleagues to really try and put some of this technology through its paces. And I’m absolutely sure for some of these things they will say ‘Nice idea, but it’s just not going to work’ – but I’m equally sure that our field engineering colleagues will come up with other use cases as soon as they see the art of the possible. Our engineers are very, very innovative.”

With innovations in automation, both for telecoms software and for robotics, the impact upon the telecoms environment and its cost bases over the coming decade stand to be considerable.

The initial benefit will likely be felt by major network infrastructure owners, improving the valuations of infracos by reducing operational costs. However, the same forces will also simplify the act of deploying and managing networks, lowering barriers to entry for new local and national players – all of which might make that predicted 50% rise in infrastructure spending much more impactful.

Photo by Derrick Brooks on Unsplash