Last week BT issued a press release about its trials of hollow core fibre at its labs in Adastral Park, Ipswich. This project used a 10-km hollow core fibre from Lumenisity, a Southampton University spin-out company, and Open Radio Access Network (O-RAN) mobile vendor Mavenir. The essential message from the press release is that glass slows the movement of light compared to its speed in air, so that lower latencies can be achieved by sending light through air-filled fibre rather than solid glass. 6GWorld™ spoke to Professor Andrew Lord, BT’s Head of Optical Research, to dig behind the headlines.
The press release highlights up to a 50% reduction in latency through hollow fibre and outlines a few use cases. Can you give us a sense of what the main commercial driver is for cutting latency in the fibre network?
There are various use cases for this hollow core, not just the latency one. The latency one is a quick win because the other parameters are not quite there yet – they will be, but latency is already there because there’s no glass in the middle. That’s not something that’s going to improve, it’s just there.
For us, latency isn’t really a big driver in the optical network. The time taken to come and go across the optical network is often in the noise compared to the processing time you have at the ends. You don’t need to do much better there.
The one use case that we really latched onto was in the RAN space where there is, in 5G, a latency ”budget” assigned to the RAN which you shouldn’t exceed because services don’t work as well. That budget is taken up in the equipment and the fibre. If we can do better in the fibre, that means there’s more latency budget for the equipment or you can move your equipment further away because a longer length of fibre will have the same latency. That means you can consolidate your resources and have one building that serves multiple radio heads. That’s a cost benefit for sure, so that’s the thinking that caused us to build this test network with Mavenir.
You mention that there are a variety of potential use cases beyond latency. Can you give me an idea of what other opportunities this would open up?
Fibres are currently focused on a core band of wavelengths in the 1.5 micron window, and that’s getting full. There’s a lot of work going on around the world, including in my lab, to try and expand that. It’s very much like what’s happening in the radio spectrum.
Fibre is very transparent, so you could potentially have a much broader range of wavelengths, but a variety of technologies play into that and they all have to be developed. So, there’s a whole global research challenge at the moment to ask “What happens beyond the C-Band? Where do we go next?” Hollow core could really play into that, because there’s no glass.
When you put potentially hundreds and hundreds of wavelengths into the glass at a very high power level of a watt or maybe more – that doesn’t sound high but in glass it’s very highly concentrated into this core – the glass starts to complain and stops working, and in the worst case it starts to melt. That’s a worst case, but before that happens it generates non-linear effects and messes up your signals. The signals all start interacting with each other. I have people on my team focused on trying to solve this problem but it’s hard. With hollow core fibre it’s not difficult because none of these wavelengths interact, because there’s no glass. Is hollow core a route to future very, very high-bandwidth systems? That would be one use case which I think would be really exciting.
So, unlike with current fibre-optic cables, you wouldn’t be sending light through the glass outer at all?
No, not at all. The glass is there as a very, very fancy mirror. I mean incredibly fancy. The glass is there to guide, but don’t just think of it as a cylinder; It’s a lot more than that. Essentially, though, the glass is there to keep the light in the core.
The holy grail is to get the attenuation of the fibre down. The attenuation at the moment is not as good as standard fibre, but it’s on a trajectory to get there. That attenuation is driven by how good these mirrors are. What Lumenisity is really trying to do is perfect these really crazy mirrors, because that’s their route to unlocking excellent attenuation.
But the idea here is really to make that core take as much capacity as possible. Because it’s empty air you can fill it with a wide range of wavelengths, a wide range of optical powers, and it just works. With standard glass you have to think hard about “Is this wavelength going to work? Am I sending at too high a power, because the glass might react?” Where there’s no glass you don’t get any of those problems.
Does using a different kind of fibre require any difference in the way that signals are managed or treated at either end?
Yes and no. Firstly, you can plug whatever equipment you have on standard fibre into this hollow core and it will still work. I don’t think we’d be engaging otherwise – if you have to change your transceivers you’re going to be struggling to get any traction because the transceiver market is massive.
So these hollow cores just have to plug in… but, because there’s no glass, there are no bad effects of glass. A lot of today’s transceivers have been developed to cope with or counteract those bad effects, so maybe you don’t need such complicated transceivers to manage hollow core fibres. Maybe you can use cheaper transceivers. You might take a transceiver which is currently designed to do a kilometre or a few hundred metres and now it might work over longer distances because the fibre is so benign.
We haven’t tried that, but it’s definitely something we’re looking at – is this fibre going to unlock a cheaper endpoint? In which case the fibre may be more expensive, but the system as a whole won’t be.
We’ve talked a bit about how the opportunity for pushing new bandwidths and higher powers through a hollow fibre intersects with your research on expanding into new parts of the spectrum. Are there other elements of your research that this intersects with?
We’ve been publishing a lot recently about quantum key distribution and quantum security, but, again, hollow core could be really good for this use case. There’s no glass, so there’s no interaction between quantum signals and anything else. The quantum channel can stay unaffected by anything else that’s in the fibre at the same time. So we’ve been pondering and doing some experiments – which we will publish shortly – looking at the benefits there as well.
If there are many benefits to transmitting through air rather than glass, could you replace fibre in some cases with point-to-point optical communications through the air?
Yes; in fact, I’m leading an Innovate UK project on exactly this at the moment. It’s called AIRQKD, where we have a bunch of industrial partners and universities doing exactly what you’ve just said, because you don’t always want to install a fibre. There are situations where this would be a much quicker solution. Also free-space comms from satellite, where the satellite is not sending radio signals but optical signals. It’s actually sending a laser beam focused on a ground receiver and sending single photons for secure comms that way.
Unfortunately, as soon as you go away from radio to optics, alignment becomes much more challenging. With radio signals you just need to point your radio antenna towards roughly the right direction and it will work. That’s not the case with optics. I don’t know if you’ve used a telescope, but it’s very hard to line anything up. So the research is all about how to build alignment systems. You can use GPS so you know roughly where you are and where the other end is. If you can get close using basic systems like that, then can you do some active alignment to do better? It’s tough, but I’m hopeful.
What is the future for manufacturing hollow core fibre, then, and do you see a timeline for its adoption?
For actually making the fibre – the industry already knows how to do this. It’s a different fibre, but the techniques are the same, where you make a pre-form and pull it. The materials are the same. It’s just a different structure. The way you do strength testing, the way you cable the fibre – all of that we’ve learnt to do over decades. The industry knows how to do this. So I really hope to see a quick ramping-up to volume.
The other answer to your question is a question of motivation, really, from the fibre manufacturers. How do they see this? Do they want to sell this as a premium product to high-frequency traders, or do they have a mission to make this available for wider markets?
To be honest, that’s partly why I’m engaging – trying to encourage the industry to take this as far as they can, because I can see such benefits in the wide-scale market, not just these niches. It would be amazing if this could have a bigger foothold. It’s completely possible and there’s nothing to stop it technically. It just needs that motivation.
