Innovation Blog: More light under the sea

Innovation Blog: More light under the sea

Polina Bayvel FRS.jpg

Look, telecoms industry, can you slow down a bit? You’re growing so much that scientists and engineers are having a hard time keeping up, says Alan Burkitt-Gray

This industry is annoying. The world’s economy throws all sorts of things at it – a financial crash, a dotcom crash, a pandemic, crazy politicians, several wars – and it keeps on growing. In fact, it’s growing at a rate that makes it hard for carriers and their technology suppliers to keep up with demand.

Fibre – the material running under the sea and into our homes that has powered the digital revolution for the past 25 years – is running out of the ability to meet the demands we’re putting on it.

A TeleGeography report published in September said that global internet bandwidth will have risen by 28% in 2022, reaching 997Tbps.

And this year isn’t exceptional. Over the past four years annual growth has averaged 29%, including a spell before the pandemic.

These figures come from TeleGeography’s latest Global Internet Geography research, which tracks the return to normal following the pandemic-generated increase in data demand in 2020. Africa experienced the fastest growth in international internet bandwidth, increasing at an average of 44% a year between 2018 and 2022. Asia follows, rising at an average of 35% a year during the same period.

Another report from Ericsson, released in June, said that total global mobile data traffic will quadruple in the next five years and reach 282 exabytes a month.

Mobile data traffic was 67EB a month at the end of 2021, and will grow by a factor of 4.2 over the next five years.

Including fixed wireless access (FWA), which added 17EB a month in late 2021, will take the total traffic carried by wireless technology to 84EB a month.

One of the factors helping the rise of data is the rapid growth of 5G. Ericsson’s report says there’ll be more than 1 billion mobile subscriptions this year.

“Oh, it’s wireless,” I hear you say. “So, there’s lots of radio spectrum, isn’t there?” Well, no, there isn’t. And every base station or street corner nanocell needs backhauling with fibre. And then there are the onward connections between data centres to handle all that material being sent between cities and under the sea.

Meanwhile, the arrival every decade of a new generation of mobile wireless technology continues. Much of the world has 4G, some of us have 5G, but the bright sparks in the labs of our universities and equipment companies are on the way to specifying 6G.

Each G brings another upward twist in usage, and makes another set of applications possible that weren’t before. Hence, the need for more fibre will continue.

The big cable-laying companies, major infrastructure investors and international carriers are working as fast as they can to satisfy demand. But there’s a shortage of fibre cable and ships to lay it, although there is no shortage of financing yet, even though margins are tight.

The intriguing thing about fibre is that we use only a small fraction of its capacity: around 10%, suggests Polina Bayvel FRS (pictured).

Bayvel is one of the world’s most qualified people to talk about sending laser light through fibres. She’s professor of optical communications and networks at University College London, and earlier worked at Alcatel Submarine Networks in Greenwich, back when it was STC.

In 2016, the Royal Society recognised her expertise by electing her as one of its 1,700 fellows – a title granted to just 198 women since the Society was founded in 1660.

In September, Bayvel spoke at SubOptic Foundation’s Wave conference about fibre’s limits, mainly those under the sea. She said that the main problem is that current designers of subsea fibre limit themselves to using bandwidths of about 10THz.

Fibre’s limiting factor isn’t just attenuation in the glass fibre, but the performance of devices at each end – specifically, how much of the spectrum they make use of.

The middle (yellow) part of the spectrum is best at transmitting data, but devices could be made to operate at the far red and violet ends of the spectrum.

So, there’s a shopping list for device and cable makers, including new optical modulators and improving signal-to-noise ratios. Though more bandwidth means more electricity will need to be fed down a cable’s core to drive the amplifiers that boost fading signals for each legs of their journey.

And all of that needs to be translated into network design to make the future’s new, more powerful, subsea networks.

Subsea carriers have already done a great job in improving their cables’ capacities from just a couple of pairs per cable in 2000 to 24 pairs now. But, while 24 fibre pairs can carry half a petabyte of data, they need power.

And there’s a limit to how much electricity can be fed along a 7,000km cable.

Remember, the first transatlantic cable, laid in 1858 between Ireland and Newfoundland, failed after a few days because engineers tried to improve its performance by increasing the voltage running through it.

So, will we get there? Probably, although it’s hard to say how. In early September, NTT said it had increased the capacity of optical devices by 50%, so fibre could carry 1.2Tbps of data, while using 10% of the power current commercial systems do.

Even Bayvel found it hard to work out from the report how NTT achieved this, but no doubt it’ll become clearer.

Meanwhile, back to your labs everyone. You’ve got a lot of work ahead of you.

Gift this article