Using lasers to beat, and speed up, the clouds
Big Interview

Using lasers to beat, and speed up, the clouds

Vicki Warker Laser Light.jpg

A seven-year project to finance a bold move to link laser satellites with fibre networks is close to fruition, the team behind Laser Light tells Alan Burkitt-Gray

A new project to connect hyperscale data centres with a software-defined combination of fibre networks and laser-based satellites should be going ahead in the next few months.

A US company, Laser Light, says it is close to finalising the finance for the first software-defined network (SDN), that will build a global network out of established terrestrial and subsea infrastructure. But Laser Light also plans to order a fleet of 12 medium Earth orbit (MEO) satellites from Thales that will be part of the same SDN.

“We will be connecting to large data centres that can do compute at scale, moving data around the globe,” says Vicki Warker, chief commercial officer of the company. “We’re going to use the SDN routing to make our network cost-effective.” Warker has long experience in the satellite industry, at Intelsat and SES’s O3b, and in telecoms, for Sprint and – when it was a BT-MCI joint venture – Concert.

Data will be moved via satellite or by terrestrial or subsea cables, she adds — or a combination, as typically each territory will have a cluster of laser ground stations networked together to avoid cloud-cover.

That’s the biggest challenge of using laser light for satellite links: they don’t get through the clouds. But fortunately weather satellites keep track of those and Laser Light’s software can swap instantly between ground stations. “We’ll use the operating system to route around the weather,” says Warker.

Customers will pay by use and will be able to choose between fast delivery or lowest cost, she adds.

Warker works with Laser Light CEO Bob Brumley, who says by email: “Laser Light’s focus is not to connect the unconnected or to serve a particular geography or target vertical. It is to provide a global utility for the distribution and management of data pools at scale.”

Warker tells me: “As the majority of content is generated in North America and Europe, our build-out will start there as origin points for our global infrastructure.” However, “demand is strongest where connectivity at scale is weakest and we are following customer demand”, she says.

Africa and Middle East

“We are in active build discussions for points of presence (PoPs) in Africa and the Middle East now and have started similar efforts in Australia and Latin America. Asia will follow shortly.”

It’s clear that content delivery networks (CDNs) are the main target, with their desire to shift huge files around the planet to hyperscale data centres for local delivery.

“But there will also be a need for latency-sensitive applications that run off 5G,” says Warker, who lists uses in artificial intelligence (AI), biomedical applications and manufacturing.

Laser Light’s launch date, depending on the company’s signing of its financial backing, will be late 2021 or early 2022, she says. That will be for fibre-based services, with the optical satellites following 30-32 months later. The satellites won’t be ordered until the money is in place. But once in service, the satellites will be able to deliver data at speeds of 200Gbps over laser beams, she adds.

The technology is not “bleeding edge”, she says. “Lasers are new to the commercial market, but not to the military market.”

Brumley, the CEO, has military experience, according to his LinkedIn profile. He’s been a lawyer, and has worked as such in the US Department of Commerce. But at the same time, he spent 23 years in the US Marine Corps attached to a reserve unit specialising in aviation command and control.

And he has telecoms experience too: for example, heading corporate development for Deutsche Telekom in the Americas in the late 1990s. Later he spent three years as president and CEO of TerreStar Networks, a project to offer integrated satellite and terrestrial services.

Next-generation infrastructure

Two decades later, “Laser Light is pursuing a convergence strategy”, he says in an email before my interview with Warker, “where a broad variety of complimentary, next-generation data infrastructure is deployed — built or bought — under an elastic network operating system where each mitigates and exploits their respective strengths and weaknesses”.

It is “putting together proven technology, but for the first time together”, says Warker. At the heart is Laser Light’s own operating system, which will select the best route and swap between satellites and different fibre routes within milliseconds. “It’s a complete green-field SDN,” she says, and Laser Light has essential patents. “It gives us the advantage of being able to use the most appropriate route.” Not just for least cost, but also swapping between satellite and subsea or terrestrial fibre, she notes.

The green-field nature of the system “gives us a huge advantage”, says Warker. “It’s not built on top of existing IT. We’re going to use SDN routing to make our network cost-effective.”

She adds: “This is a pay-for-use service.” The service-level agreement will say, in effect, “get the data to this place either as soon as possible, or at the lowest cost, or as a back-up file. And all of it is pay-for-use. The added value to the user is they will pay for volume or for time commitments.”

Sitting in on the call with Warker is Steve Saunders, a British former journalist who is now vice-chairman of Laser Light’s board. He founded the telecoms news site and newsletter Light Reading and its associate, Heavy Reading, in 2000 and sold the business in 2005 to UK publisher UBM for $33 million. Then he bought it back in 2014 and sold it again, two years later, to another UK group, Informa, though the price isn’t recorded. But a couple of years later Informa spent £3.9 billion on UBM itself.

Distance-independent cloud

“This is the first global example of a massive enterprise cloud infrastructure deployment,” says Saunders. “I look at this as an inevitable future.” What’s inevitable? “A distance-independent cloud service,” he responds.

The Laser Light project started seven years ago, but it’s taken a long time to get it to this stage. “This year, we seem to have turned the corner,” says Saunders. There are “four major activities that are being lawyered”, but he ducks out of detailing them.

Warker says that “we have a number of agreements” and “we will be announcing them pretty shortly”.

I prod a bit harder. It seems clear to me that companies such as Netflix would be in line to be customers. Walker smiles over the Zoom link but doesn’t fall into the trap of naming potential customers, except to agree that major content providers have a “need to distribute data”. And that’s when she goes on to talk about 5G and IoT connections, biomedical data, and so on. Companies are using giant video files to detect manufacturing defects, she notes.

“The first layer” of users is “large content providers”, she agrees, adding: “Amazon and Microsoft both have cloud services for CDN [content data network] providers.” And she reminds me of a pact announced in 2017 to make data centre company Equinix its “strategic interconnection provider” for its new global network. Four years is a long time in this business, but Warker assures me the deal is still in operation. “Equinix has some of the best on-ramps in the business.”

Back in 2017, Brumley said: “A partnership with Equinix permits Laser Light to become a truly tier-one global carrier with access to facilities, and incumbent local fibre providers, in a ‘one stop shop’ partnership.”

Saunders jumps in on this year’s interview with Warker to note that government and the financial community also need to move giant files around, “and they need the security of the network”. Lasers can be aimed precisely between satellite and ground station, minimising risk of interception.

Search for clear skies

The number of ground stations in practical use will vary according to the local weather. “Australia has lots of clear skies,” says Warker. The UK will probably need about five.

Brumley shone more sunlight — if you’ll forgive the expression — on the subject last year in an interview with Capacity. “We can achieve five-nines in the UK,” he told me then. “The weather in the UK is highly predictable. At any given instant there will be a node open.” For example, if it’s raining in Newport, the south Wales town that is home to one of Europe’s largest data centre, it’ll probably be fine in Kent, in south-east England. “With weather, you try not to fight it. It clears behind.” With enough nodes you can achieve that five-nines standard, he said last year.

In the US, Warker notes, there would be about 20, to serve major markets, and they’d be co-located with a large data centre.

Scale is the key, says Brumley by email. “Chasing scale has always been the bane of infrastructure business cases. The often-asked question in the boardroom is when will the model breakout of the capex/opex cycle and achieve margins worthy of the investment?”

Now, he says, “we are seeing this dynamic play out in real and consequential terms in the LEO wars ongoing today”. When and how will SpaceX’s Starlink or OneWeb achieve scale, he asks.

“The ‘chase for scale’ is a key factor affecting not just satellites and but all terrestrial business cases alike — equally. By bringing infrastructure together which was funded for a single purpose, into use on a converged basis for multi-purposes, the ability to achieve scale sooner is real. Also, the cost to achieve and maintain scale is less.”

Gift this article