Elon Musk’s satellite plan: Project Loon without helium or latency

Elon Musk dropped a bomb from near-earth orbit on Friday at an event in Seattle: Instead of working with mini-satellite startup OneWeb to build an internet network in the heavens (as was widely expected) Musk told Bloomberg he plans on creating a globe-spanning constellation of his own, launching hundreds, if not thousands, of interconnected satellites each weighing as much as a Vespa.

This is the kind of bold plan we’ve come to expect from Musk, but unlike his past grand-scale projects, the idea for this one isn’t entirely new. The most obvious example is [company]OneWeb[/company]’s planned constellation of 648 satellites. Formerly know as WorldVu, the company was founded by ex-Google satellite chief Greg Wyler and has backing from [company]Virgin[/company] and [company]Qualcomm[/company].

But there are also already satellite constellations in the sky supplying internet access to any point on Earth, most notably Iridium and Globalstar’s networks, though neither one is offering what you would consider broadband speeds. And if we’re looking to make comparisons to other internet projects out there, we need look no further than [company]Google[/company].

The voyage of Loon balloon I-167 as it circumnavigates the globe (source: Google)

The voyage of Loon balloon I-167 as it circumnavigates the globe (source: Google)

When a balloon looks like a satellite

There are surprising similarities between Project Loon and Musk’s proposed SpaceX network as well as OneWeb. The two projects not only appear to share the goals – to connect the farthest corners of the Earth with low-cost internet – but the basic architectures of the networks would be the same.

Google is building a vast network of balloons that surf the stratospheric winds 12 miles up in loosely defined latitudinal orbits around the world. Those balloons use a radio broadband link to connect to transmitters on the ground and mesh networking techniques to link to the other balloons on the horizon, creating a kind of floating internet in the sky. Data is passed from balloon to balloon until it’s within site of a ground receiver, which offloads that data into the internet proper.

Musk’s plan calls for essentially the same scheme, just 740 miles higher up. The original talk of 700 orbiters has now turned into plans for a a constellation with as many as 4,00o satellites. In low-earth orbit, those satellites would be skimming the top of the Earth’s atmosphere, 30 times closer to the surface (and your PC or smartphone) than the geostationary satellites that today carry the bulk of our orbital internet traffic. They won’t be floating like Loon Balloons, but those satellites are still bound by the laws of physics. At that altitude, they’ll need to travel at 16,000 mph and would complete a full orbit of the Earth in a little less than two hours – otherwise they’d fall out of the sky.

That means from a vantage point on Earth these birds will be whizzing overhead. So as with Loon, an Earth-based transmitter won’t be connecting to a single Musk-built orbiter, but multiple. Both balloon and satellite would pass your connection on to the next balloon or satellite as they pass overhead. Data would then flow from balloon to balloon and from satellite to satellite until they found their appropriate ground-based links.

An Iridium flare

An Iridium flare

If you want to get a more concrete of visual, you need only look to the night sky (with a little help from this website). [company]Iridium[/company]’s network of satellites have highly reflective antennas, which produce “flares” when they reflect the Sun’s light, making them resemble shooting stars.

A new spin on the orbital constellation

As those Iridium flares readily demonstrate, there’s already plenty of hardware in the heavens dedicated to providing global internet access. What will Musk or OneWeb’s constellations do that Iridium or [company]Globstar[/company]’s won’t? Or for that matter what Project Loon or other sky-bound internet projects like Facebook’s drones?

While Iridium and Globalstar may have pioneered the globe-spanning internet constellation, they also have limited number of satellites in the sky (66 for Iridium, 32 for Globalstar). Putting more birds in orbit is the equivalent of adding more towers to an urban cellular network: fewer people are connecting to the same cells so every user can tap faster speeds and there’s more overall capacity throughout the entire system.

An Iridium Next satellite

An Iridium Next satellite

Iridium and Globalstar are also focused on providing mobile internet connectivity from satellite phones and modems to a network far above. That’s very useful for leaving a GPS breadcrumb trail for a lost airplane or maintaining contact with dog sleds racing in the Iditarod, but Musk appears to have more stationary transmitters in mind. A high-power antenna aimed at a satellite can produce a lot higher data speeds than one you carry in your backpack.

And while Globalstar and Iridium may have had cutting edge technology at one point, they leave a lot to be desired today. Iridium’s current network is slower than a dial-up modem, and the new Iridium Next network Iridium is launching into space starting this year – ironically on the back of SpaceX’s Falcon 9 — will support a 15 Mbps to a stationary dish antenna.

It ain’t easy running satellites

It takes a long time to plan, design, build and deploy a satellite network. The birds that Iridium is rolling onto the launch pad this year were designed several years ago, and the network won’t be fully operational until 2017 when the last of 66 orbiters are in place.

Musk will be working with newer technology (The Information has it that SpaceX is weighing using optical lasers instead of radio spectrum), and being Elon Musk, he’ll likely be facing a much shorter development timeline (it helps to have the resources of SpaceX at your disposal). But once he gets those birds in space, he’ll likely face many of the same hurdles as other satellite providers. His technology will be frozen in time. These satellites aren’t exactly easy to fix if they break, and upgrading a satellite usually mean sending to fiery death in the Earth’s atmosphere and replacing it with a new one.

A look at the various satellite orbits. Musk's proposed network would sit in low-earth orbit.

A look at the various satellite orbits. Musk’s proposed network would sit in low-earth orbit.

One of the advantages to Project Loon is that Google’s network will be much more accessible. A Loon balloon will circumnavigate the world three times before coming down for regularly scheduled maintenance. Even [company]Facebook[/company]’s drones can be flown down for repairs and upgrades.

But Musk seems to be counting on his orbital network doing something those atmosphere-hugging projects can’t: create a better, faster internet. I’m not talking about speed here, but latency – the delay data undergoes when traversing the globe. When connecting in San Jose to a server in Sydney, your request is hitting multiple routers before it even arrives at the undersea cable to begin its long journey across the Pacific, and all of those steps introduce latency.

Musk claims he can build a purer, simpler internet in the heavens. Though any traffic would have to got through the Earth’s atmosphere twice, once that data stream is 750 miles up, it would make only a few satellite hops across a near vacuum, through which electromagnetic waves travel much faster than through a fiber optic cable. So what Musk is promising to do is not only build an internet to connect the furthest corners of the planet, but a create a network that would draw those far corners much closer together.

There’s enough of a difference between Loon and Musk’s plan, that Google may view them as complimentary technologies, and according to the Information’s report, Google is considering investing in the SpaceX project.

This post was updated on Jan. 21 to note that the size of SpaceX’s planned constellation has grown from 700 to 4000 satellites.