Race Is on to Deliver Internet from Space | Hacker News Reader

Race Is on to Deliver Internet from Space

wsj.com

26 pointsgwallens31 comments

Slow race. Motorola tried this once already and failed with the Iridium Constellation in the 90s.

https://en.wikipedia.org/wiki/Iridium_satellite_constellatio...

The venture went bankrupt and Motorola was actually on the verge of burning them up in the atmosphere when the Pentagon stepped in: http://abcnews.go.com/Technology/story?id=119255

Another interesting thing about the Iridium Constellation is that the satellites have an unusual arrangement of large, flat antennae. This, along with detailed data about their position and orientation means that they create very predictable flares of reflected sunlight that are quite visible from Earth. You can usually a fairly bright flare every couple days:

http://heavens-above.com/IridiumFlares.aspx

Oh, right, it was tried before.

Hey, remember the Newton? It's kind of a shame we'll never see an improvement on tablet computers since that failure proved it wasn't a viable idea.

The costs of building and launching small satellites has dropped significantly since the 90s. People are also traveling more and demanding more connectivity, so it seems that Iridium may have just been a bit to early to the game.

Early, or late. One of Iridium's problems was that by the time they actually became operational, a huge chunk of their potential customer base had been eaten up by terrestrial cellular phones. The number of people who needed something more than a cell phone was a lot smaller in 1998 than it was in 1987.

Iridium service was also unbelievably expensive and not all that useful. (For that matter, still is.) If you had a choice between Iridium and a cell phone, the cell phone won every time. Iridium was a product for those with no other choice. If they had come in earlier, they might have had enough people with no other choice to make it. Later, and they might have been able to offer something that could actually compete with terrestrial services. As it was, their timing ended up being really bad.

Interestingly, they're still popular with some very adventurous groups, e.g. Round-the-world sailors, motorcyclists going through remote locations, climbers and hikers in remote places / foreign countries, etc.

Very useful if it's the only thing that will work.

Exactly. For some cases it's an amazing service. It just wasn't enough to sustain a company that had to spend $5 billion up front to get things started. Once that got discharged in bankruptcy, it managed to do OK.

I've used Iridium sat phones in Iraq, and am currently prepping for deployment to Antarctica, where we'll have Iridium phones as backup to a satellite data link.

Slow being the operative word.

Iridium failed at the Internet partly due to being slow as molasses. The data rate is 2400 bits per second.

Iridium wasn't intended for internet -- it was intended for voice only. And by the time they got enough birds in the air, most areas were covered by at least one voice provider, so the only potential market was people that traveled frequently to remote areas. Not to mention that you could never have a handset that fit in your pocket.

Now the way I thought Iridium should have been constructed, is to have the uplink be a portable micro cell unit, that regular phones could roam to. That way you would only need one Iridium phone, maybe located outside, and others could use their existing phones to connect to it.

FWIW this is kind of what they do on e.g. racing sailboats offshore; a central uplink hooks into the satellites and distributes internet onboard via wifi. It's slow and expensive, but let's you do things like pull weather reports from the middle of the Atlantic, post your progress back to race control and your family / friends, etc.

Has anyone figured out how to deal with the issue of latency for space-based internet? Things like online games and video calls will just not work if the traffic hits space. I could see an ISP offering a service where traffic that needs low latency gets routed on land, while web pages and the like get routed through space. At a cost, of course.

All of the past versions of "space internet" have been geostationary orbit, which is around 25,000-26,000 miles above sea level. The new iterations, such as SpaceX's idea, is to use many many more microsats and put them in LEO. Low Earth Orbit is closer to 750 or so miles above sea level. Since the ping latency of geostationary sat internet is around 500-700ms, it would be massively less for low earth orbit.

The two technical issues I'm aware of are the sheer number of satellites required for LEO internet, and the fact that you can't point your dish at a single place. There would need to be some sort of actuator or omnidirectional receiver for tracking to satellites at every client site. This makes the installation a bit trickier, but the ping times should be entirely reasonable provided someone gets the funding to put hundreds of satellites into an internet constellation.

As someone who has worked in this industry and seen a lot of these ideas fail first hand - another big challenge is going to be ground entry points. For GEO satellites, each satellite serves many customers and may only need maybe 2-3 (for redundancy) groundstations.

For LEO constellations, each satellite can only see a small portion of customers at any time and will quickly move out of coverage of a single point on the earth, requiring many groundstations.

Alternatively, the satellites can crosslink and eventually hit a groundstation, but these handoffs and trip lengths quickly get back to the ping latencies of making a single trip to GEO.

The article reads almost like a press release for ViaSat which uses a handful of big geostationary satellites. It even disparages the OneWeb LEO initiative saying it will be too expensive and take too long to build. I think the article was prompted by the dispute between American and Gogo. I'm not sure what it's doing here on HN, doesn't seem that interesting.

This actually exists in some countries. E.g. in Europe there are (were?) several companies selling modem/ISDN+satellite bundles. All the upstream and parts of the downstream would travel through land-line while larger downloads would be routed via the satellite. This also saves you from needing to have relatively expensive satellite upstream equipment.

Otherwise, with GEO/GSO satellites there is no way to get below ~280ms round-trip times.

The new services are set to use satellites in low Earth orbit. That only adds a few hundred miles to the distance your data has to travel. The added latency is measurable, but not very high.

The problem with existing high-speed satellite internet is that the satellites are all in geosynchronous orbit, which is about 22,000 miles up. For a roundtrip, your data has to go up to the satellite, down to the ground station, then back up to the satellite and back down to you, for a total of almost 100,000 miles, or more than half a second at the speed of light.

Games and video calls, no, but it should work just fine for things like watching streaming movies.

What I think they should do (not sure of what this is called ) is UDP-broadcast "caches" of popular sites such as reddit, (hacker news :) ) wsj, etc, continuously.

I'm willing to bet that a very large portion of people only access a relatively small number of sites. If they streamed a cache this way, it could probably cut their bandwidth substantially.

multicast

ViaSat works great. Yes there is actual packet latency but practically speaking voip/videochat just works. Download speeds are fast, absolutely no problem watching streaming HD.

The biggest problem is that you get a fixed amount of traffic. They are starting to come around to throttling once you go over, but so far not for everyone.

Edit: need to amend - As a technology it works great. - I use it at a remote cabin and it's fine for weekend use. - Getting the Internet in middle of nowhere, yes+++. - If I had to live with it for day to day use I would not be happy because of the traffic limits.

tl;dr geo sync internet without traffic limits would get the customer 99% where they wanna be.

Any thoughts on how to get around the paywall?

The google search trick isn't working for me.

http://canmua.net/world/race-is-on-to-deliver-internet-from-...

I wonder whether China would consider any internet-supplying geosynchronous satellites over their borders as an affront to their sovereignty. Satellite internet (without a land-based backhaul) is surely the most effective way of subverting the Great Firewall.

They will probably be able to convince the service providers not to offer service within China. Flying the satellites over the country is fine, various international treaties take care of that, but once you start actually doing business in the country (even at a remove) then you'll run into trouble. It would be interesting for an outside company with no ties to China to just say "screw it" and ignore Chinese demands for restrictions or bans, but no company large enough to provide this service would start a fight like that in the first place.

You can see this at work already with existing services. Random example, check out the satellite coverage map here:

https://www.united.com/web/en-US/content/travel/inflight/wif...

Note the distinctly China-shaped hole in East Asia.

> Note the distinctly China-shaped hole in East Asia.

Quote: "Please note: Satellite coverage may experience outages for reasons such as government regulations"

I guess this is intentional because China may retaliate against violators (e.g. prohibit fly-overs).

People would still have to subscribe to the service, and I'm sure China could control their population's purchases to prevent them from paying for it. They could also use detectors (or jammers) and threaten their population enough to keep them from utilizing the service.

> People would still have to subscribe to the service, and I'm sure China could control their population's purchases to prevent them from paying for it.

China is iirc (one of) the biggest Bitcoin usage countries.

> They could also use detectors (or jammers)

Jamming doesn't work with a proper directional dish, and detecting who in the country surfs via satellite doesn't work if there is proper encryption on the air protocol.

Geosynchronous wouldn't be over their borders. They fly around the equator.

Geosynchronous satellites don't have to be over the equator. Geosynchronous satellites that are over the equator are called geostationary.

https://en.wikipedia.org/wiki/Geosynchronous_satellite#Defin...

Interesting. Til I guess. But what's the purpose of not being geostationary?

Covering areas far from the equator ?

Reply on news.ycombinator.com