[SEJeff]

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.

[adamkaz]

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.

[guelo]

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.