[mozumder]

The developing world needs fiber instead of satellite. This isn't going to change much over their current LTE speeds. This type of satellite internet is far too expensive far far too limited bandwidth to be useful for them. The energy cost of RF broadband is so much higher than fiber.

Really, the audience for this are ships and other remote regions, not anyplace that has a road going to it.

[Robotbeat]

Nah, You’re thinking of different constellations like OneWeb. This constellation is being way overbuilt by SpaceX. They want gigabit speeds and way cheaper data than LTE. They want to compete with fiber (at least in non-super-dense areas). From Spacex’s perspective, they’re happy to build it so large that it only breaks even (even though I think it’ll do better than that) because it gives them something to launch on their huge, fully reusable rocket that they’re building.

[mozumder]

Nah, they can want infinite speeds, but they're only going to get a few megabits of speed.

It will be much slower than LTE, since the satellites are going to be about 100 miles away. LTE towers are only a few miles away. How do you expect it to have higher bandwidth than LTE with communications a hundred miles away? It doesn't matter that they're putting in 4000 satellites, since those satellites are still 100 miles away from the terminal.

Again, I want everyone to be clear: This is going to be MUCH slower and more expensive than LTE.

Like Iridium, this project is only useful for remote areas that can't be reached by roads. Very few people will benefit from this project. It won't change society.

This is another one of Elon Musk's garbage projects that doesn't really advance the state-of-the-art.

[lorenzhs]

Latency and bandwidth are two separate things, and the distance to an LTE tower has approximately nothing to do with latency to the server you’re connecting to, which btw is likely much further away than 100 miles. Now factor in that light moves 50% faster in air or vacuum than it does in fibre optic cables. The satellites are also going to be at a height of around 1000km (600 miles). But I t takes light only about 3.3ms to travel that distance. 30ms latency end-to-end should be realistically achievable this way. That’s more than a good residential connection but good enough for just about any application other than latency-sensitive multiplayer games.

There’s valid skepticism of SpaceX’s plan, but none of your points are among it.

[mozumder]

This is a discussion about low bandwidth. Why are you talking about latency?

[Robotbeat]

Because otherwise your comment makes no sense. You can already get satellite data rates (100Mbps on a home plan) equal to or greater than LTE, from ~100 times further (36,000 km altitude Geosynchronous orbit) and the distance itself does not fundamentally limit the speed, but the latency there is terrible and is fundamentally limited. LEO satellite constellations address that.

SpaceX's particularly large constellation (allowing streaming from multiple satellites at a time and with large phased array antennae on both ends) will allow much higher bandwidth than existing satellite internet systems, but the biggest deal about these LEO satellite constellations over GSO satellites is the far lower latency.

[Robotbeat]

How? Because both the satellites and the receivers will use large phased array (beam forming) antennas with hundreds to thousands of elements unlike LTE. No one will have deployed phased arrays to users at the scale of this constellation, therefore SpaceX has invested a lot in phased-array/beamforming technology.

“Another one of Elon Musk’s garbage projects that doesn’t really advance the state-of-the-art.”

Okay, so you’re just trolling, then. Never mind.

[mozumder]

OK but now you've limited your applications to non-mobile devices, unlike LTE.

And if they're targeting this for cellular base-stations, they're now on the hook to provide an even larger aggregate bandwidth to share among hundreds or thousands of local users per tower.

[Robotbeat]

Correct, the entire point of SpaceX's constellation is non-mobile. Always has been.