Re: Plugging satellites together
The problem is the speed of light.
Four big communication satellites in geostationary orbits can cover the whole world (except the poles). E.g. Inmarsat's original satellites follow this model. Unfortunately, geostationary satellites are a long way away from the Earth, so signals travelling at the speed of light take at least a quarter of a second to go from Earth to the satellite and back to Earth. For telephone conversations, or TCP/IP networking, that delay is very long and causes problems.
To reduce the delay, you bring the satellite closer to Earth. That reduces the area it can cover, and also means the satellite is travelling across the surface of the earth instead of appearing motionless. So you need lots of satellites to get coverage.
With geostationary satellites, a missing satellite means that a fixed area of Earth doesn't get any coverage, which is acceptable - you can just sell your service to the people who have coverage, and that's how Inmarsat started with a single satellite. But because these closer satellites are moving, so the area of no coverage moves, so everyone everywhere complains about the regular dropouts. So with these new constellations, they will have to launch all of the satellites to get global coverage all at the start, plus they need some in-orbit spares so failed satellites can be replaced quickly.
Also, the smaller the spot that's covered the higher the bandwidth can be. Simplistically, if each satellite can do 1 gigabit/sec, then 50 satellites gives you a total of 50x the bandwidth compared to 1 satellite. (In reality it's not that simple, the gain isn't that big, but the general point stands). You can also reuse radio frequencies on satellites that are "far enough" apart, the same way that mobile base stations reuse frequencies. So 50 satellites can be significantly faster than 1 satellite, with the same frequency allocation.