3 ms·

5G is going exactly the opposite way to provide higher bandwidth, more cells with each less users. With satellite internet you're sharing the medium with every

5G is going exactly the opposite way to provide higher bandwidth, more cells with each less users. With satellite internet you're sharing the medium with everyone, and that doesn't scale well. Beamforming probably helps a lot, but I don't how accurate it can be on that distance.

Retric 2y ago
You can have essentially unlimited satellites the issue is the percentage of time each one spends over high density areas. Design for NYC density and 99% of satellites would be redundant at any given moment. The solution for increasing density is dropping costs so it’s viable for satellites to be idle 95% of the time. At least as a first approximation, there’s some tricks with how you setup orbits after the basic network is done.
- CorrectHorseBat 2y ago
  You're still sharing the medium (the sky), so at some point interference is going to be an issue. A quick search tells me the beams a satellite uses is measured in km, so you can only have one satellite serving several square km (you could have more channels of course).
  - Retric 2y ago
    Let’s be pessimistic and say it’s ~25km^2 with current designs. Surface area of the earth is 510,064,472 km² so you estimate limits things to ~20 million satellites X however they can slice up the available spectrum. IE essentially unlimited satellites. As of September 2024 they have 6,371 operational satellites and ~ 4 million customers globally.
    - CorrectHorseBat 2y ago
      How is that relevant? Adding satellites over the Pacific doesn't improve New York's service. The question is how dense of a population they can serve, not how many satellites Starlink could theoretically have.
      - Retric 2y ago
        The only way to add one satellite over NYC on average is to also add several satellites over the ocean and other low density parts of the earth. If you want low latency individual satellites must be in LEO which means they spend most of their time over water and low density bits of land. Which gets back to my original point where increasing the maximum density inherently reduces the average utilization of each satellite. There simply aren’t enough people living in Iowa etc to balance the east coast.
        CorrectHorseBat 2y ago
        Sure, but my point was that there's a maximum density you can get which won't be enough for NYC or even a much less dense city.
        Retric 2y ago
        The current network can handle 1 home per square km on average but averages over a very large area. So ~2,000 as many satellites and you can handle NYC’s population. It’s actually less than you expect because you can make use of satellites a hundred miles out to the sea, over the Hudson River, and even suburbs. And that’s before considering how few people would pick Starlink when they can use cheap fiber.
        CorrectHorseBat 2y ago
        Well no, that's my whole point. Because you're sharing the medium you'll get interference if multiple satellites service the same area. >And that’s before considering how few people would pick Starlink when they can use cheap fiber. Obviously we're talking about how dense an area Startlink can service on its own.
        Retric 2y ago
        > Because you're sharing the medium you'll get interference if multiple satellites service the same area I’m going to simplify because you seem to misunderstand what’s going on. Your eyes allow you to see a clear image of your surroundings because photons come from even slightly different angles aren’t interfering with each other. Starlink uses a phased array antenna to achieve a similar effect where satellites in different locations broadcasting on the same frequency can be clearly distinguished by two base stations physically next to each other. And similarly two different satellites can receive clear signals when two Starlink antennas are broadcasting when physically next to each other at the same instant. There’s physical limits and the phased array antenna in use are much worse than theoretically possible. But, the technology they are currently using really does scale vastly beyond what’s economical viable.
- Sparyjerry 2y ago
  Musk himself has said Starlink is not designed for cities and fiber is a much better option there. To supply say NYC, you would need also be supplying bandwidth to the entire world underneath the orbit that the satellites. That is a lot of wasted satellite time as it flies around the earth with zero users most of the time just for the brief moment while users spike to the scale of a city like NYC.
  - Retric 2y ago
    Yea, fiber + 5G + Starlink is more efficient than any of those technologies alone. It’s always worth remembering that efficiency and technical limitations are different things. We could absolutely blanket the US in 5G towers but the cost is vastly higher than the benefit.

Retric 2y ago

You can have essentially unlimited satellites the issue is the percentage of time each one spends over high density areas. Design for NYC density and 99% of satellites would be redundant at any given moment. The solution for increasing density is dropping costs so it’s viable for satellites to be idle 95% of the time. At least as a first approximation, there’s some tricks with how you setup orbits after the basic network is done.

CorrectHorseBat 2y ago
You're still sharing the medium (the sky), so at some point interference is going to be an issue. A quick search tells me the beams a satellite uses is measured in km, so you can only have one satellite serving several square km (you could have more channels of course).
- Retric 2y ago
  Let’s be pessimistic and say it’s ~25km^2 with current designs. Surface area of the earth is 510,064,472 km² so you estimate limits things to ~20 million satellites X however they can slice up the available spectrum. IE essentially unlimited satellites. As of September 2024 they have 6,371 operational satellites and ~ 4 million customers globally.
  - CorrectHorseBat 2y ago
    How is that relevant? Adding satellites over the Pacific doesn't improve New York's service. The question is how dense of a population they can serve, not how many satellites Starlink could theoretically have.
    - Retric 2y ago
      The only way to add one satellite over NYC on average is to also add several satellites over the ocean and other low density parts of the earth. If you want low latency individual satellites must be in LEO which means they spend most of their time over water and low density bits of land. Which gets back to my original point where increasing the maximum density inherently reduces the average utilization of each satellite. There simply aren’t enough people living in Iowa etc to balance the east coast.
      - CorrectHorseBat 2y ago
        Sure, but my point was that there's a maximum density you can get which won't be enough for NYC or even a much less dense city.
        Retric 2y ago
        The current network can handle 1 home per square km on average but averages over a very large area. So ~2,000 as many satellites and you can handle NYC’s population. It’s actually less than you expect because you can make use of satellites a hundred miles out to the sea, over the Hudson River, and even suburbs. And that’s before considering how few people would pick Starlink when they can use cheap fiber.
        CorrectHorseBat 2y ago
        Well no, that's my whole point. Because you're sharing the medium you'll get interference if multiple satellites service the same area. >And that’s before considering how few people would pick Starlink when they can use cheap fiber. Obviously we're talking about how dense an area Startlink can service on its own.
        Retric 2y ago
        > Because you're sharing the medium you'll get interference if multiple satellites service the same area I’m going to simplify because you seem to misunderstand what’s going on. Your eyes allow you to see a clear image of your surroundings because photons come from even slightly different angles aren’t interfering with each other. Starlink uses a phased array antenna to achieve a similar effect where satellites in different locations broadcasting on the same frequency can be clearly distinguished by two base stations physically next to each other. And similarly two different satellites can receive clear signals when two Starlink antennas are broadcasting when physically next to each other at the same instant. There’s physical limits and the phased array antenna in use are much worse than theoretically possible. But, the technology they are currently using really does scale vastly beyond what’s economical viable.
Sparyjerry 2y ago
Musk himself has said Starlink is not designed for cities and fiber is a much better option there. To supply say NYC, you would need also be supplying bandwidth to the entire world underneath the orbit that the satellites. That is a lot of wasted satellite time as it flies around the earth with zero users most of the time just for the brief moment while users spike to the scale of a city like NYC.
- Retric 2y ago
  Yea, fiber + 5G + Starlink is more efficient than any of those technologies alone. It’s always worth remembering that efficiency and technical limitations are different things. We could absolutely blanket the US in 5G towers but the cost is vastly higher than the benefit.