4 ms·

(This is all from memory as a fan, so take it all with a grain of salt and please correct me if i'm wrong on anything!) So Starlink (the name of the network Sp

(This is all from memory as a fan, so take it all with a grain of salt and please correct me if i'm wrong on anything!) So Starlink (the name of the network SpaceX plans to build) is different than most other satellite internet projects. While most try to put their sats in GEO orbit (which is about 35,000 km above the earth), Starlink is going to LEO (their sats are planned to be from roughly 350 km to 1,500 km above the earth, those numbers aren't typos). Starlink is also planned to have multiple orders of magnitude more sats than any other network (mostly due to the lower orbit, since lower orbits must move "faster" relative to your position on the ground, each individual sat may only be within signal range for a short amount of time). The latency from a "normal" GEO internet sat is on the order of a second or so for a round trip, the latency for a Starlink sat is expected to be on the order of 100ms for a round trip! They even have some plans which show that it can have SHORTER latency in some cases than ground-based communications! Throughput is still yet to be shown entirely, but with such a large number of sats which all communicate between the others around them at any given time, and some clever routing algorithms, they should be able to get a worthwhile amount of throughput from the system. As for future-proof-ness or updates/upgrades, the path is pretty elegant if not a little crazy sounding. Their plan is to just let them deorbit on their own after some time (I believe on the order of a year or so), and constantly replace them with new sats by leaning on the cheap launches that their rocket tech provides. Since the sats are in such a low orbit, the natural drag of the atmosphere will slowly pull them down over time, and I believe they also have self-deorbit capabilities. This solves the issue of upgrades (don't upgrade them individually, upgrade them as a group over time!), and future-proofing it with those upgrades (as the tech improves, so will the constellation over time!) Hopefully this wasn't too rambling and difficult to understand!

brandmeyer 8y ago
Their are some downside risks to this endeavor. For starters, there is a low utilization factor for each satellite. Humanity is concentrated onto a fairly small percentage of the Earth's surface. Earth's radius is about 6400 km, so at that low altitude you can only get so much out of pivoting your antenna's towards populated areas. Each satellite will be spending much of its time without any users (oceans and deserts), or with very few users (rural areas). Second, there is a belief that the next billion new Internet users are going to be mobile users. So the market size of satellite Internet isn't "all the people that don't currently have Internet," its much smaller than that. In the developed world, Internet access is frequently bundled with other services, so the potential pool of flexible customers that could be wooed away from their current service isn't that large, either. Finally, eventually the folks digging ditches are going to catch up. Once fiber gets to your premises, it will be much cheaper ber byte than satellite Internet. It doesn't matter how the last mile is traversed. Even if its fiber to an LTE tower, it'll still be cheaper per customer than satellites. I could be wrong about the economics. Presumably they've done the numbers, and the plan still closes to profitability. SpaceX isn't the only one perusing it, after all. Nevertheless, I think there is also reason for skepticism.
- Klathmon 8y ago
  >there is a low utilization factor for each satellite So we are getting to the limits of my knowledge here, but I believe that this won't be as big of an issue simply because sats aren't JUST communicating with the ground, they are communicating with each other as well. The routing system will take sub-optimal paths (from what you'd expect) to ensure speeds stay as high as possible and latency as low as possible (with i'm guessing some tweaking of those 2 values when near the limit), and a lot of those are going to include crossing those large oceans and deserts which will keep those sats busy, and will even route toward those large underutilized areas to spread the load out. >Second, there is a belief that the next billion new Internet users are going to be mobile users I think this is where the concept of base stations really shine. Rather than pulling fiber to all of your cell towers, you can get those "fiber quality" speeds by just having power and a place to put the tower. This means propping up new "traditional" cell towers is much cheaper (for customers of Starlink anyway!) and those mobile users could be using Starlink without even knowing it. >In the developed world, Internet access is frequently bundled with other services, so the potential pool of flexible customers that could be wooed away from their current service isn't that large, either. This is entirely anecdotal, but I'm 28 years old, and out of all of my friends and family, only my Father has cable. TV over the internet, and streaming services are taking over, and public sentiment isn't exactly high for the traditional internet service providers. Many (at least in the US) feel they don't have a choice, and if Starlink ends up being as well loved by the public at large as SpaceX or Tesla is, then I have a feeling not only will people be willing to switch, but they will jump at the chance! Though a lot of that is assuming that tesla targets most home users(which I'm not sure if they are). >Once fiber gets to your premises, it will be much cheaper ber byte than satellite Internet. It may be cheaper, but it might not be as low latency! Turns out Starlink might be able to get lower latencies across continents than traditional fiber can currently achieve. Obviously it's still up in the air if they will be able to deliver on these numbers, but if they can then there might be a small portion of users who will be willing to pay top dollar for reliable low latency connections. But taking off my fan hat, I agree that there are still a lot of unanswered questions about starlink, and if it ends up like most of SpaceX's other promises, it will be downgraded at least once in "performance", end up about 10% more expensive than expected, and a few years late, but still amazing. It may not be the next comcast powering a huge portion of the "last mile" of the internet as we know it, but I'm hopeful it will be a great alternative that can improve things for a significant number of people, and hopefully pays off well for SpaceX.

brandmeyer 8y ago

Their are some downside risks to this endeavor. For starters, there is a low utilization factor for each satellite. Humanity is concentrated onto a fairly small percentage of the Earth's surface. Earth's radius is about 6400 km, so at that low altitude you can only get so much out of pivoting your antenna's towards populated areas. Each satellite will be spending much of its time without any users (oceans and deserts), or with very few users (rural areas). Second, there is a belief that the next billion new Internet users are going to be mobile users. So the market size of satellite Internet isn't "all the people that don't currently have Internet," its much smaller than that. In the developed world, Internet access is frequently bundled with other services, so the potential pool of flexible customers that could be wooed away from their current service isn't that large, either. Finally, eventually the folks digging ditches are going to catch up. Once fiber gets to your premises, it will be much cheaper ber byte than satellite Internet. It doesn't matter how the last mile is traversed. Even if its fiber to an LTE tower, it'll still be cheaper per customer than satellites. I could be wrong about the economics. Presumably they've done the numbers, and the plan still closes to profitability. SpaceX isn't the only one perusing it, after all. Nevertheless, I think there is also reason for skepticism.

Klathmon 8y ago
>there is a low utilization factor for each satellite So we are getting to the limits of my knowledge here, but I believe that this won't be as big of an issue simply because sats aren't JUST communicating with the ground, they are communicating with each other as well. The routing system will take sub-optimal paths (from what you'd expect) to ensure speeds stay as high as possible and latency as low as possible (with i'm guessing some tweaking of those 2 values when near the limit), and a lot of those are going to include crossing those large oceans and deserts which will keep those sats busy, and will even route toward those large underutilized areas to spread the load out. >Second, there is a belief that the next billion new Internet users are going to be mobile users I think this is where the concept of base stations really shine. Rather than pulling fiber to all of your cell towers, you can get those "fiber quality" speeds by just having power and a place to put the tower. This means propping up new "traditional" cell towers is much cheaper (for customers of Starlink anyway!) and those mobile users could be using Starlink without even knowing it. >In the developed world, Internet access is frequently bundled with other services, so the potential pool of flexible customers that could be wooed away from their current service isn't that large, either. This is entirely anecdotal, but I'm 28 years old, and out of all of my friends and family, only my Father has cable. TV over the internet, and streaming services are taking over, and public sentiment isn't exactly high for the traditional internet service providers. Many (at least in the US) feel they don't have a choice, and if Starlink ends up being as well loved by the public at large as SpaceX or Tesla is, then I have a feeling not only will people be willing to switch, but they will jump at the chance! Though a lot of that is assuming that tesla targets most home users(which I'm not sure if they are). >Once fiber gets to your premises, it will be much cheaper ber byte than satellite Internet. It may be cheaper, but it might not be as low latency! Turns out Starlink might be able to get lower latencies across continents than traditional fiber can currently achieve. Obviously it's still up in the air if they will be able to deliver on these numbers, but if they can then there might be a small portion of users who will be willing to pay top dollar for reliable low latency connections. But taking off my fan hat, I agree that there are still a lot of unanswered questions about starlink, and if it ends up like most of SpaceX's other promises, it will be downgraded at least once in "performance", end up about 10% more expensive than expected, and a few years late, but still amazing. It may not be the next comcast powering a huge portion of the "last mile" of the internet as we know it, but I'm hopeful it will be a great alternative that can improve things for a significant number of people, and hopefully pays off well for SpaceX.