5 ms·

The reason they specified sparsely populated areas is that Starlink is designed for sparsely populated areas. In a denser area the performance would drop drasti

The reason they specified sparsely populated areas is that Starlink is designed for sparsely populated areas. In a denser area the performance would drop drastically because you'd need far higher satellite density to sufficiently distribute the load.

labcomputer 3y ago
That may be true with current space vehicles, but there’s no fundamental reason that two or more SVs can’t form a distributed phased array (for MIMO a la WiFi). At least from a physics / spectrum perspective. The engineering might be tricky. With SVs separated by hundreds of km, the beam width would be absolutely tiny, and channel reuse would be effectively infinite.
- dotnet00 3y ago
  Hadn't heard of that concept before, sounds pretty interesting! Will have to read up on it more.
- rini17 3y ago
  The distributed array will however share one link and satellites don't have unlimited capacity either. Wonder why you'd rather fill LEO to the brim, emitting vast amounts of CO2 in the process and filling stratosphere with metal oxides, than figure out how to improve infrastructure in your neighborhood.
  - LeifCarrotson 3y ago
    I've never heard anyone complain about filling the stratosphere with metal oxides, do you have somewhere I can read more about this risk? Naively, I would have expected that even thousands of 250kg satellites would be insignificant on the scale of the whole stratosphere (we're not talking about surface-level CO2, where we're pumping out gigatons per year), but I have no idea what that ecosystem is like. Is the metal oxide concentration measurable? I do think this is a temporary issue; once the cable ISP monopolies are no longer valuable because anyone can just get Starlink, that competitive pressure will push down hard on those cable monopolies and infrastructure laws will have to return to sanity. Terrestrial ISPs with fiber-sharing requirements and appropriate infrastructure build-out could absolutely out-compete Starlink, but right now they can just collude together and lobby against changing the status quo. I would much prefer to have a functional bureaucracy and intelligent, efficient infrastructure! But it's probably easier to fix that by launching the largest satellite constellation ever than by playing politics against some of the wealthiest media companies on the planet.
    - Clubber 3y ago
      "Space junk," was a boogie man the news threw at the wall and didn't quite stick. Many of the pictures/diagrams of the satellites made them appear about 1000x bigger than they actually are. Example: https://www.nationalgeographic.com/science/article/space-junk https://www.nationalgeographic.com/science/article/space-jun... Beware the murder hornets! https://www.nytimes.com/2020/05/02/us/asian-giant-hornet-washington.html https://www.nytimes.com/2020/05/02/us/asian-giant-hornet-was...
      - chargingmarmot 3y ago
        I agree with the sentiment about media sensationalism but I don't think "space junk" is an issue we should be so quick to dismiss. Just because the problem hasn't gotten too bad yet doesn't mean it isn't real (especially because the problem involves a positive-feedback "domino effect" and junk might be infeasible to clean up). https://en.wikipedia.org/wiki/Kessler_syndrome https://en.wikipedia.org/wiki/Kessler_syndrome With regard to Starlink: I was a little bit concerned because of the large number of satellites they're putting up, but as far as I understand, Starlink isn't actually a big risk right now as far as space junk goes because the satellites orbit at low altitudes (340 miles) and are expected to naturally de-orbit relatively quickly due to atmospheric drag. (The decision to put them at low altitudes had to do with network latency.) If Elon were to say tomorrow, "hey we're moving Starlink out to 500 miles" I would hope some kind of regulation would be able to prevent that.
        rini17 3y ago
        Everything we launch has to pass through these low altitudes. If these altitudes become full of debris we might have to stop all launches for several years [1] and no human presence at space for a bit longer. Quite scary, better to handwave it away "aint happen space is big". https://space.stackexchange.com/questions/59559/how-long-can-the-spacex-starlink-satellites-survive-before-they-deorbit https://space.stackexchange.com/questions/59559/how-long-can...
        sliken 3y ago
  - WalterBright 3y ago
    Compare the CO2 emissions of launching a satellite with the CO2 emissions of digging trenches.
    - cozzyd 3y ago
      The right thing to compare it to is building a cell tower
- anderspitman 3y ago
  How would you synchronize them?
- HWR_14 3y ago
  Each satellite has N routers, and can support a certain bandwidth. Even if you hardwired a long LAN cable to it, more users drops the speed.
- wayfinder 3y ago
  I imagine the total cost of installation and ongoing maintenance of all the infrastructure required for this proposed setup for a substantial amount of urban users dwarfs the equivalent total cost of just installing new cabling and the lobbying required to get that done, especially over a 20 year period.
  - jtchang 3y ago
    Don't underestimate lobbying costs in dense urban areas. Whenever you have to deal with people the alternative is often easier.
    - MichaelZuo 3y ago
      I highly doubt it would 10x the cost though.
asynchronous 3y ago
It’s actually not the satellites as the bottleneck typically from what I understand, it’s the ground stations that serve those areas getting too congested.
- reportingsjr 3y ago
  The satellites may not currently be the bottleneck for certain areas, but they definitely don't support anything resembling a densely or even sparsely populated area. Typical calculations (using only publicly available info as a caveat) generally point towards the current starlink constellation supporting a really low density of antennas. If you want the details, the person who created starlink.sx has a great blog with some incredible calculations[0]. The section that is most relevant is under the heading "Simulation with TDM and beam spread combined". Basically, the result is that starlink could support around 15 antennas/user terminals per "cell" at 75mbps, where a cell is about 252 square km. This is roughly 1 terminal per 17 square km. If you allow for 10x oversubscription, fairly common for ISPs, that gets you to 1 terminal per 1.7 square km. Not really that impressive when you compare it to the density and services where 99.9% of people live, but it clearly helps out the small number of people way out there. 0: https://mikepuchol.com/modeling-starlink-capacity-843b2387f501 https://mikepuchol.com/modeling-starlink-capacity-843b2387f5...
  - asynchronous 3y ago
    This was a very interesting read, thank you for correcting me. Enjoyed the deep dive.
esskay 3y ago
If they can get the interconnect working that issue goes away (its mostly groundstation bandwidth limiting things). The 'end goal' is to have it use the laser link to send the traffic across to a satellite close to the final destination, that should in theory help massively redistribute the ground station traffic. It's obviously not foolproof, there's still bandwidth to be shared there but using a combination of some clever routing you could theoretically spread traffic across multiple satellites and groundstations with minimal latency impact.

labcomputer 3y ago

That may be true with current space vehicles, but there’s no fundamental reason that two or more SVs can’t form a distributed phased array (for MIMO a la WiFi). At least from a physics / spectrum perspective. The engineering might be tricky. With SVs separated by hundreds of km, the beam width would be absolutely tiny, and channel reuse would be effectively infinite.

dotnet00 3y ago
Hadn't heard of that concept before, sounds pretty interesting! Will have to read up on it more.
rini17 3y ago
The distributed array will however share one link and satellites don't have unlimited capacity either. Wonder why you'd rather fill LEO to the brim, emitting vast amounts of CO2 in the process and filling stratosphere with metal oxides, than figure out how to improve infrastructure in your neighborhood.
- LeifCarrotson 3y ago
  I've never heard anyone complain about filling the stratosphere with metal oxides, do you have somewhere I can read more about this risk? Naively, I would have expected that even thousands of 250kg satellites would be insignificant on the scale of the whole stratosphere (we're not talking about surface-level CO2, where we're pumping out gigatons per year), but I have no idea what that ecosystem is like. Is the metal oxide concentration measurable? I do think this is a temporary issue; once the cable ISP monopolies are no longer valuable because anyone can just get Starlink, that competitive pressure will push down hard on those cable monopolies and infrastructure laws will have to return to sanity. Terrestrial ISPs with fiber-sharing requirements and appropriate infrastructure build-out could absolutely out-compete Starlink, but right now they can just collude together and lobby against changing the status quo. I would much prefer to have a functional bureaucracy and intelligent, efficient infrastructure! But it's probably easier to fix that by launching the largest satellite constellation ever than by playing politics against some of the wealthiest media companies on the planet.
  - Clubber 3y ago
    "Space junk," was a boogie man the news threw at the wall and didn't quite stick. Many of the pictures/diagrams of the satellites made them appear about 1000x bigger than they actually are. Example: https://www.nationalgeographic.com/science/article/space-junk https://www.nationalgeographic.com/science/article/space-jun... Beware the murder hornets! https://www.nytimes.com/2020/05/02/us/asian-giant-hornet-washington.html https://www.nytimes.com/2020/05/02/us/asian-giant-hornet-was...
    - chargingmarmot 3y ago
      I agree with the sentiment about media sensationalism but I don't think "space junk" is an issue we should be so quick to dismiss. Just because the problem hasn't gotten too bad yet doesn't mean it isn't real (especially because the problem involves a positive-feedback "domino effect" and junk might be infeasible to clean up). https://en.wikipedia.org/wiki/Kessler_syndrome https://en.wikipedia.org/wiki/Kessler_syndrome With regard to Starlink: I was a little bit concerned because of the large number of satellites they're putting up, but as far as I understand, Starlink isn't actually a big risk right now as far as space junk goes because the satellites orbit at low altitudes (340 miles) and are expected to naturally de-orbit relatively quickly due to atmospheric drag. (The decision to put them at low altitudes had to do with network latency.) If Elon were to say tomorrow, "hey we're moving Starlink out to 500 miles" I would hope some kind of regulation would be able to prevent that.
      - rini17 3y ago
        Everything we launch has to pass through these low altitudes. If these altitudes become full of debris we might have to stop all launches for several years [1] and no human presence at space for a bit longer. Quite scary, better to handwave it away "aint happen space is big". https://space.stackexchange.com/questions/59559/how-long-can-the-spacex-starlink-satellites-survive-before-they-deorbit https://space.stackexchange.com/questions/59559/how-long-can...
        sliken 3y ago
- WalterBright 3y ago
  Compare the CO2 emissions of launching a satellite with the CO2 emissions of digging trenches.
  - cozzyd 3y ago
    The right thing to compare it to is building a cell tower
anderspitman 3y ago
How would you synchronize them?
HWR_14 3y ago
Each satellite has N routers, and can support a certain bandwidth. Even if you hardwired a long LAN cable to it, more users drops the speed.
wayfinder 3y ago
I imagine the total cost of installation and ongoing maintenance of all the infrastructure required for this proposed setup for a substantial amount of urban users dwarfs the equivalent total cost of just installing new cabling and the lobbying required to get that done, especially over a 20 year period.
- jtchang 3y ago
  Don't underestimate lobbying costs in dense urban areas. Whenever you have to deal with people the alternative is often easier.
  - MichaelZuo 3y ago
    I highly doubt it would 10x the cost though.

asynchronous 3y ago

It’s actually not the satellites as the bottleneck typically from what I understand, it’s the ground stations that serve those areas getting too congested.

reportingsjr 3y ago
The satellites may not currently be the bottleneck for certain areas, but they definitely don't support anything resembling a densely or even sparsely populated area. Typical calculations (using only publicly available info as a caveat) generally point towards the current starlink constellation supporting a really low density of antennas. If you want the details, the person who created starlink.sx has a great blog with some incredible calculations[0]. The section that is most relevant is under the heading "Simulation with TDM and beam spread combined". Basically, the result is that starlink could support around 15 antennas/user terminals per "cell" at 75mbps, where a cell is about 252 square km. This is roughly 1 terminal per 17 square km. If you allow for 10x oversubscription, fairly common for ISPs, that gets you to 1 terminal per 1.7 square km. Not really that impressive when you compare it to the density and services where 99.9% of people live, but it clearly helps out the small number of people way out there. 0: https://mikepuchol.com/modeling-starlink-capacity-843b2387f501 https://mikepuchol.com/modeling-starlink-capacity-843b2387f5...
- asynchronous 3y ago
  This was a very interesting read, thank you for correcting me. Enjoyed the deep dive.

esskay 3y ago

If they can get the interconnect working that issue goes away (its mostly groundstation bandwidth limiting things). The 'end goal' is to have it use the laser link to send the traffic across to a satellite close to the final destination, that should in theory help massively redistribute the ground station traffic. It's obviously not foolproof, there's still bandwidth to be shared there but using a combination of some clever routing you could theoretically spread traffic across multiple satellites and groundstations with minimal latency impact.