Slashdot Mirror
Can We Get Global Broadband From Low-Earth Orbit Satellites? (blogspot.com)
"The internet is unavailable to and/or unaffordable by about 50% of the world population," writes Larry Press (formerly of IBM), who's now an information systems professor at California State University. But he's also long-time Slashdot reader lpress, and reports on new efforts to bring cheap high-speed internet to the entire world.
SpaceX, Boeing, OneWeb, Telesat, and Leosat are investing in very large projects to deliver global, high-speed Internet service [using low-earth orbit satellites]. This could be a significant option for developing nations, rural areas of developed nations, long-haul links, Internet of things, and more by the mid-2020s.
Parts of Alaska could see internet-via-satellite as soon as 2020, according to Larry's article, which adds that the technology could even be used to bring high-speed internet access to ships at sea.
Parts of Alaska could see internet-via-satellite as soon as 2020, according to Larry's article, which adds that the technology could even be used to bring high-speed internet access to ships at sea.
We already get internet by satellite from a dozen companies.
Here are the top ten.
http://www.toptenreviews.com/s...
Getting a two-way connection from a moving satellite is a nightmare. You get all kinds of frequency-shift, Doppler, atmospheric, and localized multi-path problems. You'll need a big chunk of spectrum for all the error correction and sync signals required. You'll either need a tracking dish, which will be expensive, or a phase-array, which is cheaper to build but will require a more complicated and expensive front-end.
It may work for niche cases for low-bandwidth applications in remote areas. I'm guessing the uplink hardware will be so expensive that you'll have micro-ISPs serving small areas.
My Other Computer Is A Data General Nova III.
LEO Satellites at 1200 miles up will have a minimum Earth-Ground latency of 24 milliseconds and Earth-Ground-Earth Latency of 48 milliseconds because of the speed of light ---- this is a major latency issue unless there are MANY infrastructure Earth stations at major colocation facilities AND the traffic can be efficiently routed, so we're not landing traffic in a NEW YORK internet exchange that then needs to be routed to SAN FRANCISCO, or Atlanta, and thus appending another 50 milliseconds of ground latency after the satellite hop, for example.
Being somewhat familiar with SpaceX's plans here are a few advantages of their approach, I guess compared to traditional satellite providers:
They're planning to deploy thousands of cheap, small, short-lived satellites in LEO, which means:
-They get the advantages of cheaper production due to economies of scale, orders of magnitude better than something like GPS or Iridium.
-So many units means they can just over-provision, use less hardened, cheaper components, and just replace units as they fail.
-Being in LEO means they have a shorter lifespan due to atmospheric drag, so they stay up for maybe 5 years, drop into the atmosphere and are replaced by newer, better hardware.
-I did a back of the envelope calculation once and I think I came up with something like 1/3 the latency of fiber when going halfway around the earth, due speed of light in glass vs air/vacuum, and the various geographical features cables need to contend with.
-One of the reasons I remember being mentioned for SpaceX getting into building their own satellites when their rocket reuse program was just getting off the ground is they'll eventually end up with a supply of rockets that's larger than the entire launch market is going to need, at least in the short term, so this is a way for them to be their own customer and amortize the cost of the rocket by reflying it 10 times with cargo they can afford to lose.