• Welcome! The TrekBBS is the number one place to chat about Star Trek with like-minded fans.
    If you are not already a member then please register an account and join in the discussion!

SpaceX Planning 12,000 Satellites For Worldwide Direct Internet Access

Snaploud

Admiral
Admiral
Elon Musk is at it again.

SpaceX’s Latest Launch Marks Its First Step Toward Beaming Internet Down From Space

By Neel V. Patel

Feb 22, 20182:45 PM

SpaceX launched two of its own satellites into space on Thursday morning, and if you watched the webcast of the launch, you probably only remember hearing about the primary payload: a Spanish Earth-observation satellite called Paz. While the primary payload is important, it’s the two other satellites that deserve more attention. Unlike the last thing Elon Musk sent up into space, a cherry-red convertible, these two satellites, called Microsat-2a and Microsat-2b (nicknamed Tintin A and Tintin B, respectively) are a pretty critical part of the company’s next venture: to beam the internet from orbit back down to Earth.

SpaceX launched two of its own satellites into space on Thursday morning, and if you watched the webcast of the launch, you probably only remember hearing about the primary payload: a Spanish Earth-observation satellite called Paz. While the primary payload is important, it’s the two other satellites that deserve more attention. Unlike the last thing Elon Musk sent up into space, a cherry-red convertible, these two satellites, called Microsat-2a and Microsat-2b (nicknamed Tintin A and Tintin B, respectively) are a pretty critical part of the company’s next venture: to beam the internet from orbit back down to Earth.

Tintin A and B are prototype instruments built by SpaceX to test out technologies that would provide internet access to people down on the surface. The company is moving forward with plans to create a giant internet service called Starlink, comprised of 12,000 satellites organized high in two different altitudes (200 miles and 700 miles above the surface of the planet), operating at different radio frequencies, to cover individuals nearly anywhere on Earth.

It’s a massive, ambitious plan, but also a lucrative one if SpaceX can pull it off. The Wall Street Journal reports the company plans to have more than 40 million subscribers by 2025, which at the point should be enough to generate some $30 billion in revenue.

Why send internet down from space at all? It could be a great equalizer: Broad swaths of the world are still hampered by poor access to broadband internet. If you’re living in a town comprised of just a few dozen people hundreds of miles from the city, or you live in a developing part of the world that hasn’t been graced by a telecom giant, you probably have bad internet, if you have access to the web at all. Satellite internet can solve that problem, beaming connectivity down to even the most remote places on the planet.

https://slate.com/technology/2018/0...-toward-beaming-internet-down-from-space.html
 
Interesting but it leaves a few questions unanswered:
  • What will download bandwidth, upload bandwidth and latency be like, not to mention the data limit and service cost?
  • Will the rates be significantly cheaper than existing satellite phones?
  • Will some services such as audio and video comms be allowed.
  • Are additional uplink ground stations required? If not, how many simultaneous uplinks/downlinks are possible per satellite?
I assume a Falcon Heavy might be able to deploy hundreds of these satellites at a time. Mass production will drive down cost per satellite.
 
Don't know the answer to those questions, but Wikipedia says the satellites are at 625 kilometer orbit, or 2600 km full round trip (ground A → satellite → ground B → satellite → ground A), assuming you're texting with the classmate on the desk behind you. That's 8.7 light milliseconds, as a theoretical maximum.

Now, realistically:
1. Unless SpaceX also run communication service of their own, communications would go through a ground server, so the classmate may be twice that distance away (but not if they are using a ground station, and only you're on a satellite). That would be particularly bad if you're video gaming (assuming the game server won't be in orbit), where an exchange would travel 8 times between ground and orbit to get back to you. ⁽¹⁾
2. To go back to the real world of today, the theoretical classmate behind me would be electronically as far as someone in this lovely park in Torino, which is 1300 km from me (2600 round trip). I've got delay of 45-81 ms to sites related in Torino, although that's not a perfect comparison – ground communications don't go in a straight line, and some of those sites are in Toscana or France, for some confusing reasons. That's way more than 8.7 ms, but I think we could do better.

It would still be insanely better than current geostationary satellite internet. A friend of mine was using such back in the day, and her latencies were insane. And that was just a silly proxy that only send the downstream over the satellite. Can't find any real-life figures on it right now, but speed of light math says 239 ms if its sat is one-directional like hers. That's my exact delay to Yahoo.com, so even GEO internet is not downright horrible, although when bi-directional it goes to the whooping 477 ms – which would make your video chat like one of those TV interviews.

But here's the good stuff: You can host sites in orbit, which could be theoretically faster, especially if the BFR can launch a lot of servers. Now, that's only good if the server is flying right above your head. However, even if the physical server needs to move to stay in orbit, a virtual one does not - it may be possible to live migrate virtual servers between orbital servers so that they stay in one place, and the laws of physics don't apply to them. Live migration does have a considerable cost in bandwidth, and delay would be detrimental to it, so I hope your satellites are really close to one another.

P.S. Wild guess - the bandwidth will suck. At first. But we need to fix the Martian bandwidth problem in the long term, so some of that can be reapplied to those satellites as well. Hopefully.

(1) For reference, human reaction times: 160-180 ms (unverified), cinema: 42 ms per frame – the answer to life, universe and everything. My money says you won't notice 17 ms extra, but then I'm not a gamer.
 
My guess is that you will be able to use your ordinary mobile on a Musk-owned network and the automated ground station implemented by the network will compute the optimum route using some algorithmic combination of bandwidth and latency, which might not involve using any satellites at all. I doubt the satellites will host any web sites although I expect they and the ground stations could perform some caching.
 
Wouldn't satellite based internet have the same shortcomings as satellite based television?
i.e. It wouldn't work very well on a rainy or cloudy day.
 
i assume the ground stations will have powerful transmitters/receivers. Coverage will probably be restricted to remote, mountainous or arid regions where there is no infrastructure for access currently? Monsoon-prone areas such as India, south China and Indochina might well be problematic. I don't see an economic case for establishing infrastructure if the population density is low.

The Register has an article and discussion thread about the proposal:

https://www.theregister.co.uk/2018/02/22/spacex_broadband_satellies/

One comment claims a latency of a few tens of ms as compared to several hundred ms for geosynchronous.
 
Musk has tweeted that his network won't be based on IPv6, will use updatable hardware-based encryption, and will be inherently peer-to-peer. I assume one would need to tunnel IPv6 and other protocols over it for access to the Internet. Also, people would be able to buy their own uplink relays if they chose (and had the money) except in countries that insist on being able to snoop traffic. He thinks it unlikely that his network will be allowed in China, which is a big proportion of the world's population.
 
For a guy so terrified of AI taking over the world, Elon Musk sure seems determined to usher in the coming of Skynet...

Always remember to clap at the end of the show!

Applause.gif
 
Wouldn't satellite based internet have the same shortcomings as satellite based television?
i.e. It wouldn't work very well on a rainy or cloudy day.
Starlink (and the constellations planned by Bezos and Google as well) are Low Earth Orbit satellites as opposed to the GEO sats which are much further away.

While weather makes little importance (there isn't going to be rain between Low Earth Orbit and geosynch orbit), it will affect latency. Likewise the ability to have multiple satellites over the same location will allow the signal to "punch through" from multiple points. Your device will find multiple attach points, and the network will be robust enough to switch as needed. This has already been the case with cellular networks for some time. I can't remember the last time I lost connection on a cellphone from weather.

These LEO constellations are not really a new idea. They were going to be a thing in the late 90's, and had inspired a slew of launch company startups like Pioneer, Kistler, and others hoping to catch a bit of the launch business. But the moment wasn't right and apart from one fleet that did ultimately get finished in a much smaller capacity, it took another 20 years or more to really reach the starting point.

And yes, dark skies at night are history. It was going to happen. It is happening. Astronomy will have to move off-earth completely.
 
Teledesic was one of the first planned constellations.

I like the old ORBITAL ANTENNA FARM concept.
Fewer, large comsats higher up--out of the way of third world ASATs
https://ui.adsabs.harvard.edu/abs/1977AsAer..15...20E/abstract

p-comsat-egge-hudson.jpg

Medium-sized, next-generation, "large" communications satellite platform. The beams of the platforms, the large antennas, bulk radiation shielding (not pictured), and fuel propellants for stationkeeping could come from asteroidal and/or lunar materials. Eventually, the solar cell power plant and heat radiators, too. In any case, who can compete with services from this? Not Teledesic or Iridium. It will make a lot of competitors obsolete, so if you can't beat us, join us.

Artist: David Egge. Copyright owner: Gary C. Hudson (Hudson Engineering, Menlo Park, CA); used in his newsletter "The Commercial Space Report". This is an illustration of an orbital antenna farm as proposed by the Director of COMSAT
https://www.permanent.com/space-products-and-services.html


One such concept:
http://www.astronautix.com/g/globis.html

4 to 6 such satellites would make up a global information system, solving congestion problems in geosynchronous orbit. It was felt that only heavy platforms in an integrated system could operate high-volume optical inter-satellite data links.

A very funny cartoon
https://xkcd.com/2264/
 
Last edited:
Starlink 6 launch on March 11. I looks like they are getting into a cadence of 2 launches a month for this satellite fleet.
 
If you are not already a member then please register an account and join in the discussion!

Sign up / Register


Back
Top