SpaceX and OneWeb internet satellite comparison

[watermod]

With a SpaceX internet why would new observational and scientific sats in orbit around the Earth need custom electronics equipment to talk to the Earth?   They could just be designed to become web-nodes on the SpaceX internet.

Sats in orbit around Earth use standard equipment, changing it to another type of equipment to talk to the "SpaceX internet" is not necessarily the best option. Maybe for some applications it could be useful, in some sense serving a similar role as TDRS or EDRS. I guess there are a few questions:
- Is SpaceX actually going to provide such service ?
- what type of equipment is needed (compared to current solutions) ?
- what would be the type of service and how is the data transmitted  (what about quality of service, reliability, security...) ?

1) I think SpaceX would be missing a big market if it didn't provide a such a service.   It would be good practice for their Mars colony's future sats and data backbone too.

2) A rad-hardened SpaceX Internet subscriber unit would, hopefully, be what the sat needs.   As a subscriber to an existing service the satellite wouldn't need to get bandwidth (frequencies) allocated to it either.  A lot of paperwork and red-tape could be thrown away with that.  SpaceX might need to modify their proposed satellites to handle signals from above too.

3) with a few hundred sats in a field of view robust high bandwidth communications should be possible.    Security, quality etc. could be handled at higher level protocols just like it is currently done on the Internet.   Scatter-gather cloud stuff etc... all in play.

If I had an observational satellite(s) I would consider the paper work red-tape savings a major plus and a simplified land side major.   If you needed really big pipes - multiple floating buoys (with subscriber units) at sea tied into undersea fiber etc would permit big and robust links for uphill/downhill distributed transmission.  You could scatter links all over the land connected to the wired Internet too.   It provides a decentralized land side more immune to natural and man made disruptions.   To break it one would need to break everybody's use of the sats.   

[MP99]

With a SpaceX internet why would new observational and scientific sats in orbit around the Earth need custom electronics equipment to talk to the Earth?   They could just be designed to become web-nodes on the SpaceX internet.

Sats in orbit around Earth use standard equipment, changing it to another type of equipment to talk to the "SpaceX internet" is not necessarily the best option. Maybe for some applications it could be useful, in some sense serving a similar role as TDRS or EDRS. I guess there are a few questions:
- Is SpaceX actually going to provide such service ?
- what type of equipment is needed (compared to current solutions) ?
- what would be the type of service and how is the data transmitted  (what about quality of service, reliability, security...) ?

1) I think SpaceX would be missing a big market if it didn't provide a such a service.   It would be good practice for their Mars colony's future sats and data backbone too.

2) A rad-hardened SpaceX Internet subscriber unit would, hopefully, be what the sat needs.   As a subscriber to an existing service the satellite wouldn't need to get bandwidth (frequencies) allocated to it either.  A lot of paperwork and red-tape could be thrown away with that.  SpaceX might need to modify their proposed satellites to handle signals from above too.

3) with a few hundred sats in a field of view robust high bandwidth communications should be possible.    Security, quality etc. could be handled at higher level protocols just like it is currently done on the Internet.   Scatter-gather cloud stuff etc... all in play.

If I had an observational satellite(s) I would consider the paper work red-tape savings a major plus and a simplified land side major.   If you needed really big pipes - multiple floating buoys (with subscriber units) at sea tied into undersea fiber etc would permit big and robust links for uphill/downhill distributed transmission.  You could scatter links all over the land connected to the wired Internet too.   It provides a decentralized land side more immune to natural and man made disruptions.   To break it one would need to break everybody's use of the sats.

It's my guess that any switching centre for the terrestrial network will end up with an uplink to the sat constellation, so there's really no need for any one customer to build such links for themselves.

Cheers, Martin

[Zed_Noir]

Got a question. Will either of the SpaceX or OneWeb constellations be encrypted enough to deter the US government from doing missive data grabs routinely? The constellations will have difficulties attracting  customers if the NSA have easy access (backdoor) to them.

[Robotbeat]

That's kind of irrelevant. The only solution to that is client-side encryption or there's no way to know that your provider doesn't have a backdoor.

But if you encrypt (with good enough algorithm) on your end, then it doesn't matter if NSA has a backdoor or not, they aren't getting your data from the network (though there are other ways of getting it, but that's not relevant to the constellation network).

[guckyfan]

It's my guess that any switching centre for the terrestrial network will end up with an uplink to the sat constellation, so there's really no need for any one customer to build such links for themselves.

Cheers, Martin

The primary purpose of the satellite network is serving those 10% worldwide that don't have direct access to a terrestrial network. That is still a huge number of subscribers. Inter-Network long distance data is secondary.

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

[Robotbeat]

It's my guess that any switching centre for the terrestrial network will end up with an uplink to the sat constellation, so there's really no need for any one customer to build such links for themselves.

Cheers, Martin

The primary purpose of the satellite network is serving those 10% worldwide that don't have direct access to a terrestrial network. That is still a huge number of subscribers. Inter-Network long distance data is secondary.

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

Source?

BTW, there exist constellations whose whole business case is "backhaul."

[guckyfan]

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

Source?

BTW, there exist constellations whose whole business case is "backhaul."

Elon mentioned that pricerange in his announcement speech in Seattle. That would be an end consumer device. A device connecting to the backbone would be different. Or am I wrong with that assumption? I don't think so.

[MikeAtkinson]

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

Source?

BTW, there exist constellations whose whole business case is "backhaul."

Elon mentioned that pricerange in his announcement speech in Seattle. That would be an end consumer device. A device connecting to the backbone would be different. Or am I wrong with that assumption? I don't think so.

Everything we know about the constellations is in the original post, together with references.

If you think there is something I've missed then provide a reference and I'll add it.

[guckyfan]

Everything we know about the constellations is in the original post, together with references.

If you think there is something I've missed then provide a reference and I'll add it.

Actually no, nothing new. That price range of 100 to 300$ is mentioned there.

I am just a little surprised that you added "depending on capability". I missed that in the speech.

[MP99]

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

Source?

BTW, there exist constellations whose whole business case is "backhaul."

Elon mentioned that pricerange in his announcement speech in Seattle. That would be an end consumer device. A device connecting to the backbone would be different. Or am I wrong with that assumption? I don't think so.

I don't know if you're wrong, but that is also my assumption.

Cheers, Martin

[Robotbeat]

Oh, you were referring to the terminal. Makes sense. No doubt the service over a few years will be much more expensive. More like $10-30/month.

[sghill]

I honestly don't think that New York is anywhere close to being the market for this constellation.

With such a massive population density, 4G or 5G or whatever has the economics to fill the area with as many small cell transmitters as necessary to provide massive mobile bandwidth. Similarly, the cost per home to provide wired or fibre services will never be smaller than at such density (and provides low cost backhaul for the mobile service).

While some people may use the sat service, it will be a fraction of the total population.

Cheers, Martin

I couldn't agree more, which is in direct contrast with advocates on this thread of this constellation somehow providing ubiquitous broadband.

[Ludus]

It's my guess that any switching centre for the terrestrial network will end up with an uplink to the sat constellation, so there's really no need for any one customer to build such links for themselves.

Cheers, Martin

The primary purpose of the satellite network is serving those 10% worldwide that don't have direct access to a terrestrial network. That is still a huge number of subscribers. Inter-Network long distance data is secondary.

As proof Elon Musk has mentioned a subscriber data unit in the price range of 100-300$.

http://www.geekwire.com/2015/elon-musk-plans-get-mars-via-seattle-spacex-founder-said-private-event/
I don't agree that's what he's saying. The reference to the subscriber unit is an answer to a question, not part of his presentation. The answer starts with saying well, it can't be free. My take is just the opposite. He's being polite and is saying they do intend to be a direct ISP but the details aren't very central to the idea.
The 10% reference is to the percent of the global local market ie direct ISP they expect NOT a percent that don't have direct access to a terrestrial network. He specifically says they expect to compete directly in areas with service. In the presentation itself I think the emphasis is very much on this being a new internet backbone (which is radically different from Wyler's so a split is understandable).

[Nomadd]

I honestly don't think that New York is anywhere close to being the market for this constellation.

With such a massive population density, 4G or 5G or whatever has the economics to fill the area with as many small cell transmitters as necessary to provide massive mobile bandwidth. Similarly, the cost per home to provide wired or fibre services will never be smaller than at such density (and provides low cost backhaul for the mobile service).

While some people may use the sat service, it will be a fraction of the total population.

Cheers, Martin

I couldn't agree more, which is in direct contrast with advocates on this thread of this constellation somehow providing ubiquitous broadband.

In case you hadn't noticed, high speed cell data service is still being sold at a price where you can use your entire $65 monthly allotment in about 20 minutes flat at maximum speed. The phone companies are also still making fortunes off of commercial customers with T1s and partial T1s at ridiculous prices by refusing to run fiber to many commercial locations.
 You better believe the urban market will be a factor.

[ChrisWilson68]

It is also unwise in my opinion to assume that the satellite division will be charged less than external customers for launches. They could charge the same for many reasons, not least because the satellite division may be spun off into a separate company.

The launch business has fairly high fixed costs compared to its marginal costs (and with reuse the costs will only move farther in that direction), and SpaceX has no serious lower-cost competitors.  That's a situation that's just begging for price discrimination, and accounting lower costs to an internal division than to external customers is a dream scenario for price discrimination.

[ChrisWilson68]

This reminds me of 2 approaches to wireless in a city. Use some big cell phone towers, or use lots of 802.11 hot spots. The big cell phone towers strategy won out.

Cellular won for exactly what it was designed for (carriers covering large areas) and 802.11 won for exactly what it was designed for (local coverage in a home or business).

[ChrisWilson68]

In the 90s Xerox Park looked at the problem from this perspective.

(1 world with X bits of data bandwidth) / Number of people = bits per person.

This of course implied 2 things:
1) more bandwidth = a few more bits per person
2) Cellularization gives a lot more bits per person by dividing N people  by M cells

#2 is completely true right up until each cell tower's signal overlaps with another's and impedes the growth of "M".  At that point #1 becomes the limiting factor.

M can be increased by putting up more satellites with narrow beams.  Robotbeat and others have been trying to get this point across, but you have yet to really address it.

This was the whole basis of my points a few days ago.  The growth of "M" is limited by spectrum channels (spectrum bandwidth) because chatter from different transmitters will overlap.  This isn't a problem at all while you've still got channels, but once you run out, you eventually get interference.

Beamforming gives a vast improvement in power use efficiency over a dumb-antenna, but it's no where near the pencil-thin beams which ground units would be required to have in order to talk to the satellite so that everyone on the ground can be served (remember; New York will look like a dot to the satellite) without overlap.

Why do you say vague things like "pencil thin" instead of using real numbers?  If the people you're arguing with are wrong, you should be able to show the math to prove it instead of continuing to wave your hands.

Cell site coverage areas are miles across, so beams would only have to be that tight.

Even then, all those beams (even lasers) would converge on a single point (the satellite) so the amount of spectrum channels available at that spot is still the critical limitation.

But that's not a limitation.  Just put up more satellites when that becomes the bottleneck.

[cro-magnon gramps]

question for the experts here... are you saying my dream of walking out my door, with Google Glass v-XXX and being able to connect with my voip contacts either singular or conference call, with internet white boarding, while surfing / reading the web, is not going to happen in the next 10 years 

Gramps

[sghill]

In the 90s Xerox Park looked at the problem from this perspective.

(1 world with X bits of data bandwidth) / Number of people = bits per person.

This of course implied 2 things:
1) more bandwidth = a few more bits per person
2) Cellularization gives a lot more bits per person by dividing N people  by M cells

#2 is completely true right up until each cell tower's signal overlaps with another's and impedes the growth of "M".  At that point #1 becomes the limiting factor.

M can be increased by putting up more satellites with narrow beams.  Robotbeat and others have been trying to get this point across, but you have yet to really address it.

This was the whole basis of my points a few days ago.  The growth of "M" is limited by spectrum channels (spectrum bandwidth) because chatter from different transmitters will overlap.  This isn't a problem at all while you've still got channels, but once you run out, you eventually get interference.

Beamforming gives a vast improvement in power use efficiency over a dumb-antenna, but it's no where near the pencil-thin beams which ground units would be required to have in order to talk to the satellite so that everyone on the ground can be served (remember; New York will look like a dot to the satellite) without overlap.

Why do you say vague things like "pencil thin" instead of using real numbers?  If the people you're arguing with are wrong, you should be able to show the math to prove it instead of continuing to wave your hands.

Cell site coverage areas are miles across, so beams would only have to be that tight.

Even then, all those beams (even lasers) would converge on a single point (the satellite) so the amount of spectrum channels available at that spot is still the critical limitation.

But that's not a limitation.  Just put up more satellites when that becomes the bottleneck.

To all three of your points: 

Beamforming is a straw man argument which is why I don't discuss the width of a beam.  It helps with power management only (which is good for the satellite or tower build-out cost).  It does nothing for channel availability planning because you have to plan for spectrum use as if the sky is saturated underneath any given satellite or groups of satellites sweeping across the sky.

You don't get to add more satellites to solve a spectrum channel scarcity problem any more than you get to layer towers one on top of each in an attempt to reuse channels in a terrestrial application without causing interference.  Those cell towers you see with lots of carriers on top of them are all using different channels, and there are a finite number of channels. Channel reuse efficiency helps, but it doesn't solve the issue.  It's not a capacity problem with the equipment, it's a capacity problem with the physical spectrum.  The solution is to buy more allocated spectrum, not to launch more satellites. 

[Nomadd]

question for the experts here... are you saying my dream of walking out my door, with Google Glass v-XXX and being able to connect with my voip contacts either singular or conference call, with internet white boarding, while surfing / reading the web, is not going to happen in the next 10 years 

Gramps

Just make sure the video of you walking into the lamp post makes it to Youtube.