A lot left on the table by OneWeb and the Google/SpaceX constellations

[nadreck]

I think building the network that Elon has proposed, and that we now have a rumor that Google may participate in, is a game changer not in what they provide as a potential communications service, but instead how they change the landscape for other people with other business ideas.

I would note that Elon likes competition, and in some cases I think he deliberately courts competition as tool to help his efforts along (including using the public appearance of his competitor(s) to build the climate of opinion he seeks for his own efforts).

So to me the fact that Elon, at least for now, is eschewing mobile with his satellite communications constellation leaves a bigger meal on the table for his competition than he is taking himself. Whether Elon's satellite factory ever builds to order satellites for the applications of other companies, his factory will force (but also enable) other companies to duplicate whatever economies his constellation has. While I personally see someone, and I suspect it will be an Elon Musk company, creating a $100,000  satellite bus with the storage batteries, solar panels, hall effect thruster, guidance, and ultra basic communications with room for 50 to 100kilos of customizable capabilities that might cost anywhere from another $100,000 for something that copies an existing application, to several million for a totally custom package. However this is between two and three orders of magnitude cheaper than current state of the art, and it will be five years minimum before it is here with a few things which may or may not develop the way Elon intends (the revolution in space hardened technology that he is looking for, the performance of satellites using GPS for their own positioning, the hall effect thrusters, and of course the communications hardware including data switching that his application needs), but however the technology does develop it is sure to be a game changer.

So what applications exist for manufacturing large quantities of satellites and positioning them on a variety of different planes to cover most or all of the earth? The fixed data, alternate GPS, GPS for space craft, and the mobile voice and data are the most obvious answers. A number of people on NSF speak about a number of imaging applications. Then, besides Elon's suggestion of a Mars network, there are probably a number of NEO apps that would make sense. If we imagine a 2.4km/s ΔV vehicle budget (just taken from one WAG on amount of hall effect ΔV a vehicle with a hall effect thruster would have) we could probably have 2 or 3 launch out of Earth orbit on an F9 as we are going to see with DSCOVR (or 1 on a trajectory to Mars, Venus or main belt asteroids) and have the energy budget to do something useful.

So what I would like to do is collect some speculation on applications for these low cost birds that we may seeing built 5 to 10 years from now.

[Robotbeat]

I personally think that if the SpaceX constellation is successful (and as long as it flies, it surely will at least pay for itself... think at least ViaSat-level revenue, and almost surely a multiple of that), SpaceX will probably be buying up cell spectrum.

...that is, if they think they can get a cell-based architecture to close. For cities, local cell sites are pretty darned optimal (although no reason they couldn't use the SpaceX constellation for backhaul!!!). If you wanted to jazz it up a bit, put the cell sites on persistent, solar-powered high altitude drones. (But ground-level would probably work just fine.)

[gosnold]

Imagine how much the Pentagon would be willing to spend for an imaging constellation with persistent view of any point of the globe (provided there are no clouds). It is only a matter of adding an optical payload, since the data handling capability is provided by the standard telecom payload with its intersatellite and sat-to-ground links.

Also, if the articles about Google investing in SpaceX are true, remember Google has bought Skybox imaging, an optical sat company. They already have a constellation in the works to do 1m resolution imaging, their 2nd generation could reuse the SpaceX bus and comms payload to provide commercial real-time coverage.

Edit:spelling

[Lars-J]

Imagine how much the Pentagon would be willing to spend for an imaging constellation with persistent view of any point of the globe (provided there are no clouds). It is only a matter of adding an optical payload, since the data handling capability is provided by the standard telecom payload with its intersatellite and sat-to-ground links.

1100 km altitude is not ideal for that kind of imaging, and would likely require a *significant* optical instrument, not something that you just add as side payload for 4000+ satellites!

Also, if the articles about Google investing in SpaceX are true

It is true: http://www.spacex.com/press/2015/01/20/financing-round

[gosnold]

I did not mean that you carry the optical sensor as a secondary payload. You can keep the same bus, cut a bit on the comms payload, add an optical payload and have the satellite work at a lower altitude (700km for instance, but with a Hall thruster you can afford to go quite low). These optical models can form a separate constellation.
Regarding the optical payload, it is not that hard to do. Skysat-1 from Skybox imaging has a mass of 100kg, and provides 1m resolution from 600km, using low-cost technologies (digital TDI). The balance between resolution, field of view, and cost will depend on customer requirements though.

[meekGee]

If (when) this works, cell companies will be unaffected - they might switch to having commX antennas at their tops, and no data connection to a terrestrial long-haul carrier.

Same with any hot-spot provider - the "last 1000 feet" will be largely unaffected.

The losers here are the cable carriers - ATT and Comcast.  Expect $hit to hit the fan on their behalf...   No sympathy here though...

[nadreck]

I edited the title to reflect that Google involvement is a done deal. I am surprised by Fidelity being in there, as well I don't credit them with creative input on the actual data constellation project (though maybe I am underestimating them).

MG, the cable guys are definitely in retreat on other fronts and opening this new front will be a last nail in their coffin. I think the market for the last 1000 feet to homes and businesses will be changed outside urban cores, but really, what I wanted to suggest with this thread is what other killer constellation or micro/mini sat apps are left out there after the two constellations mentioned in the title. I think that if someone successfully puts up a 700 or 4000 satellite constellation it will not be the only one, but its creation will have revolutionized both what organizations see as possible and of course what actually is possible since satellites at these price points are disruptors to all sorts of established applications, and of course create potential for new ones.

[Burninate]

I did not mean that you carry the optical sensor as a secondary payload. You can keep the same bus, cut a bit on the comms payload, add an optical payload and have the satellite work at a lower altitude (700km for instance, but with a Hall thruster you can afford to go quite low). These optical models can form a separate constellation.
Regarding the optical payload, it is not that hard to do. Skysat-1 from Skybox imaging has a mass of 100kg, and provides 1m resolution from 600km, using low-cost technologies (digital TDI). The balance between resolution, field of view, and cost will depend on customer requirements though.

You can actually devise surprisingly specialized designs for an imaging satellite if you're building a whole network of them.

A bunch of GOCE-shaped imaging satellites flying at 250-300km with ion thrusters and substantial propellant mass fraction could take similar resolution photos over a smaller swath range, using half the aperture (and ~1/8 the optics mass) of a satellite at 500-600km, or a quarter the aperture (and ~1/64 the optics mass) of a satellite at 1000-1200km.

[butters]

I can imagine multiple remote sensing constellations for different and/or competing services, but there's no compelling reason to have more than one packet-based communications network in space. The Internet eventually replaces all other incompatible networks in all but the most demanding niche applications. Legacy channel-based comsat services are living on borrowed time.

[nadreck]

I can imagine multiple remote sensing constellations for different and/or competing services, but there's no compelling reason to have more than one packet-based communications network in space. The Internet eventually replaces all other incompatible networks in all but the most demanding niche applications. Legacy channel-based comsat services are living on borrowed time.

Hmm, I can think of some things that might interface with communications networks but be very specialized data satellites of their own. Imagine if, at very low bandwidth and with more penetrating frequencies you have a constellation that predominantly listens for signals with far more sensitive receivers than are being used for general communications, the signals it listens for might be animal tag trackers, a new generation of EPIRB type transmitters, lowjack devices etc.

[butters]

I can imagine multiple remote sensing constellations for different and/or competing services, but there's no compelling reason to have more than one packet-based communications network in space. The Internet eventually replaces all other incompatible networks in all but the most demanding niche applications. Legacy channel-based comsat services are living on borrowed time.

Hmm, I can think of some things that might interface with communications networks but be very specialized data satellites of their own. Imagine if, at very low bandwidth and with more penetrating frequencies you have a constellation that predominantly listens for signals with far more sensitive receivers than are being used for general communications, the signals it listens for might be animal tag trackers, a new generation of EPIRB type transmitters, lowjack devices etc.

Yes, tracking is a distinct application, a unique hybrid of communications and remote sensing. Although there are some benefits to having one universal tracking system, the "network effect" is not as pronounced as it is in more typical communications networks, so there could very well be multiple competing asset tracking systems in space. For example, maybe one for tracking a modest number of big/expensive things and another for tracking a huge number of small/inexpensive things.

[Razvan]

Imagine how much the Pentagon would be willing to spend for an imaging constellation with persistent view of any point of the globe (provided there are no clouds). It is only a matter of adding an optical payload, since the data handling capability is provided by the standard telecom payload with its intersatellite and sat-to-ground links.

Also, if the articles about Google investing in SpaceX are true, remember Google has bought Skybox imaging, an optical sat company. They already have a constellation in the works to do 1m resolution imaging, their 2nd generation could reuse the SpaceX bus and comms payload to provide commercial real-time coverage.

Edit:spelling

It kind of opens the Pandora's box. Of course, Pentagon would be xxxinterested especially now, under budget strain, in low cost imaging assets and other Agencies will manifest a sudden appetite for listening to the whispers  of this new Space hypernet. Question I see is what guarantees that other uninvited guests will not benefit of it?
Probably, Russia, China, Iran, North Korea, Venezuela and few other Countries wouldn't allow it on their territories. Well, never mind that, will be reasonably easy to snick in over the border a small box like phased antennae...

[savuporo]

Yes, tracking is a distinct application, a unique hybrid of communications and remote sensing. Although there are some benefits to having one universal tracking system, the "network effect" is not as pronounced as it is in more typical communications networks, so there could very well be multiple competing asset tracking systems in space. For example, maybe one for tracking a modest number of big/expensive things and another for tracking a huge number of small/inexpensive things.

at least one such a tracking system is already in orbit, "hosted" on an existing commercial constellation. "GPS" that works underground.

last I heard, the operator was looking into ways of commercializing this but their coverage and scalability was not ideal

[Oli]

MG, the cable guys are definitely in retreat on other fronts and opening this new front will be a last nail in their coffin.

I think people need a reality check.

Lets look at the OneWeb constellation for example since there we actually have the relevant specs.

8gb per sat. One sat covering a region of roughly ~1000 miles across (there is more overlap when you get to the poles), so easily 100m+ people per sat.

4k streaming will be standard in a few years, which requires at least a steady 20mb/s. So with one OneWeb sat you can serve 400 households streaming a movie or a tv show in the evening.

400 out of 100m+.

Meanwhile Petabits per sec have been demonstrated for a single bloody fiber. Remember, you only have to lay a pipe once, not every 5 years.

[Jarnis]

I am worried about the bandwidth, but until exact details about the payload on these sats is available, it would be fairly pointless to speculate.

Until info becomes available... I sincerely doubt Elon and Google would be doing this if we'd just get a big cloud of bandwidth-starved sats that will be able to do good approximation of a dialup/slow DSL for a bunch of users in far away places. That won't get subscribers. Only way this plan makes sense is if they push the state of the art as far as bandwidth is concerned.

[JasonAW3]

The thing I find interesting is, that with the economies of scale and the mass production of satillites, the cost of losing a sat will come way down to an almost disposable level.  so satillite lossesnot only become managible for large companies, but small businesses could then create dedicated sats, based on the mass production sat bus that SpaceX will develop, for whatever mapping, communication or experimental purposes that the user may want.

It would not suprise me in ten years that a lunar orbiter and lander would be sent to the moon for around $500,000 by a private company.

[Danderman]

It would seem that the greatest consequence of this effort may be the advent of extremely low cost 100 kg satellite buses.  Then the issue arises of launching such platforms in small numbers, ie for Mars missions.

[nadreck]

It would seem that the greatest consequence of this effort may be the advent of extremely low cost 100 kg satellite buses.  Then the issue arises of launching such platforms in small numbers, ie for Mars missions.

If the buses are that cheap one of the reasons they are that cheap is that there is a big enough demand for them so ones for common orbits, destinations, will be able to share launchers (the constellation deploy systems could mix and match different customer's needs). Obviously Mars or specific other BEO destinations would require more co-ordination and likely delay, though I can envisage (given a 2.4Km/s hall thruster ΔV budget) a highly eliptical parking orbit for craft going to different BEO destinations.

[meekGee]

MG, the cable guys are definitely in retreat on other fronts and opening this new front will be a last nail in their coffin.

I think people need a reality check.

Lets look at the OneWeb constellation for example since there we actually have the relevant specs.

8gb per sat. One sat covering a region of roughly ~1000 miles across (there is more overlap when you get to the poles), so easily 100m+ people per sat.

4k streaming will be standard in a few years, which requires at least a steady 20mb/s. So with one OneWeb sat you can serve 400 households streaming a movie or a tv show in the evening.

400 out of 100m+.

Meanwhile Petabits per sec have been demonstrated for a single bloody fiber. Remember, you only have to lay a pipe once, not every 5 years.

"Last nail in the coffin" is probably an overstatement, but it will hurt their business a lot.

The Wyler constellation you bring up is a LOT less ambitious, than the SpaceX one, which is what was being discussed.

Also, the SpaceX constellation is just the first one.  If it works well, you can upscale it very easily.  You might end up with cell towers in all but city-centers that don't have a connection to the back-haul, and if you're only talking city-centers, WiMax might do the trick for last 1000' distribution.

This is definitely a game changer, but exactly how disruptive remains to be seen.

A first clue will be if ComCast/Verison/ATT start fighting it.


 

[Oli]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

[TrevorMonty]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

Technology has improved and new players are better financed. Here is hoping it works this time.

[oldAtlas_Eguy]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

Technology has improved and new players are better financed. Here is hoping it works this time.

Actually the problem back then was the players were not financed! A few milion vs $1B is a big difference.

[nadreck]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

A) Well, the electronics are different we have had about 8 - 10 Moore's law generations since then (256 to 1024 times the performance).

B) The market is different, the average consumer of video content watched broadcast television not VOD over IP.

C) The launcher market just got another large step cheaper (Musk vs Russia) with promises of further price cuts which, if they do not appear, will be just like Pegasus and other launchers that were supposed to revolutionize but failed to.  So we don't know if that step is different yet, but we will in a few years. 

D) The Chinese and Indian economies, as well as other developing nations, can easily absorb 10 times the amount of services that they could have in 1998.

E) We are not on the brink of an inventory surplus from over manufacturing in anticipation of Y2K and then crash of dumping/discounting that inventory (including massive overbuild of fiber capacity) post Y2K

[Oli]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

A) Well, the electronics are different we have had about 8 - 10 Moore's law generations since then (256 to 1024 times the performance).

B) The market is different, the average consumer of video content watched broadcast television not VOD over IP.

C) The launcher market just got another large step cheaper (Musk vs Russia) with promises of further price cuts which, if they do not appear, will be just like Pegasus and other launchers that were supposed to revolutionize but failed to.  So we don't know if that step is different yet, but we will in a few years. 

D) The Chinese and Indian economies, as well as other developing nations, can easily absorb 10 times the amount of services that they could have in 1998.

E) We are not on the brink of an inventory surplus from over manufacturing in anticipation of Y2K and then crash of dumping/discounting that inventory (including massive overbuild of fiber capacity) post Y2K

a) So have electronics in ground-based systems, and they are not power-limited.

b) The demand for internet was there, with less infrastructure in place than today.

c) Sure and I think at some point in the future new, lucrative markets will appear. I don't think its satellite internet though.

d) I don't know about India, but countries like China and Russia won't allow their data to go over an American satellite network. The NSA will have access anyway, SpaceX will comply and not talk about it like everyone else. I think you're right though, there is definitely more potential in the developing world than back then.

Btw, Google is part of a consortium that is currently building an undersea fiber optic cable from the US to Brazil for $60m. Capacity: 64 terabits per sec.

[savuporo]

Technology has improved and new players are better financed. Here is hoping it works this time.

Actually the problem back then was the players were not financed! A few milion vs $1B is a big difference.

Uh, no. Iridium spent like 5-6 billions. Globalstar had billions. They were all backed by at the time big tech giants - Motorola was a godzilla.  EDIT : Oh, and Teledesic hit near $1B in '99
Those who cannot remember the past are condemned to repeat it

[meekGee]

b) The demand for internet was there, with less infrastructure in place than today.

I don't think that's true.  The demand today for mobile data (e.g. cars) is increasing, and it is characterized by wide geographical dispersion.  This is a specific kind of demand that is difficult to meet with land-based infrastructure.

Mobile towers are only a partial answer, they too need to link to the backbone, and it is difficult to scale them.

I fully expect, for example, that each and every car will be an internet node within only a few years.

[nadreck]

The whole situation is reminiscent of the 90s. Reduction of launch prices (Russia's entry back then), companies working on RLVs, all kinds of satellite constellations planned (with prominent support as well).

I don't see what's different this time.

A) Well, the electronics are different we have had about 8 - 10 Moore's law generations since then (256 to 1024 times the performance).

B) The market is different, the average consumer of video content watched broadcast television not VOD over IP.

C) The launcher market just got another large step cheaper (Musk vs Russia) with promises of further price cuts which, if they do not appear, will be just like Pegasus and other launchers that were supposed to revolutionize but failed to.  So we don't know if that step is different yet, but we will in a few years. 

D) The Chinese and Indian economies, as well as other developing nations, can easily absorb 10 times the amount of services that they could have in 1998.

E) We are not on the brink of an inventory surplus from over manufacturing in anticipation of Y2K and then crash of dumping/discounting that inventory (including massive overbuild of fiber capacity) post Y2K

a) So have electronics in ground-based systems, and they are not power-limited.

b) The demand for internet was there, with less infrastructure in place than today.

c) Sure and I think at some point in the future new, lucrative markets will appear. I don't think its satellite internet though.

d) I don't know about India, but countries like China and Russia won't allow their data to go over an American satellite network. The NSA will have access anyway, SpaceX will comply and not talk about it like everyone else. I think you're right though, there is definitely more potential in the developing world than back then.

Btw, Google is part of a consortium that is currently building an undersea fiber optic cable from the US to Brazil for $60m. Capacity: 64 terabits per sec.

A) yes but the increase in capability in electronics is missing from a lot of existing satellite systems and makes things that were impossible in 1998 possible today.

B) Lets agree to disagree, the demand for internet was not what it is today and as someone else pointed it it was no where near WHERE it is today. I could list many applications that are possible with the network SpaceX is proposing and more importantly with the potential they are creating for other satellite systems that just weren't possible back then. People are starting to get used to (and not even knowing that they are) the cloud being there for them all the time.

C) I don't see it as the only market, and I do see several niches of satellite internet that, if it creates a service at the costs suggested, will create market that will only exist if delivery of internet by satellite at those volume levels can exist.

D) Ah - I will wager that China, Russia and the non-islamic form Soviet sociallist republics will accept this, I think it will be the hard line islamic countries that resist the most. And what difference does having NSA tapping the comms have for russian commercial use? Their governments will tap them from the ground end anyway. Anyone who wants secure communications encrypts end to end. Then the NSA only knows where your packets went, not what was in them. NSA taps would not block or invalidate TOR for example, or more importantly, those secret systems similar to TOR but that are not disclosed.  So let's agree to disagree here.

Bonus round: Google project between US and Brazil (and the follow on one between EU and Brazil) are great for the end points of the cable but don't help further off. As well cable infrastructure over land is a lot more expensive because of the breaks for router points and the land rights.