How VLEO Can You Go?
November 15, 2018 11:03 PM   Subscribe

FCC tells SpaceX it can deploy up to 11,943 broadband satellites — Initial launch of 4,425 satellites to be followed by 7,518 closer to the ground. [Ars Technica, 11/15/2018]. SpaceX "proposes to add a very-low Earth orbit (VLEO) NGSO [non-geostationary satellite orbit] constellation, consisting of 7,518 satellites operating at altitudes from 335km to 346km," the FCC said in the draft of the order that it approved unanimously today. The newly approved satellites would use frequencies between 37.5 and 42GHz for space-to-Earth transmissions and frequencies between 47.2 and 51.4GHz for Earth-to-space transmissions, the FCC said. More details, WP, Elon Musk on making Starlink (YT).
posted by cenoxo (34 comments total) 9 users marked this as a favorite
 
Starlink/Skynet. That name’s a little too close for my comfort.
posted by greermahoney at 11:13 PM on November 15, 2018


That's certainly a lot of satellites. Good that they have "debris mitigation" as part of the requirements for the permits, but I didn't quite get what that entailed. Is it simply a side effect of the satellites having a short lifespan?
posted by Harald74 at 11:46 PM on November 15, 2018


This is your reminder that all the terrible lag issues with satellite connections only apply to geosynchronous orbits at 35,786 km. These VLEO sats will be at 1/100th that distance.

(One side effect of the low altitude is each sat can see less of the ground, so you need a lot more satellites to make a full constellation.)
posted by ryanrs at 11:47 PM on November 15, 2018 [8 favorites]


I think space debris at 335 km will clear out in a year or so just from atmospheric drag. Polluting VLEO is not a long-term problem the way it is in geosynchronous orbit.
posted by ryanrs at 12:09 AM on November 16, 2018 [3 favorites]


It's worth mentioning that the ISS orbits between an altitude of 330 and 435 km, and requires regular boosts to keep itself from de-orbiting. This constellation has an expiry date, assuming there aren't going to be refueling missions to approx 4.4k individual satellites.
posted by figurant at 12:11 AM on November 16, 2018


Depends on their propellant and/or power budget. If they can run an ion thruster, then they can stay up there for quite a long time.

The ISS used to boost about once a month (though it could go a lot longer before falling out of the sky). But once the shuttle was retired, they moved the ISS to a higher orbit (400 km), so there's less drag.
posted by ryanrs at 12:24 AM on November 16, 2018 [2 favorites]


Thousands of them will orbit above 1,000 km so very low drag, which is where probably where the urgency to nail down a deorbit plan is coming from.
posted by RobotVoodooPower at 4:58 AM on November 16, 2018 [2 favorites]


Reddit Starlink FAQ
posted by cenoxo at 5:01 AM on November 16, 2018


Oh bloody great. Now the Musky One will be even MORE insufferable...
posted by Samizdata at 5:17 AM on November 16, 2018 [1 favorite]


There are currently 2 test satellites currently in orbit , reports are they have been used during testing to play video games that require fast internet connection. Probably just a few minutes due to the fast orbit period, the ISS is visible for only about 6 minutes max and the orbits are similar, but it's not just one of the many pie in the sky proposals that have fizzled for decades.

Once launched it will bring cable like (high bandwidth, low latency, non-stupid-expensive) communication literally anywhere. And Musk has his own trucks (Falcon 9).

They do need a large antena, about the size of a pizza box that is outside with a line of sight view. Pricing guestimates suggest that it'll be a serious competitor to the cable companies, a very good thing.
posted by sammyo at 5:24 AM on November 16, 2018 [2 favorites]


Pretty cool. Elon musk is still an ass.
posted by nikaspark at 5:52 AM on November 16, 2018 [1 favorite]


Initially I considered the problem of air traffic control at 335Km becoming very interesting... I mean, how do you put up that many satellites in VLEO and not artificially narrow launch windows for anything flying into geospatial - or creating a shadow and blocking signals, but assuming that isn't a problem... lets do some different math.

The earth's diameter is 12,742K. Min Orbit is 335. Max is 346. That means the Min VLEO Orbit is 13,412K and the max is 13,434 - not terribly different once we calculate the surface area at each 565,115,166K^2 (rounded up) or 566,970,628K^2 (also rounded up). With a total deployment of 11,943 satellites, that means each satellite - if spaced equidistantly would have 47,317K to 47,473Ks between them. In all, that's a pretty good launch window - even if you are traveling at 28,000KpH, but at that coverage, I wonder if we have any three dimensional traffic wave issues where some of these satellites could cluster and service is blacked out in a given footprint. The only way you could avoid that is really to double up coverage, and that's Musk's problem... somedays I wish I had his problems...
posted by Nanukthedog at 6:15 AM on November 16, 2018 [1 favorite]


I'm intrigued by the geopolitical import of this. Apparently they plan to work with governments to essentially disallow connections/censor content based on the geo-location of the ground antenna. But, that is a software switch. Once a system like this exists there is theoretically no reason it couldn't provide the same level of service anywhere - and in fact with orbital patterns it would make business sense to maximize data flow throughout the orbital path.

Also, I assume this means a potential for much better connectivity on airplanes, ships and maybe even from buses and automobiles (if you can accept the intermittent nature of disconnects due to overhead obstructions).
posted by meinvt at 6:15 AM on November 16, 2018


I work in HFT, this thing made the rounds last week. How could this thing possibly pay for the astronomical startup and maintenance costs for such a monstrous constellation?

By being the lowest latency connection between New York and London, that’s how. For better or worse, that is probably worth more money than all the other applications put together. They want to build this to DO THAT, and other concerns are secondary.

For those coming in late, sending signals over fiber only goes at about 2/3 c. So even with the couple of hundred miles detour to get into VLEO and back, the radio signals (which are going at full lightspeed) will get there significantly faster than the fastest fiber connection. This is why there are strings of microwave towers between Chicago and New York, and between London and Frankfurt. But you can’t put up microwave towers in the middle of the ocean.

To be fair, one of the larger secondary concerns for a business plan like this is the massive number of maintenance launches this would require to continue operating, and that would be a lovely business-building move for SpaceX. But there has to be somebody willing to PAY for those launches for that business-building to actually happen, and that’s where the HFT comes in.
posted by notoriety public at 6:35 AM on November 16, 2018 [8 favorites]


Like most of Musk's proclamations, I'm extremely skeptical of this (even though it'd probably be a great thing if it actually worked).

240,000 Gbps is not a lot of bandwidth, when you're talking about the theoretical maximum for a planet-wide network. Bandwidth limits of individual satellites are going to be a lot lower, and I would imagine that it's going to be incredibly easy to overcrowd a single node. The reasonable answer to this constraint is that the system is going to be expensive to access, and bandwidth on it is going to be metered.

Iridium is a pretty similar network that needs to operate within these same constraints. It's eye-wateringly expensive to use (the network would be overwhelmed if it was any cheaper), and is infamously unprofitable. Falcon 9's cheaper launch costs will help a bit, but I think you need to do an awful lot of hand-waving to get to that $50/month cost estimate quoted on Reddit.
posted by schmod at 6:42 AM on November 16, 2018 [3 favorites]


I like the concept of this; it's sort of like someone deploying a new cell tower network, only in space so it works everywhere. I'm absolutely unequipped to determine whether it makes practical or financial sense.

Project Loon is still out there experimenting. That works with balloons at 18km altitude instead of satellites at 340km. So very different, but some similarities in what it looks like accessing it from the ground.
posted by Nelson at 6:47 AM on November 16, 2018 [1 favorite]


At least he's an ass about, like, electric vehicles and space travel and PV, rather than, say, gambling and fracking and Republican politicians. It would still be better if he were less of a cock about it, but as far as billionaires go there are certainly worse ones out there.
posted by Anticipation Of A New Lover's Arrival, The at 7:05 AM on November 16, 2018 [2 favorites]


So - I have been working in Canada's far north for the last 6 months, and everyone (business, government, citizens) are looking forward to ANY level of promised increased connectivity - we have been watching the LEO launch news very closely. The existing options have horrible latency and bandwidth is extremely constrained. While undersea fiber is being investigated, the reality of extreme conditions and distance will make that very challenging.
posted by jkaczor at 7:26 AM on November 16, 2018 [1 favorite]


but I think you need to do an awful lot of hand-waving to get to that $50/month cost estimate quoted on Reddit

You could double-that and in remote areas it would still be a better option than what is currently available . Most of us who live in urban areas, might not realize how bad existing satellite - or for rural areas, WISP - technologies are. And there are typically no competitive options for which provider you can use.
posted by jkaczor at 7:30 AM on November 16, 2018 [2 favorites]


If they can run an ion thruster, then they can stay up there for quite a long time.
ESA were showing off their air-breathing ion drive a few months ago. With one of those, a VLEO sat could stay up indefinitely. Not that SpaceX will be using this of course; I think the short lifespan is part of the goal to improve their flight rates to get into the kinds of economies of scale that would be needed to support something like the BFR.

the lowest latency connection between New York and London
Seems like it would affect latency advantages between Europe and Asia too?
posted by rhamphorhynchus at 8:18 AM on November 16, 2018


A point nobody on this thread has brought up is the FCC demanding that Musk launches half his cluster in six years—and sticking to that point.

A quick BOTE says that SpaceX have to launch at least ten satellites a week to get within spitting distance of this (remember, not all satellites end up in the right orbit or work properly). Even if they're small satellites, in the 300kg range plus a maneuvering bus, that's got to be at least one Falcon 9 per week just for the initial build-out. So this is going to top the entire current planetary launch market in one project, within 5-6 years.

Another point that isn't being mentioned (much) is that Musk is evidently aiming not merely to corner the London-NYC data channel (and presumably all the other major stock/bond markets) but to disrupt the global smartphone roaming market. If you're a US citizen you probably don't travel overseas much, but trust me on this—international roaming charges are a big problem for most of the planet.

I got a lesson in this over the past five years: I'm with Three, a UK carrier, who offer all your inclusive minutes, texts, and data at your standard national rate in about 30 countries. Pre-Trump I used to spend about 6-10 weeks a year in the USA; when I got this tariff it cut my annual phone charges by about $300, with immediate effect. It's wonderful and it revolutionized my approach to international travel ... until I spent three weeks in Canada last month, where data costs CAD $10/megabyte because the Canadian cellphone networks are frankly deplorable.

If Musk gets this working with handsets at even a competitive price point, he will clean up the market among anyone who travels internationally more than about 2 weeks in any year. And that's a big market: think students and tourists, not just businessmen, or people who live in smaller nations with nearby land borders to other countries they need to travel to.

What's the global smartphone data market worth? Can't be less than a couple of hundred billion a year ...
posted by cstross at 8:39 AM on November 16, 2018 [4 favorites]


>the lowest latency connection between New York and London
Seems like it would affect latency advantages between Europe and Asia too?


Yes indeed, and there is money to be made there too, but it’s less compelling for a couple of structural reasons. The biggest one is that there is no overlap between Asia trading hours and US trading hours, and very little overlap between Asian and European hours. That is another trading angle that this could address, because it would be theoretically possible to do microwave towers between the Asian and European exchanges, but it would be a hellacious expensive logistical nightmare to attempt.

Many futures trade nearly 24 hours, which increases the utility, but futures volume is still concentrated close to the market hours the future tracks, so when you go too far away from those hours it is less useful.

To return a little closer to the actual topic, I should add that at our office we didn’t put too much faith in something like this really happening. This network will be expensive, and we doubt that even the HFT firms could make enough off of it to pay for it. But we couldn’t think of any other clients that could even come close to justifying the costs.
posted by notoriety public at 9:27 AM on November 16, 2018 [1 favorite]


That's because you're overestimating the cost. (Supposedly, it's not as if it would be the first or last vaporware satellite license if it ends up on the "shit Elon Musk says" list)

That said, I think there is a very high chance of getting an extension. The FCC has a long history of doing so, even when pathetically little progress has been made.
posted by wierdo at 9:57 AM on November 16, 2018


If Musk gets this working with handsets at even a competitive price point
AFAIK, they're only talking about pizza-box size terminals right now. Still needs a phased array antenna.

Even if they're small satellites, in the 300kg range plus a maneuvering bus, that's got to be at least one Falcon 9 per week just for the initial build-out
My BOTE gets about half that. You're right that they're small birds (386kg according to the FCC filing), but a block 5 F9 can chuck at least 20 of them (FH could do nearly three times that by mass, but it'd be volume limited by the fairing). So, a F9 every two to three weeks. That roughly equals their flight rate from this year, which is already a big chunk of the global launch market, so yeah, a lot of launches.
posted by rhamphorhynchus at 10:03 AM on November 16, 2018 [1 favorite]


Oh this is going to bankrupt him like it did Motorola. Oh, it won't be compatible with iPhone? Hahahhaha. Oh he's fucked.
posted by sexyrobot at 10:13 AM on November 16, 2018


It may not be compatible with the end user devices, but it will be extremely compatible with stationary relays, i.e cell towers and relay stations. This does two interesting things. One, it could remove the requirement currently for large bandwidth providers to lay terrestrial cable between long distances. That is huge. No more trans-oceanic fiber cables to be cut or sabotaged. Two, areas that currently are sparsely populated, which tend to have little to zero cell or data services, could get coverage for a fraction of the infrastructure investment.

More thoughts later...
posted by daq at 10:26 AM on November 16, 2018 [3 favorites]


I don't disagree about this being a vastly better option for rural areas (although terrestrial wireless has made some huge strides toward covering those areas recently), but the question remains about how you make this service accessible to the people who need it, while steering away the people who don't (which you'll need to do, because it isn't scalable) .

As far as global latencies go, I'm curious about how many uplinks/downlinks there will be, where they'll be located, and how much "inter-satellite" routing there will be. In all likelihood, it's going to be far more efficient for the network to offload traffic to the terrestrial internet backbone as early as possible (ie. you probably won't be routing packets from New York to Tokyo via satellite, simply because it makes too many hops and ties up too much bandwidth that could be allocated to local traffic instead).

An interesting requirement for Iridium was that the network needed to be able to operate without any ground stations. You could call somebody around the world, and if the recipient of your call was also an Iridium subscriber, the call could be routed entirely through the satellite network.

This is an incredibly difficult engineering problem, and it's kind of cool that they pulled it off. But, also, that kind of redundancy is incredibly difficult to scale.
posted by schmod at 10:33 AM on November 16, 2018


Oh this is going to bankrupt him like it did Motorola
Likely. SpaceX and Tesla were also pretty likely to bankrupt him. The guy makes big bets.
posted by rhamphorhynchus at 10:53 AM on November 16, 2018 [1 favorite]


One, it could remove the requirement currently for large bandwidth providers to lay terrestrial cable between long distances.

I do not think that is the case. Unless the satellites have very highly-directional antennas for satellite-to-satellite connections, you have a shared-media problem.

There's a reason why point-to-point microwave links, which were the dominant technology for "long lines" communication for a while in the postwar period, fell out of favor with increasing bandwidth demands, and optical fiber took over. Even at extremely high frequencies, the bandwidth is much less than you can pump over optical fiber, and as long as you're pulling one optical fiber you can pull a bunch, giving you multiple physical layers in a way that you can't do with RF.

Quite the opposite; I think this could create a greater demand for backbone connections, because the network will work best with lots of downlink / ground stations, to minimize the number of hops the data takes via RF. Ideally, they would build a network of "supernodes" proportional to customer density which can take traffic off the satellite constellation and drop it onto optical backbones. That avoids the scaling problems that almost always befall mesh networks (saturation around and to/from backhaul nodes). The fewer hops each packet needs to take within the mesh, the more users you can support.

I am not shitting on the project, though. It's ambitious, it's risky, it's going to scare the crap out of the lazy wireline internet providers, who are basically awful in every conceivable way, and only exist because of their corrupt monopolies over last-mile connectivity. To say nothing about the existing satellite broadband providers, to whom basically everyone in rural areas who wants Internet service is locked to.

Even if the network is not capable of actually absorbing all the traffic from residential broadband (which it likely is not), it could save consumers many times what it will cost to launch, just by forcing the wireline providers to drop their prices. We have seen in the past that the cable/telco providers price their services not at the cost of service delivery (efficient market model) but instead at the maximum they can extract from consumers (inefficient / monopolized market / market failure model). The second they get any sort of competition, prices fall—witness what Google has done by going into, or even threatening to go into, new markets. To be able to do that to the entire planet, at once, is amazing.

I hope Musk funds himself by taking up short positions on the existing ISPs; it seems like the sort of thing he would do.
posted by Kadin2048 at 11:54 AM on November 16, 2018 [7 favorites]


Oh this is going to bankrupt him like it did Motorola.

That is emphatically not what happened, assuming you are referring to Iridium's very public bankruptcy. How Motorola came to make a bunch of money on the deal and why Iridium went bankrupt so quickly is an interesting and sordid tale, previously told by others far better than I ever could.
posted by wierdo at 1:07 PM on November 16, 2018 [2 favorites]


wierdo, you left the thread hanging! The story when I worked at Moto seemed straightforward enough; they made a $5B bet that launched late, right into the teeth of GSM taking over most of the world, and had to write almost all of it off when they sold Iridium for tuppence ha'penny. What's the juicy conspiracy version?
posted by rhamphorhynchus at 2:42 PM on November 20, 2018 [1 favorite]


In short (from memory), Motorola wrote off some debt on the deal, but made a significant cash profit with the money from Iridium's lenders, investors, and gateway operators. The space operations service, building the ground stations, and running the billing, among other things, resulted in payments back to Motorola in very large amounts that continued during much of the bankruptcy. In the end, they made back far in excess of their cash outlays on the project, though not nearly as much as they were owed, hence the accounting loss.

There were also the typical shenanigans that you see all the time in government contracting where costs that should be allocated to a fixed price contract are actually billed to the client on another variable price contract that combined with the terms of the debt between Motorola and Iridium along with Motorola's delay in making the constellation actually work reliably meant that there was zero chance of Iridium not failing right out of the gate.

The whole saga is laid bare in the book Eccentric Orbits by John Bloom with actual figures to go along with the rather interesting narrative of how the project came together and the insane arm twisting and strange bedfellows that were involved in keeping the satellites from being deorbited. I have a vague memory of one of David Cay Johnston's books mentioning it as well.

Needless to say, I'd be quite interested to hear any critiques of the work by anyone in a better position to know the story than I do. ;)
posted by wierdo at 7:50 PM on November 21, 2018 [2 favorites]


Turns out my memory was faulty, I read it as an ebook and just now found it. Motorola was owed approximately $400 million on opening day, out of a total of about $6 billion in debt and equity, the majority of which had already been paid to Motorola for building, launching, and operating the space system.

That said, it didn't help that interest rates back then were much, much higher than today. Their outlook was poor regardless of whether Motorola had soaked them for everything they could raise and then forced them into bankruptcy or not. Given the biggest losers were Chase, the House of Saud, a few other banks, and some abusive monopoly telecoms, my sympathy is limited at best regardless of anyone's ethics or lack thereof in the deal.
posted by wierdo at 8:07 PM on November 21, 2018 [1 favorite]


Thanks for the book recommendation. I did try googling before asking, but all I could find was shallow reporting on the bankruptcy and relaunch with nothing in between...
posted by rhamphorhynchus at 6:39 AM on November 23, 2018


« Older “I CHOOSE YOU!”   |   Özgür Baba - Dertli Dolap Newer »


This thread has been archived and is closed to new comments