I'm in the (slow) process of acquiring and fitting a sailboat for living at remote places. While I acknowledge that the cheapening of satellite internet will be nice, practical, and make my plans more realizable money-wise, I'm also a bit sad on what this implies to the taking a real distance to the information overloaded society.
> Can I travel with Starlink, or move it to a different address?
> Starlink satellites are scheduled to send internet down to all users within a designated area on the ground. This designated area is referred to as a cell.
> Your Starlink is assigned to a single cell. If you move your Starlink outside of its assigned cell, a satellite will not be scheduled to serve your Starlink and you will not receive internet. This is constrained by geometry and is not arbitrary geofencing
It is for the US Army. They are the initial customer anyway. Civilians are the best beta testers. The internet was originally created to have a network in place for the military and I see this as the natural evolution beyond hard wired systems or intermittent wireless towers.
I think it was dreamed up as a way to monetise the increasing capacity of SapceX launches. They might run out of customers and so they need ways to make money out of launching. A constellation of 10k+ satellites is a good way to need lots of launches!
Also a sailboat (even a big one) is not stable and from experience satellite receivers can lose connection on slightly strong winds. So I guess this wouldn't work even if Starlink allowed it.
The starlink antenna may look like an ordinarry satelite dish -- but it is a super-advanced phased array antenna which uses a lot of signal processing to electronically steer the beam. This electronic beam steering should have no problems keeping up with boat motion.
From what I gather beamforming operates only in limited range, i.e. tracking satellites as they move small fraction of a degree per second. For bigger adjustments the dishes turn mechanically. It's not clear if beamforming could handle relatively small waves, but for big ones mechanical adjustment would be needed. I'm speaking out of my ass, but if I had to guess moving the antenna mechanically would not keep the link up in wavy waters.
The beamforming is fast enough. It can probably steer at a few kHz of steering-bandwidth, if not more.
Yes, mechanical adjustment for big waves, but that's a simple passive 2-axis gimbal and having the center of mass slightly below the gimbal axis crossing.
Where did they demonstrate it? Latest news was that they filed for a license [0] to install 5 user terminals on up to 5 Gulfstream planes - application that is still pending.
Speed of beamforming tracking is not the problem - what is a problem is the lack of compensation for movement of the base station (and necessary hardware for it) which is quite evident when people test it in motion [0] - which is quite understandable, SpaceX had to think about the features it actually needs to support to get current price point and use in motion is not the intended target for the current rollout. Tracking is much easier if you can assume that the ground station is immobile.
And no, putting it in a passive gimbal as some of the other commenters are suggesting is not enough to fix this as the direction of the antenna beam is not even close to being parallel to ground (like it is when a gimbal is used with a marine radar).
Plus there is the fact that it does not appear to be sufficiently waterproofed to begin with and the motor mount is rather flimsy (both a problem if you want to use it at sea instead of lakes...).
Marine antennas for Starlink will come, no doubt - but the current home user terminal one is not it.
This is absolutely true as long as they built the IMUs and control loops into the software to do that. My guess is by the time the service is available for moving users that will be the case.
i'm pretty sure boats (especially big ones) are one of the potential markets for starlink, so even if current receiver will lose a connection, there will be special receiver for boats eventually. Stabilizing phased array shouldn't be that hard, even without adjusting motors constantly.
I wonder if they'll reserve a segment for global use, probably at higher expense. I could see a lower tier strata that requires planning but allows for relocation on a planned schedule.
It’s because most of the satellites do not have laser links yet, so the satellite you use must be able to see a downlink station in order to provide internet.
When they have laser links widely deployed, you will be able to move about.
10K users, one ten users per launched satellite, is not "mass market" as defined by the FCC. Latency issues are queuing issues at the scale (and thus load) of users needed to achieve breakeven.
When Verizon 4G first launched I had a 45mbps connection all over silicon valley. Then adoption spread and my data rates on the same device dropped to 5mbps. Loading matters!
Our Wireless NBN install in rural(ish, 1 hour from a capital city) Australia went from 25mbps to 0.1-0.5mbps and has sat there for 2 years until we gave up calling them monthly to complain, and switched to a 4G booster and just tethered to our phones.
Due to blatant corruption there’s zero competition for broadband here, just resellers of the same system.
If Starlink can even sustain 20mbps+ and rolls out to rural Australia they’ll pretty much own the market, let alone similar stories across the entire globe.
Considering it doesn’t require any infrastructure, I can’t see how our tech-illiterate decrepit government can stop it. Just sell the modems online and ship directly worldwide.
> Due to blatant corruption there’s zero competition for broadband here, just resellers of the same system.
Can you tell us more about this blatent corruption? Are you saying corruption has stopped multiple players entering into the highly profitable internet market of the sparsely populated rural Australia?
In 2009 a Labor government planned an Australian "National Broadband Network" (NBN) which was to serve 93% of Australians with direct fiber connections, the remainder to be covered by Wireless and Satellite.
Unfortunately in 2013 the coalition government won power and immediately set about to destroy the NBN.
It has since has been widely reported that this was at the behest of Rupert Murdoch who saw that his Cable network would be threatened by a high-speed broadband network.
Whatever, do your own research. There have been thousands of words written on this.
Is it not ok to ask people making bold statements to backup their claims?
The wikipedia article (which I admittedly only quickly scanned) doesn't appear to mention corruption, murdoch, or newscorp. The other is a hyperbolic opinion piece that doesn't provide any references whatsoever.
Regardless of all this, you completely missed my point. The grandparent was complaining about lack of competition whilst being in a semi-rural (read low population density) and there being only one real option, with multiple resellers. Corruption isn't stopping rollout of companies from rolling out multiple competitive networks to rural australia. There is not going to be any return on investment.
For a country like australia we need internet as publicly owned infrastructure, with "zero competition...just resellers of the same system".
The coalition really ballsed-up what should have been a great resource for the so called "clever country". However, if starlink lives up to the marketing promises "our tech-illiterate decrepit government"s opinion that it would be a waste to provide fibre coverage to 93%* of our population, and their decision to instead to wait for technological advances to provide solutions better suited to australia is suddenly not a bone headed as people make out.
> Considering it doesn’t require any infrastructure, I can’t see how our tech-illiterate decrepit government can stop it. Just sell the modems online and ship directly worldwide.
Make unlicensed operation of the dishes illegal, detect violators using an AWACS plane if detector vans are not sensitive enough to detect beams formed directly to the sky. Even if a country's government can't touch SpaceX directly, the users are still a good target.
I always think about uplinks when I hear about "guaranteed" bandwidths to users, fiber to the home and such. Maybe I can get 1 Gbps to the next hop but can you realistically expect the uplink to be n times 1 Gbps? Also I guess it wouldn't be economically feasible to provide uplinks that can cope with that. Can anyone expand on the technological assumptions (average saturation etc) and the economics of such ISP setups?
Reading it cynically/critically/literally, one could interpret it as counting users, not customers (so a family of four sharing a link would count as four users).
They also mention “thousands of customers”, so I guess that reading may be correct.
Also, as to economics, Wikipedia (https://en.wikipedia.org/wiki/Starlink) says there will be up to 30,000 of these satellites. Even if (from https://ecfsapi.fcc.gov/file/1020316268311/Starlink%20Servic...) “hundreds of thousands of individuals spread over diverse locations across all 50 states signed up to register their interest at Starlink.com even without any formal advertising” turns into a million paying customers, that’s about 30 customers per satellite. Can 30 families in rural locations (so, presumably, not in the highest income brackets) really afford to pay for launching and maintaining a satellite in low-earth orbit? (Real number of satellites probably will be lower, initially, but so will the number of paying customers)
Edit: thinking a bit more about it. These satellites weigh about 260kg. That would be 10-ish kg/customer. They fly, mostly, at 340km, or 30-ish as high as jets. Not trying to model this precisely (to say it mildly), let’s guess, with highly reusable rockets, it’s as expensive to move a kg up a km as with a commercial jet. If so, it’s 30 times as expensive per kg to bring such a satellite in its orbit as it is to use a jet airplane to take a kg to cruising height. So, each customer would have to pay the equivalent of moving 300-ish kg by long-distance jet. That’s 3 personal flights with luggage. Guessing at a satellite lifetime of 6 years, that’s ½ a flight a year. Seems affordable for many in the USA, even taking into account that the satellites have to be built and managed.
30 paying customers over the 5-year lifetime of a satellite is $180k. This is more than the estimated cost of a satellite plus the cost of lifting it into orbit assuming they have a fully reusable launch vehicle.
However, even if they don't achieve that you are ignoring a rather important point: that is, most of the time the satellites are not over the US, and there are paying customers elsewhere, too. Even though the profit is not going to be as great elsewhere (generally, in most of the world you have to pay for spectrum, and the customers probably can't support as high a price), each Starlink satellite will spend roughly 50% of it's time within reach of a populated area it can serve. (Even over the southern pacific, with a low angle to the horizon the satellites will be in range of a populated island quite often.)
I've read somewhere that Starlink has the potential of replacing private optic cable connections between world stock exchanges, providing its customers (read micro trading companies) latency advantage.
Early on there was also a lot of talk about High Frequency Trading firms being a prime customer. Once they have inter-satellite laser links, the latency over large distances can be significantly less than via undersea fiberoptic cables.
I believe this is to do with the difference in speed of light in the cable and also indirect path that light takes as it bounces along a cable. Over inter-continental distances, the benefit is magnified.
Laser links can bring in the HFT traders, but having laser links might lock starlink out of some markets.
Some governments would be happy to let a foreign company run internet services for their people if they can guarantee the data must flow through a ground station within the country. If the satellites have lasers though, the government cannot guarantee that, and therefore may ban the service entirely.
Even if you did line-of-sight links on the ground, and had a "perfect" route that didn't require deviating to avoid any obstacles, the curvature of the Earth means you'd need far more relays introducing latency.
But speed of light in space is faster than the speed of light in fiber making it an overall win. And high-speed traders will pay a lot for a millsecond.
I’ve had StarLink installed for a week now. We live on a floating house in Oregon, and up until now, DSL was our only viable option. That’s been hard given my wife and I both work remotely and do a lot of video calls.
Our impression: it’s fast, latency is awesome, but “downtime” is common and disruptive. Having the internet go offline at random intervals through the day, often for 10+ minutes, is a thing. We expected some of this given it’s currently in beta and it should get better as they make improvements and add satellites.
This is only a dozen users per active starlink satellite. I wonder how many customers each satellite can actually support in terms of both available bandwidth and the ability for the phased arrays on the satellite to target the customer. I can't really see how Spacex will scale their system to reach the millions of customers they need to cover the costs of launching and maintaining all those satellites.
They don't have satellite-to-satellite links yet, so to serve a customer the satellite also has to be able to see a downlink base station below it at the same time, and there's only a handful of those in one geographic region.
If they doubled the number of base stations in a new area (with satellite coverage) they would double the number of customers they could serve right now. So the constraint at the moment is mainly the number of base stations. The thing is once they have the sat-to-sat laser links working, they will be able to scale up with a lot fewer base stations, so there's no point deploying masses of base stations right now.
They are hoping to launch roughly 600 per Starship launch with the launch itself eventually costing roughly 2M. Don’t know how much each satellite costs.
GP is not questioning the feasibility of Starlink, yes you can make it work, yes you can provide 60Mbps links. But there is physical limit per satellite, so all of this goes out the window beyond maybe a few million users?
If you create all this infrastructure (with the impact it has) to then just serve a couple million people then this service will be unaffordable for mortals and limited to the wealthy and most likely the military. [0]
Wifi can only serve one computer at a time, but that doesn't make it expensive. You won't want to use Starlink if you have other options. It's a surprisingly nice service compared to existing satellite providers, but most people will prefer fiber if they can get it. Fiber is faster and most people won't care that they could get slightly better latency by using Starlink. The people most likely to use Starlink are in rural locations where they are less likely to have decent internet available.
There is (practically) no limit to how many satellites you can put in place. You just position the next satellite a bit up-north and the dishes adjust their position.
Practically even LEO ain't "unlimited" space, theoretically an oversaturation of objects in LEO could lead to Kessler syndrome [0]
Practically the presence of Starlight satellites already has an impact on the work of astronomers, an issue that will only get worse as more of them are sent up [1].
That's not to belittle how cool the whole thing is, but it's kind of disheartening how we've already managed to pollute pretty much every place on the planet, and are now on our way to even polluting the space around our planet.
Kessler syndrome might start to be a problem once they reach a billion satellites or so. Space is big, all of these satellites have active control, and if a couple lose active control the orbit is so low they will deorbit in a couple of years.
Kessler syndrome would destroy SpaceX, so they're more aware of it than random internet people.
As an amateur astronomer and astrophotographer, I'm biased against the massive cost of the StarLink pollution fleet to our ability to gaze at the wonders above. I think the Vera C. Rubin Observatory is a much higher priority project for humanity, than StarLink.
That said, I think Elon needs to travel the world a little more. Several of his attempts to solve purported world-relevant issues may be founded in limitations present in the US.
1) I have been to the jungles of Cambodia, trekked across Myanmar, stayed in villages of central Iran, driven across the deserts North Africa. Never have I felt deprived on good connectivity, even in the deserts of Iran. Everyone glued to their smartphones like everywhere else. And I had a pretty strong case of FB-itis myself back then. Where there is a human settlement, there seemed to be a connection. As an anecdote, I recall having Skype calls with my folks from the Shwedagon Pagoda - it had excellent wifi ~7 years ago. Walking across a salt lake one can and should expect to drop our for a while. The only exception I recall is some village in Tunisian Sahara where locals made a point of not being connected (mostly expats). Ksar Ghilane, perhaps?
How is it that in rural US there is no alternative to satellite internet? There seems to be some local systemic failure. The rest of the world will be left holding the bag, so that people do not have to dig down some cables in Arkansas.
2) Building the self-driving car. It already exists everywhere. It is called the metro, tram, bus, tuk-tuk, rickshaw, whatever. It exists, and works great, servicing billions and if you exclude the latter options, safe. But not in N.America, where you cannot even cross a street without a vehicle. Thankfully, this seems to only use up VC money, without any externalities for the rest of the world.
While I mostly agree with you, as I too have traveled a lot (I've been to all 7 continents). I've been to many places with very poor internet.
Still, I think you're missing one very important point. The likely biggest user of this system will be the US military and Department of Defense. Just how they use the iridium constellation, only phone calls aren't enough in the 2020 and beyond.
They need the same system but with broadband speed _internet_. They can not be dependent on some other nations infrastructure and there are a lot of boots out at a sea and places where there isn't a cell tower.
The math just doesn't add up. Mark my words, the StarLink system will never work or be profitable without the US Gov being a major customer.
About the DoD, I think Starlink will be beneficial as UAVs will be able to stream with less latency and higher quality, which eventually means better targeted democracy.
Can you demonstrate that math in a comprehensive manner? I've seen this kind of claim repeated frequently and it's never supported.
> The likely biggest user of this system will be the US military and Department of Defense.
The biggest user of Starlink by an extreme margin will be consumers. It won't be a close comparison.
There will be at least a million consumers eventually using Starlink around the world. The network has capacity to handle that globally. The military is going to push that much data through Starlink? No they're not. A million consumers pushing HD video streams through it will rather comically outweigh anything the military will be doing.
From a business perspective, what matters is $'s, not Gbps.
I'm pretty sure the US military will be paying more $'s than any other country's people.
And it'll be used for being able to fly drones with weapons to kill anyone in the world in 30 mins. With that ability, you can decommission a lot of aircraft carriers and spend that cash on decent connectivity for your drone fleets...
But I think the DoD billing would be different than you describe. Specifically, even if they're not pushing GBs of video all the time, they'd be paying a lot for guaranteed bandwidth at any time. Pure speculation on my side though - mostly based on seeing overinflated requirement docs.
There is some math on the Internet that's commonly linked. With some super pessimistic assumptions, it shows Starlink just barely breaking even (aka, not losing money, but not making much either). Most of those assumptions have already been proven to be false. Doing the math with less pessimistic assumptions shows it to be insanely profitable. I'd do it here, but this has already dropped off the front page so nobody will see it.
You can't really cross oceans with microwaves though. So a London from can gain a competitive advantage against other London firms on Wall Street exchange.
Probably not too useful for high speed traders as they can afford and are technically sophisticated enough to relocate their systems closer to the exchange, but I know that enterprise software that suffers heavily due to latency in APAC regions. Lower-latency connectivity from those regions to EU/US datacenters will be a small game-changer for many businesses.
would that help them though? i mean it appears to me they are already as close as possible to avoid latency and use microwave links and stuff so that is pretty hard to beat i guess. the only thing i can easily imagine being beneficial is getting information across the globe faster but then i lack the skills to see how one profits from that, ie. being closer would still be better, no?
Arbitration. If you can see an asset being traded cheaper in Australia than in the US 10ms before everybody else does - you basically pocket the difference at 0 risk.
Mark Handley has done a bunch of videos analysing this on his channel[0].
The TLDR is that there is a lot of room for Starlink to beat existing terrestrial options IF they can get the laser links working on-orbit. This video[1] shows estimates using what was known about proposed configurations in 2018
> ...jungles of Cambodia, trekked across Myanmar, stayed in villages of central Iran, driven across the deserts North Africa. Never have I felt deprived on good connectivity, even in the deserts of Iran.
That makes me even more grouchy about my current situation :D
You don't have to get too far off a major highway in rural Saskatchewan to have miserable connectivity. You can generally connect to an LTE tower, but between the distance to the tower and how oversold they are, it's a total crapshoot as to whether you'll actually be able to use it. Because the towns are too far away from the switches, the only wireline connectivity you're going to get is 56K dial-up (maybe only 33.6K...).
When they announced the "Better Than Nothing" Beta, I was very excited because "Better Than Nothing" is "Better Than What I've Got" at our rural spot most days.
Edit: I missed your comment in the middle there:
> How is it that in rural US there is no alternative to satellite internet?
Well, the RM (~= county) we're in has a land area of 336 sq.mi. and a population of 275 people (0.8 people/sq.mi.) I have no idea what the population density is like in the Cambodian jungle, but looking at a table in Wikipedia suggests that the least-dense province has about 6 people/km^2 compared to our 0.3/km^2. It's just really really uneconomical to build out wireline service.
Yeah, I never understood this either. If electrical lines, or phone lines offering dial-ups can be laid down, and the major cost is digging, then why can ADSL or Fiber be laid down too?
DSL slows over distance. If you're far enough out, you're basically back to dial-up speeds unless they are willing to build cabinets closer, which no North American carrier I've heard of is still doing for DSL.
BPL (Broadband over Powerline) also requires heavy investment [0]:
- Only low and medium voltage power cables can be used for access BPL.
- Signals need booster equipment to make them travel long distances.
- Transformers, circuit breakers, and surge protectors can interfere with broadband signals.
Again, can't the exact same argument be made for why some citizens shouldn't be connected to the electrical grid?
While I don't dispute that electricity is more vital than a fiber connection, the point I'm trying to make is that this is primarily a political issue and not an economical one.
I don't think anyone was suggesting BPL as an alternative.
Ok, I mistook your mentioning of electrical connection as suggesting BPL as an alternative delivery mechanism.
But again, no NA carrier I'm aware of is pushing out new DSL DSLAMs anywhere in the country or even building out copper capability at all in new neighborhoods. Its a legacy technology that isn't being invested in, for good or ill.
As an amateur astronomer you should be cheering for Starlink.
Yes, the first Starlink satellites were horrible for light pollution, especially during orbit raising. However Starlink has taken measures to mitigate this. They've added an shade to darken the satellites and they've modified their orbit raising path to avoid reflecting as much sunlight back to earth.
Starlink's goal is to be invisible to the naked eye when in final orbit. They're not quite that dark during orbit raising, but they're a lot better than they were 6 months ago. Orbit raising takes a couple of months for 1/3rd of the satellites in each launch.
Another advantage of Starlink is that they are in super low orbits. This means that they are shadowed by the Earth except for the half hour after sunset and the half hour before sunrise. There is so much refracted sunlight during these times that they aren't useful star gazing times for most.
Starlink's competitors (OneWeb, Kuiper) have none of these advantages. Kuiper may add some, but none are announced yet. They have no sunshades, are cubical so orbit raising perturbations don't help much, and are in higher orbits so are visible during the darker periods of the night.
Your shadowing comment depends heavily on location. In large portions (and heavily populated bits at that) of the world they'll be visible all night for 1/3 to 1/2 of the year.
The issue you describe is Elite Projection. Musk has it bad. Eg. He has a problem with traffic congestion when he drives his car, and he thinks solving that problem for him will address real, broad transportation problems. It won't.
I see comments on latency every so often when Starlink come ups. I would love some clarity on how satellite connectivity is expected to perform. Example: Streaming services will be great, but syn ack dependent SAAS apps have to be optimized for local cache and data retrieval? What ways are they looking to overcome latency? I could soap box, but this community is much smarter than me.
The satellites are very low altitude. They are not geosynchronous which would require them to be very far from the earth's surface. Instead they orbit at low altitude and the antenna base station uses a complex antenna arrangement (a phased array) to follow the satellites as they move rapidly overhead. Once the inter-satellite laser links are fully deployed, starlink will be the fastest way to ping a signal to the other side of the planet.
the "original" satellite internet [0] used geosynchronous satellites.
geosynchronous satellites (~35,000km) have an inherent speed-of-light latency of ~250ms. that means even if every other link along the route imposes <1ms latency, you've still got that absolute minimum quarter of a second on every single packet.
Starlink is in LEO (~500km) with a corresponding reduction in the minimum possible latency. that lower altitude also requires much more complex antennas and other ground hardware, compared to just pointing your antenna at a (relatively) stationary point in the sky.
10,000 * 20Gbps = 200Tbps, and that all assumes that the traffic is going in and out directly from customer to internet with none of that capacity used for anything else like groundstation-to-groundstation hops or anything like that.
You'll find a *lot* more than 200Tbps of usable bandwidth in even a moderately sized national operator. At a small ISP (~90k users) we have about 25Tbps of capacity in our equipment.
A single mid-range carrier router can easily handle a dozen terabits; something like a Juniper MX2020 has 80Tbps capacity. A Nokia 7950 XRS-20e has 96Tbps. These are single racks worth of kit that deliver half the entire constellation's capacity.
Headroom is particularly important for usable high speed services as traffic becomes burstier. Statistical multiplexing gets you so far.
Now bear in mind that's with 10,000 perfectly working satellites and no degradation from 20Gbps based on spectrum availability and I'd be surprised if SL carries more than a million or two people - maybe 10 at the highest level.
Don't get me wrong, service for the hardest-to-reach million or two places on Earth with good latency and throughput would be awesome, though I think they have a lot of work to do on their impact on ground based scientific astronomy if they really want to launch and maintain 10,000 sats. But a mass-market service it ain't.
AIUI per-satellite bandwidth is power constrained.
The beauty of a five-year lifespan for the satellites is that when, for example, multilayer PV panels (i.e., more efficient/higher output) get cheap enough, it'll be simple to roll them out.
Likewise with more power-efficent routing chips, and everything else.
It's not 10,000 satellites. Starlink has FCC support (with the ITU) for over 40,000 satellites.[1]
With a steady-state demand of about 10k/year, Starlink will be able to get fairly good prices from semiconductor fabs for ASICs. Is the market for Juniper MX2020s the same size? (I honestly have no idea.)
No doubt Starlink will be researching furiously to optimise transmit power, modulation schemes, antenna design, et cetera as well.
A final point is that Starlink's markets are additive, not mutually exclusive.
They can do low-latency links over the North Atlantic and North Pacific for HF traders/arbitrageurs, and when the satellites are over other parts of the globe, they can serve village WISPs, or individual consumers (and the US military if you insist). The one market Starlink can't serve well is exactly that which is best served by fibre: urban areas in the developed world.
Looking at the business case, the risk is all upside for Starlink.
On 15 October 2019, the United States Federal Communications Commission (FCC) submitted filings to the International Telecommunication Union (ITU) on SpaceX's behalf to arrange spectrum for 30,000 additional Starlink satellites to supplement the 12,000 Starlink satellites already approved by the FCC.
Note that the cost of an MX2020 likely exceeds Starlink's cost of building and launching one of their satellites.
Now there is likely to be an relatively few MX2020s in a mid-size carrier (conceding that Starlink will be launching many more satellites) but Starlink has essentially 0 cost infrastructure for last mile deployment other than the satellites themselves.
But lets say Starlink gets to your upper limit of customers (which I assume they will, eventually, assuming they aren't already up against a Shannon limit on their bandwidth,) that is a billion dollars a month at $100 vs ~1.2 billion in launch revenues total in 2020, not even counting high revenue customers like the Department of Defense.
This is going to be a money printing machine for SpaceX once their main cost is just replacing deorbiting satellites and bandwidth on the ground.
Sure those big routers are costly but they're maybe the cost of a dozen satellites and serve half the total constellation size. Buried infrastructure has a 20-30 year guaranteed lifespan if built with fibre. Routers tend to last 10 years minimum. The opex cost of constantly replacing satellites is going to be pretty substantial.
Also Starlink doesn't have last mile costs as such (they just push that to the consumer) but they do have ground station requirements which will grow linearly with demand. It'll be interesting to see how they manage that as they grow. And of course, they're going to need those big routers too...
Will any one ground link station actually require the big routers though? The traffic will be distributed through each connection point to possibly push the throughput per ground station to a lower level of cost hardware.
It wouldn't surprise me if they put ground stations in every country. Doing a satellite to satellite hop across 10 countries needlessly wastes the limited bandwidth of your satellites. It makes more sense to just use the satellites for last mile delivery and hard to reach places.
A lot of these comments read kind of like the ones when dropbox did a ShowHN. "Why not just use rsync?"
I suspect starlink will change the course of human history in substantial ways. It already stands to be the most positive technological development for my life since I've been living.
At the very least, until the market adjusts, it will provide a verdant field of lifestyle arbitrage.
RE: Starlink changing the course of human history.
Exactly. We've already seen Russian authorities say that people will be arrested for using Starlink. All of a sudden, governments have lost the ability to shut off the internet without some serious jamming.
Just thinking of the post-disaster communications potential is pretty staggering.
(I feel compelled to mention that I'm not a "fan" of Elon Musk.)
I'd love to see a multi-link mesh appear across the world, with Starlink stations, wifi, 4/5G, fibre and cable, etc etc. I honestly think that basic low-broadband speeds should be free to everyone in the world, and so should the ability to serve up content from a cheap Raspberry Pi.
> Exactly. We've already seen Russian authorities say that people will be arrested for using Starlink. All of a sudden, governments have lost the ability to shut off the internet without some serious jamming.
I'm not sure that's the case. I believe you have to licence spectrum in the same way that you would if you were offering a terrestrial based service, and if you don't have a licence then you'd have to turn off the beams as the satellites passed over that part of the ground.
All international law is effectively diplomacy (politics) that has been pre-agreed.
The current status quo (as I understand it) is that the US Gov is unlikely to allow a private company to breach international law by making unauthorised transmissions into another country. It would be a brave foreign policy decision for the US Gov to decide to permit that.
A logical response from Russia or China (or any other nation with the capability) would be to start shooting down satellites. They would have reasonable grounds to do so, and I think it very unlikely that anyone wants to start a space war.
There are 3 billion people without internet. It's not a human right, but it's the next tier.
There's a clear path to hooking many of these people up. For instance aid organisations. Unlike installing the scam computer labs where they can just forget upkeep and training, it's known upfront costs they have to be accountable for. And better, small businesses like cafes.
But the 1% are worried it might make their stars look less authentic and their bored ultrarich reddit life needs drama.
Because in reality they don't want to actually go to space and be with the stars, screw that tech of cheap launches and using space, it's just about the books and TV and the appearance.
People are both awful to others and so uninspiring as a default setting.
[edit] For the Flagged reply, it's not logical. The now is a mass of high entropy. The battle forward is already reducing that. This is about caring about their healthcare as human beings for instance, not their crappy Facebook posts. We are about to be hit by AI to mass spam us all, can't you see that? I understand the 'I like real stars' and 'nice conversations', but this is bigger than that. They are people.
I know this example gets brought up many times but then, did the HNers get this completely wrong? Steve Jobs/Google/MS were right to consider remote storage as a feature and not a product. I see the comments from that time about rsync as a more technical variant of the same argument. Basically the argument is that remote storage alone is not going to cut it.
It is infrastructure expansion. Just like if someone built an electric grid or new road. Plenty of people may be happy with their existing roads, but making a new one opens up opportunities to people who do want to use the new infrastructure.
The counterpoint to that is that Starlink is just the latest in a series of massively government-funded projects attempting to bring global comms coverage using LEO satellites (Iridum, Inmarsat, OneWeb).
Some things about me are probably pretty atypical, but I'll say starlink beats every one of those (for me specifically!) handily. Wikipedia comes closer. Maybe I live long enough to benefit from CRISPR, but I suspect that's for other -- wealthier people. If I end up being able to afford any CRISPR treatments it'll be because of Starlink.
Starlink is going to enable me to own a home without a mortgage. It will allow me to make good money while living in one of the cheapest places in the country. This lets me move the goalposts on success so I have to do significantly less just to feed myself and my family. Suddenly retiring by 35 seems conceivable. I can do work I want instead of what's available. Nothing else on that list comes close for me.
There's a youtube of a teardown of a Starlink dish.[1] A nicely patterned printed circuit board, two servomotors, plastic housing parts and an aluminium foil back shield. There's not even an ethernet jack on the PCB, just a JST connector for the ethernet cable. First time I've seen that level of cost cutting.
Those'd have to be some seriously expensive ICs on that PCB for the bill of materials to come to $1000, let alone $2400.
And ICs are like software, printed circuit boards, aluminium foil stampings, and plastic mouldings. The expensive bit is the upfront design. The more you make, the cheaper the unit cost. EDIT: I see mschuster made the same point.
---
Totally agree about the regulatory hurdles. Those could be insurmountable, especially (sadly) in countries where people would really benefit from having access to better internet.
> I’m sure they want to get that number down before they really go into mass production.
For what it's worth scale matters a lot in pricing - the more you order of chips, the cheaper they get. I've seen 30x differences in per-unit price for 10 to 10k pieces for chips. For PCBs and injection molding of plastic cases, the situation is similar - especially for plastic cases, where the actual material worth of the plastic is negligible, but the setup costs for (sometimes multiple) rounds of milling molds until the quality of the parts is sufficient are enormous, to the range of 80-100k per mold.
One of my favorite YTers lives in the PNW. Previously they had to scrounge up wifi from somewhere or spend hours uploading their video in 720p over their awful "satellite" internet. They are one of the beta testers for Starlink and now they can live-stream in 4k. Sometimes it stutters, but it's still beta.
