Satellite IoT dreams are crashing into reality
staceyoniot.com
staceyoniot.com
I cannot say much now but we are working on something super exciting and we might be able to turn, tune, tweak and bring some major change to the cost involvement with Satellite data.
- For a lot of public utilities, the point is that building several parallel sets of infrastructure would be really wasteful. It makes sense to build it once, hopefully well, and let everyone use it.
- One might think this only is true of strictly terrestrial infrastructure like phone landlines or water districts. But of course, GPS is an example where we all benefit from a satellite constellation which is operated by the US military with an annual budget of roughly $1B -- and we can all use it for free. Had the US not provided for civilian access to GPS, and instead there were several competing private systems, they would all be worse and every application which is today based on GPS would have an additional monthly fee to access whichever private constellation that application had built around.
Admittedly, satellite connectivity comes with a bunch of complexities that GPS don't have. But still, if we were smart, would we be building one giant common-access constellation rather than multiple competing ones?
Because while i agree on GPS and all the others (and it happened anyway), i'm lost on why i would care about satelite?
I do prefer optic fiber when possible.
Examples:
* Rural cell towers that use satellite backhaul
* Airplane wifi over the ocean
* Newscasts which use video uploaded via satellite, from places like Afganistan
The other need requires low power, high bandwidth and low latencies. Satellite is completely nonviable as of now. We might need an upgrade in the physical laws of reality to make it viable.
I imagine 90% of most people's usage would be closer to the latter case than the former. Denser constellations might make the former more palatable in the long run.
Non-Americans, yes, but Americans pay taxes to use it. In fact, if you replaced tax-funded model with, say, a markup on each receiver, Americans would pay less than they already do, and only those that use it would pay for it. I find both of these things beneficial.
> Had the US not provided for civilian access to GPS, and instead there were several competing private systems, they would all be worse
Why, though? Why would they be worse? Typically, when you have even two competitors, the quality is improved.
> and every application which is today based on GPS would have an additional monthly fee to access whichever private constellation that application had built around.
You are lacking imagination. It’s like saying that if government stops providing internet connectivity for free, every application accessing the internet will charge monthly fee. Alas, government does not provide free internet, but nevertheless, apps don’t charge consumers for ingress/egress.
As it happens, businesses are pretty good at coming up with business models convenient to users. Google could, for example, easily cover the entire cost of GPS completely on its own, and offer free Google Maps, which would easily outcompete other map apps if they charged money for use. Considering how much they pay for being default search engine on Apple, $1B is a pittance.
It's hard to argue this is the case, especially for businesses that require a significant amount of upfront capital. The trivial example in utilities are ISPs. You can argue that ISPs are not a "free market" because of the government doing X, Y, Z, but if it's the case that government intervention is required anyways, then removing the profit motive and having the government just build seems preferable.
>Considering how much they pay for being default search engine on Apple, $1B is a pittance.
The $1B number is a red herring. It cost them $1B to maintain the system. The costs of acquiring rockets and designing satellites, as a private company, is likely far greater. Secondly in the case of Google Maps, that product doesn't exist as it's own standalone business, but as one that was largely subsidized by another, more grossly profitable business. The question you should ask would if a company like Waze (pre-acquisition of course) could have built the infrastructure for GPS. The answer is likely no.
Edit: Thinking about this more, there is a trivial counter-example to your point about being charged egress on the internet, and it's just Google Maps. Google did the work of mapping a huge amount of the US, and generally has the highest quality maps. If you wanted to use that data, much like a business uses GPS, Google charges an fee which has been often derided as exorbitant. Even in the face of new competitors like Apple Maps and OpenStreetMaps, Google has even increased prices. I feel we would be worse off if GPS was privatized like this as this would put companies like Uber, Flexport, and Waze at the mercy of a single profit driven "competitor". The "App Store" but for GPS doesn't sound like a better alternative.
There is plenty of capital to be had these days, and a lot of competition in many capital-intensive industries. For example, we have multiple competitors in cloud computing space, despite it requiring ten of billions of dollars in capital to enter.
The ISPs, and utilities in general are different not because they are simply capital intensive, but rather because they have low marginal return for marginal unit of capital spent, and because competition damages these return even higher. Launching a few satellites to provide service globally is a different sort of capital expense than painstakingly building out utilities every street and every block.
> The $1B number is a red herring. It cost them $1B to maintain the system. The costs of acquiring rockets and designing satellites, as a private company, is likely far greater.
Sure, which is why there has been no competition to GPS so far. However, SpaceX, now that it already has rocket and satellite expertise, could provide competition to Government GPS relatively easily.
> The question you should ask would if a company like Waze (pre-acquisition of course) could have built the infrastructure for GPS. The answer is likely no.
No, but they would simply buy license to use GPS, just like they bought license to use maps. No reason to pass these costs to users directly as a line item in a bill. That’s my point: there is no reason to believe that privately owned GPS would diminish the user experience. As I already mentioned, instead of subscription fees that people lacking business imagination keep bringing up, you could for example charge royalties per GPS receiver sold, you know, the way Android phones already pay dozens of dollars in royalties.
They'd be subscription services.
A subscription GPS system with bidirectional communication (to authenticate users) would have grossly inefficient devices (transmitting to space is not cheap, plus the incurred latency on the communication). But that's the only way to enforce a subscription.
Subscription GPS systems with unidirectional communications (down only) would be cracked in about 5 minutes. Every authorized device would have to share the same key for it to be practical, which means even with key rotation schemes (sucks to be in the wilderness without internet connectivity for an extended period in this case) the keys would be on every device and trivially recovered.
The only way a unidirectional communication scheme works is if the GPS devices are already internet connected on the ground side, which largely defeats the purpose of GPS in many situations (at least historically, before widespread cellular data networks). You certainly couldn't use this effectively at sea or when hiking through the backcountry.
1: https://militaryembedded.com/comms/encryption/securing-milit...
> Typically, when you have even two competitors, the quality is improved.
If those state services weren't available for the public they would exist anyway, they are primarily for military purpose.
Imagine the US didn't offer free GPS and civilian competition would arise. Wouldn't it just mean the US users would have to pay for two satellite services? Once for the military one, once for the civilian. Since cost does not increase with users, the military (tax supported) one wouldn't be cheaper to run either.
Apart from that there are competitors even today. My smartphone does connect to Galileo, Glonass and GPS.
Government buying from private businesses is no what “crony capitalism” means.
I think the key points are:
- because there would be fragmentation, we might have several smaller private services, each with crappier coverage because of fewer satellites in each, different variance in clocks, etc
- because GPS _doesn't_ need to worry about collecting revenue from users, you don't have to prove you're a subscriber to use it. Every GPS device you have doesn't have to be linked to your account. And you don't have to worry that e.g. your phone and your car partnered with different satnav services and so now you need two subscriptions _and_ they don't play together, etc.
I think it's not that hard to imagine casual consumer applications where the product benefit of incorporating satellite-based location services would become swamped by the overhead of connecting to your preferred location service.
GPS isn't technically designed to do that, the open signal is open. And if you did it, the cost of a receiver would go up because it would have to be more sophisticated.
Galileo tried to do this business model and failed.
Yes, but non-Americans also pay, by living in a world with America, in particular the US military, in it. So the cost is spread fairly evenly, I’d say.
> if you replaced tax-funded model with, say, a markup on each receiver, Americans would pay less than they already do
What is that based on? Americans pay a few dollars per year.
I think the fixation should be less on the 'free to use' portion of GPS, and more on the fact that for many of us there's basically one satellite-based location system to which we have access, in the same way that there may only be one set of phone landlines in my neighborhood, and one municipal water supply. Phone service and water are not free to use, but we also didn't pay to build several competing overlapping systems.
I compare this to cell service, where multiple private networks covering the same area are built and maintained, and if you're on the wrong network, you might have worse coverage in a given neighborhood than your friend -- and you both have worse coverage than you would under a hypothetical network which was the union of both.
The original article was about a plethora of companies who had aspired to create their own constellations and are now seeing that actually that's really expensive. We're still going to end up with several competing ones. Would we be better off (i.e. pay less for connectivity from a constellation with more satellites and better coverage) if there was one larger one?
We certainly would not be better off.
Because with a monopoly, you pay the highest price and get the worst service!
What is not seen is a potentially better service that doesn't cost $1 billion a year which wasn't created. GPS radios are fairly power hungry, slow to sync without ephemeris, and accuracy can vary especially near buildings and terrain. But it works well enough and is entrenched enough that a private improvement would struggle to make financial sense.
A utility of this scale also creates a single point of failure, which came to bear during the Galileo outage a few years ago. What some call needless competition is seen by others as redundancy & robustness. The water provider in Flint, MI should have gone out of business and been replaced for pumping lead into peoples' homes, but instead the situation became a political football. The only cost to bad actors was the expenditure of political capital; all but one minor charge was dropped.
There are lots of trade offs, so the ways we'd be better off change depending on the way the system is defined. I do think the consideration of what might have been is interesting nonetheless.
I think you're missing the history of utilities but also most public granted monopolies: they did grow out of several competing businesses. In the case of electricity, water, subway systems, most of these institutions will trace back to a history of privatized providers competing to establish a market. These utilities having a monopoly from the jump in any new developments is only because we've inherited that history and know better. It's only in the aftermath of providers clashing to an annoying enough degree that granted monopolies get created.
This isn’t the case though. Most “GPS” (GNSS is the better acronym) chips these days support several.
American GPS, Russian GlONASS, European Galelio, and Chinese BeiDou all provide the same basic functionality, the first three are quite common in most chipsets (like the one in your phone)
In this hypothetical you'd need to assume that there was also no civilian access to GLONASS, BeiDou, or Galileo.
When you have one big system and no competition, you get status quo solutions. That is, an order of magnitude greater cost. (And this isn’t an exaggeration… compare space shuttle costs for commercial satellites to Falcon 9. STS roughly is 10 to 100 times as expensive per launch with comparable payload.)
But it’s important to keep in mind that the satellite networks are different from GPS and one important way: they are two way communication systems not broadcast. That means that you don’t get the same economies of scale for having just one system that you might for GPS. Having two communication systems that use two different spectrums means double the throughput that a single system would’ve provided.
GPS, moon landings, space stations, space shuttles, and most of the rest of the history of space are hardly status quo solutions! It's not like space has been stuck, waiting for Elon Musk to save the day.
Could you provide an example of someone saying that here? Elsewhere?
> in certain social circles
Is it possible that people have rational ideas that disagree with yours, which should not be dismissed as outcomes of social effects? At best, Musk's personal following seems like by far the strongest social phenomenon involved.
> It took the non-status quo companies like SpaceX (and Masten Space, etc) to prove out reuse.
That wasn't the prior point and therefore wasn't what I responded to. The prior point, as I understood it, was that government is unable to innovate and only business is. Also, SpaceX is now part of the status quo.
But again, that’s precisely what happened. There was a decades-long, post-Cold-War stagnation in the government-led space industry (to the extent that NASA used a higher inflation rate to calculate aerospace project costs… ie aerospace technology was becoming MORE expensive over time for the same capability), particularly for launch technology (commercial launch was fleeing the US and going to cheaper foreign launchers like Russia), and it was SpaceX who ultimately broke the impasse and proved reusable launch was feasible. It is a huge impact, too. Consider SLS or, more to the point, the reference HLS lander versus Starship, which was selected for NASA’s commercial HLS bid. Starship is approximately 10 times larger in capability than the reference design or Apollo. NASA literally couldn’t afford any other option for Artemis to meet the 2024 landing goal.
SLS vs Starship is a very illustrative comparison of what a single, monolithic, government-led solution vs a more competitive, commercial led solution is.
(And government vs commercial ISN’T a good framing, here. NASA has been super important in commercialization of this stuff by fostering competition, serving as an anchor customer for new entrants, and providing valuable insight and consultation. And even in the Soviet era, there was competition between firms in the Soviet launch sector… single, monolithic solutions are just not a very efficient approach…).
Yes. This is something a lot of people are interested in. https://muskdeer.blogspot.com/2021/10/rethinking-gis.html
In fact India's Remote Sensing Data Policy has specifically recognized this data as a 'public good'.
Also looks like you have to sign a 5 year contract, very optimistic from them.
Generally thinking the IoT and satellite connectivity (at affordable, non Iridium level, pricing) have it's uses. But for general vehicle/machine tracking where you usually have power, need to extract data from the machine anyway (for real insights), a plain old (or even better new NB-IoT/LTE-M) cellular connectivity will (generally) be plenty.
[1] - https://hiber.global/
But it's NOT free. The fixed or variable costs might be lower than typically seen before but "smaller" ≠ "zero" just like "large" ≠ "infinite".
It's confusing these that causes most of these types of "entrepreneurship Epic Fails".
You still have to have an infrastructure which has non-zero fixed and variable cost.
When it comes to space (I used to be employed as a rocket scientist at a government Think Tank), the infrastructure costs are absolutely enormous even compared to more modest terrestrial systems. Like many orders of magnitude higher than, say, the entire interstate highway system. And it's because the risks and capital costs are that much higher.