[HN Gopher] Satellite IoT dreams are crashing into reality
       ___________________________________________________________________
        
       Satellite IoT dreams are crashing into reality
        
       Author : ozdave
       Score  : 76 points
       Date   : 2021-10-20 18:53 UTC (2 days ago)
        
 (HTM) web link (staceyoniot.com)
 (TXT) w3m dump (staceyoniot.com)
        
       | Brajeshwar wrote:
       | I'm nowhere near being an expert, but I have been dealing with a
       | whole lot to do with things about Satellite - Geospatial Imagery
       | and related data. We were intrigued and somewhat happy with all
       | the cube-satellites and micro-satellites being launched. In-fact,
       | we were contacted by many of these Satellite Startups and we are
       | happy that we will have lots of options for our data sources. We
       | could, however, never understood the cost though. We did know
       | that the data is being commoditized more and more.
       | 
       | 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.
        
       | jinzo wrote:
       | I checked the Hiber [1] IoT Solutions (mentioned in the article)
       | that are mostly focused on Asset/Vehicle/Machinery tracking. Even
       | tho their message focuses (among other things) on how cheap they
       | are - I have to disagree. Maybe if they find someone that truly
       | needs global connectivity where 2G/NB-IoT/LTE-M can get pricey,
       | but then someone needing that probably has quite a big purchasing
       | power. Also their claims of easy install and productivity/uptime
       | gains (without a CAN module, as they don't offer it currently)
       | are very much PR speak and don't work like that in reality (I run
       | a tiny asset/vehicle tracking business). Really not seeing how
       | they are even close to being competitive in this space - or able
       | to provide value added solutions later on. There are niche
       | tracking solutions which are a lot better fit than what they are
       | currently offering/focusing on. Something like yacht tracking,
       | you know, where satellite connectivity is actually needed. But
       | what do I know, maybe I'm wrong.
       | 
       | 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/
        
         | delabay wrote:
         | Have you read about Helium? its an open Lorawan network.
         | permissionless and $.00001 per 24 byte packet. coverage is
         | great in the US & europe.
        
       | abeppu wrote:
       | Maybe this is crazy, but should we be thinking of satellite
       | connectivity the same way we think about public utilities?
       | 
       | - 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?
        
         | Klimentio wrote:
         | Do you have the feeling that you miss Satelite or that you use
         | it a lot?
         | 
         | 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.
        
           | wumpus wrote:
           | You can use satellites in many ways without realizing it.
           | 
           | Examples:
           | 
           | * Rural cell towers that use satellite backhaul
           | 
           | * Airplane wifi over the ocean
           | 
           | * Newscasts which use video uploaded via satellite, from
           | places like Afganistan
        
           | ultrarunner wrote:
           | I have two specific needs for remote data connections
           | (uploads in both cases). In one, satellite connectivity makes
           | a ton of sense, and I think we are on the cutting edge of it
           | being viable. This need is fairly tolerant of low bandwidths
           | & high latencies.
           | 
           | 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.
        
         | xyzzyz wrote:
         | > which is operated by the US military with an annual budget of
         | roughly $1B -- and we can all use it for free.
         | 
         | 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.
        
           | abeppu wrote:
           | > Why, though? Why would they be worse? Typically, when you
           | have even two competitors, the quality is improved.
           | 
           | 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.
        
           | wumpus wrote:
           | > In fact, if you replaced tax-funded model with, say, a
           | markup on each receiver,
           | 
           | 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.
        
           | nemothekid wrote:
           | > _Typically, when you have even two competitors, the quality
           | is improved._
           | 
           | 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.
        
             | xyzzyz wrote:
             | > It's hard to argue this is the case, especially for
             | businesses that require a significant amount of upfront
             | capital.
             | 
             | 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.
        
           | kevin_thibedeau wrote:
           | > Why, though?
           | 
           | They'd be subscription services.
        
             | Jtsummers wrote:
             | There is, practically, no model in which GPS subscriptions
             | make sense without enabling bidirectional communications
             | with the satellite. GPS modules (ground side) do _not_ need
             | transmitters, only receivers and, potentially, a decryption
             | module (soft or hardware).
             | 
             | 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.
        
               | withinboredom wrote:
               | Little known fact, if you do have the GPS decryption
               | module, you get way better sensitivity and spoofing
               | protection. But you have to be the US military[1].
               | 
               | 1:
               | https://militaryembedded.com/comms/encryption/securing-
               | milit...
        
             | xyzzyz wrote:
             | Maybe, or maybe not (most likely not, in fact). Have you
             | read the rest of my comment?
        
           | orangeoxidation wrote:
           | > Why, though? Why would they be worse? Typically, when you
           | have even two competitors, the quality is improved.
           | 
           | > 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.
        
             | xyzzyz wrote:
             | In that world, military would just contract with commercial
             | providers and use commercial GPS, same way they already
             | spend hundreds of billions of dollars on products and
             | services provided by private companies.
        
               | jppittma wrote:
               | I can't imagine how one would think that's better than
               | the current situation. Military contracts with commercial
               | providers are an absolute nightmare. Is this not what
               | even libertarians denounce as "crony capitalism?"
        
               | xyzzyz wrote:
               | Most of technology procurement that military does is
               | through commercial providers. This means that if there is
               | a nightmare project, it will most likely be a done with a
               | commercial partner, because most projects are done with
               | commercial partners. On the other hand, there are plenty
               | military contracts with commercial businesses which
               | aren't nightmares. For example, US military has a
               | contract with Samsung for mobile devices. It is most
               | definitely not a nightmare.
               | 
               | Government buying from private businesses is no what
               | "crony capitalism" means.
        
           | teddyh wrote:
           | > _Non-Americans, yes, but Americans pay taxes to use it._
           | 
           | 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.
        
           | abeppu wrote:
           | If competition makes stuff better, and a large company like
           | Google can make one at "a pittance", it's worth asking -- why
           | isn't there a privately created global satellite navigation
           | system yet? GPS, GLONASS, GALILEO were created and are
           | operated by state institutions.
        
             | xyzzyz wrote:
             | Because there is one already, and it's free to use. If
             | government hadn't paid for it, you'd see much more
             | competition in the space. For comparison, ask why no
             | company has built a cheap, fast satellite internet yet?
             | Apparently, some are building them right now.
        
         | bigbillheck wrote:
         | > Had the US not provided
         | 
         | In this hypothetical you'd need to assume that there was also
         | no civilian access to GLONASS, BeiDou, or Galileo.
        
           | varjag wrote:
           | Which is a fair assumption. GPS going civilian in the 1980s
           | was an incredible thing, so was the lifting of selective
           | availability two decades ago. Without American govt spoiling
           | the market there'd be little incentive.
        
         | short12 wrote:
         | Where are public utilities like that? Certainly not the US
        
           | rjsw wrote:
           | The last time I went to the US there were roads, has that
           | changed ?
        
             | lloydgrossman wrote:
             | Last time I used NHS, it was free. Is everything free in
             | the UK?
        
         | abeppu wrote:
         | Ok I described this poorly. That's my fault.
         | 
         | 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?
        
           | colechristensen wrote:
           | > for many of us there's basically one satellite-based
           | location system to which we have access
           | 
           | 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)
        
           | DaveExeter wrote:
           | > 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!
        
             | ultrarunner wrote:
             | What is seen is a generally workable GNSS solution,
             | especially after the removal of selective availability
             | (note: this is a good thing, but by no means necessary;
             | BeiDou still suffers from intentional degradation).
             | 
             | 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.
        
           | musingsole wrote:
           | > we also didn't pay to build several competing overlapping
           | systems
           | 
           | 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.
        
         | wumpus wrote:
         | Electric power from the grid comes in only a couple of ways,
         | but there are a ton of different kinds of satellite
         | communications: broadband in several bands, narrowband in a
         | billion bands, satellites that talk to low power devices,
         | satellite phones in either S or L band, etc.
        
         | Robotbeat wrote:
         | If we had, we definitely wouldn't have gotten the low cost of
         | the Starlink constellation. The industry experts, including Tim
         | F, didn't think such a thing was feasible. I've talked to a
         | Guidance and Navigation expert (a civil servant working for the
         | federal government for space systems) that didn't think
         | droneship landing (which SpaceX relies on for making reusable
         | rockets--and thus Starlink--economical) was possible.
         | 
         | 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.
        
       | bsenftner wrote:
       | The economic advantages of Apple to delay any satellite
       | functionality of their devices until these IoT providers are
       | desperate is the script playing out. It requires no coordination
       | between the larger technology and tech service providers for them
       | to realize the significant value they will realize by being slow
       | to implement satellite related options to their devices and
       | services.
        
         | macintux wrote:
         | This seems like a particularly cynical take. It seems more
         | likely that Apple is, as they usually are, cautious about
         | introducing an external dependency until they're confident
         | about its reliability, quality, and battery life impact, plus
         | how to make the integration easy for users.
        
           | bsenftner wrote:
           | Cynical? yes; profitable? enormously. why partner now when
           | one can own later?
        
       ___________________________________________________________________
       (page generated 2021-10-22 23:00 UTC)