Swarm – Low cost, global satellite connectivity for IoT | Hacker News Reader

Swarm – Low cost, global satellite connectivity for IoT

swarm.space

230 pointsjsmorph156 comments

SpaceX acquired swarm in 2021: https://www.google.com/amp/s/www.cnbc.com/amp/2021/08/09/spa...

I guess that explains why the marketing site looked so familiar.

Hasn't Helium shown that there is low demand for this type of IoT connectivity?

Helium isn’t truly world wide. Swarm is more like a direct competitor to Iridium Short Burst Data (SBD) service. Works anywhere on the globe. Helium relies on people nearby running a node so it only works where people live and have an internet connection.

It's not even close to being global -

https://explorer.helium.com/

Ignoring exactly how effective a crypto it has been or will be.

Minimum order of 25, so roughly $3725 for one year unless you get the "eval kit" @ $449.

USD $5/MO PER DEVICE

Provides 750 data packets per device per month (up to 192 Bytes per packet), including up to 60 downlink (2-way) data packets

What might somebody (or some company) use this for?

192 bytes * 750 packets = 144000 bytes per month at what speed/latency?

Weather station or any kind of remote sensor that you want to keep in some remote area.

Silence sense and FCC-required logging for a mountaintop broadcast transmitter.

What is "silence sense"?

I'm speculating, but certain FCC licenses require you to transmit without significant gaps on your licensed frequencies or be fined. (The theory being that others could be making better use of your frequency since you aren't using it.) If your transmission gear goes down for whatever reason, you want to notify the FCC before others do. E.g.,

https://www.insideradio.com/free/fcc-slaps-birach-broadcasti...

It's a circuit that alerts people to dead carriers.

People in radio station studios generally don't listen to the over-the-air signal because there is a delay. A silence sense is a circuit that monitors the over-the-air signal and takes action when it's been too quiet for too long. This is usually an indication that something has failed, either at the transmitter, or in the studio-transmitter link. It is sometimes triggered by dramatic pauses in classical music and talk show content, but in those cases is ignored by the DJ/host/producer.

They've been around forever, and can be made from simple analog circuits. In the stations where I've worked, if the silence sense activated, a red light lit up in the DJ booth, and the engineering department. Some stations had a secondary silence sense that would wait a bit longer, and light up a light bulb at the receptionist's desk because she had the master list of phone numbers to call the right people in case of a transmission failure.

There are thousands of radio station transmitters that are far enough away from the originating studio that it's not possible for the studio to hear the over-the-air signal, so a silence sense on top of a mountain, next to the transmitter could send an alert packet via this satellite service back to the studio to let someone know something is wrong.

This service is high latency at present and only supports sending uplink data a half-dozen times per day when satellites are overhead. It would likely not be suitable for this application.

When it takes hours to days for someone to get from the station to the transmitter, the alert latency isn't a big concern.

I worked on a project for a carrier focused on monitoring temperature of seafood (oysters to start) from water to store.

They were anticipating some legislation that would require full auditability of supply chain.

Given the low cost and power requirements of LoRa this could be fantastic wildlife and natural habitat monitoring. Could act as a backup emergency contact mechanism for folks that spend a lot of time away from civilization.

Hopefully some resellers pop up that let you buy them onesie-twosie to tinker with.

Security is an obvious use.

In the normal case, you send a heartbeat once per (interval calibrated to threat model). Should be plenty of packets left to encode "send lawyers, guns and money" when necessary.

In-Q-Tel, the venture capital arm of the CIA, lists Swarm Technologies as one of their startups.

I bet the military/intelligence community has all sorts of ideas!

Not to mention that people could have internet connected crap without even knowing it. No way to block it using PiHole, pfsense or other local routers and access points.

Given there's mobile service where most people are spending time, this situation is already much worse with cellular data b/c of the substantially higher bandwidth capability.

eg https://marketplace.att.com/products/att-iot-dataplans-lte-n...

NB-IoT or CAT-M1/LTE-M are a far better fit for this frankly.

Inqtel also funds Anaconda, Mongodb and Gitlab, interesting

They fund much more than this, and almost all of it is cool. Check out In-Q-Tel Labs, they have some cool projects.

disclaimer: I work with In-Q-Tel investments almost daily, but am not affiliated.

That's a pretty interesting portfolio.

Do you see any interesting trends? In terms of one layer deeper than the general trends such as AI, space, etc.

It's a perfect fit for the Yo app.

https://en.wikipedia.org/wiki/Yo_(app)

Wait, this is a real app?!

https://www.youtube.com/watch?v=OVoFzu-vH4o

I could see this being very useful for scientific beacons, for example to track ocean currents or bird migrations. The latter currently use cellular connections, but can only track birds through areas where reception is present.

If I was asked to engineer such a tracker, I would definitely buffer data inside to flush when connected to celluar network. Not always real time, but much more useful.

I'm guessing you live in a relatively dense country. Since many animals will deliberately avoid more populated areas that wouldn't work many places.

Remote communication in places without cellular connectivity. Air quality, water quality, remote weather stations, aviation, industrial applications... Latency anywhere from a few seconds to minutes. Not used like normal networking, it's something you build around for very specialized applications.

How about for check points on an ultra-marathon desert run, where you are crossing the Sahara. Each check point could have an RFID reader that gives feedback on the racers. If Bob doesn't get to the next checkpoint in say 2 hours, you know he's probably lost.

Unfortunately, command and control for terrorist drone networks, sending coordinates to low-use credit-card gas pumps for night-time refills, and coordinates to targets. Give it ten years.

Hourly data collection for science like stream gages, earthquakes, air quality, precipitation, or similar in very remote areas.

Sensors along a pipeline etc to monitor remote infrastructure.

Texting in remote areas for logging, cattle, forestry services, etc.

Redundancy for boats, emergency services etc that may have primary and secondary communication channels, but still greatly benefit from the capacity for redundant txt messages.

Wouldn't a pipeline be ideal for a wire or fiber just by attaching it along it?

A wire like that is a single point of failure. Great for sending lots of data, but using two different systems is safer than using two different wires along the pipe, both of which could be severed in the same event.

It's ideal for asset tracking - you can fit a GPS fix, timestamp, and accuracy estimate (based on # of satellites visible) into 12 bytes, which is enough to ping once per minute and still only use half your allocated bandwidth. If your asset tags call in every 5 minutes or once per hour lots of options open up.

2 way communications aren't much more demanding because most of what you need can be done with lookup tables, and there aren't that many different commands you would need to send to an asset tag for local scanning. If you want to do person-to-person, 2 bytes per word is sufficient for a fairly large vocabulary (especially a tightly specified one like military brevity codes) - not exactly chatty but sufficient for many operational/emergency purposes.

It's not only bandwidth limited, it's message count as well.

The $60/yr plan allows for 750 _messages_ per month. Up to 3000 messages per month if you stack 4 plans together.

So on the base plan, you're limited to ~1 message per hour if you spread it out across the month.

> ~1 message per hour

How so? 6 messages per day is one message per 4 hours.

24 * 31 = 744, so one message per hour is just under the 750 limit. This is probably not a coincidence.

> which is enough to ping once per minute and still only use half your allocated bandwidth

Err, no - at 750 packets/month, you’d burn through your allocation in 12.5 hours at that rate.

Sorry, I should have said record it once per minute and then upload less frequently (since each packet can be 192 bytes).

A good use of IOT is agriculture. At proper cost, having satellite connected devices across huge sprawling farms would be extremely useful. There are already solutions, like LoRA, but these have their own disadvantages.

I met a person who runs a competitor with similar restrictions.

Their suggested architecture was basically to use this for notifications for very remote sensors: "battery below 20%. Come help" etc

The other way is for things like Garmin InReach: "Send message 1" (which is "I'm ok and don't need help") etc.

This seems like a surprisingly small amount of data in this day and age of connectedness.

I wonder how much total bandwidth each of these satellites can push per month. Anyone know the backend aggregate sustained data rate per sat?

I don’t have the numbers, but at least for Iridium the data service is basically an afterthought. They allocated so much of their bandwidth to voice (using older insufficient voice codecs no less) and are now essentially locked in to a suboptimal bandwidth allocation. They must didn’t realise the importance of data when originally designing their hardware. And hardware in space is expensive to change.

Provides 750 data packets per device per month (up to 192 Bytes per packet)

144,000 bytes. Almost the capacity as a Commodore 1541 disk drive.

40 years ago it would have been unfathomable that a regular schmoe could transmit a floppy disk via satellite for just $5.

I love this type of tech and want to get in on that sweet 144,000 bytes of data at 5 per month, but... it's not just 5 per month, it's a minimum investment of 500 unfortunately.

Also, not sure where to find a decent Commodore system these days to utilize this plan :)

Sure, but 40 years have passed since then and 144K peanuts for any real world application.

A quick example, you have a sensor that sends a single metric every minute. Suppose that you can pack that info (sensor id, metric type and data) in 32 bytes. You're looking at about 1.2MB/month of data (8 times your 144K), which will come to about $40/month.

Quite expensive for a single data point, but I can imagine several use cases where that cost can be easily justified.

Up to the minute sensor data for something so remote that it needs a satellite uplink is either going to be for something incredibly critical or totally overkill and a waste of money. With Swarm, you could be waiting up to 2 hours between passes that last anywhere between 10 and 50 minutes so it's not like you'll have your data instantly for large portions of the day.

Yes, but then IoT usually invokes a sense of (almost) real-time monitoring of things and lots of packets etc...

But I get your point, also you can always do a lot of things to save on bandwidth, like do not send info which is not interesting (i.e. no point in sending 100s of "no fire was detected" messages).

Btw, anyone here knows if there's compression schemes designed specifically for small data packets? Gzip and others would be overkill as headers vastly exceed payload size. Just using raw LZ77 may work but it's 2022 so there's probably a specific thing for that already.

Also, what about data that follows a specific format, like only integer numbers, it would be nice to have an algorithm that takes a "string" of 32-bit ints and gives you back a binary buffer with a smaller lossless representation of it.

Yeah and that's probably why IoT devs absolutely crank up the data rate and spew dozens of Bluetooth advertising packets in the air for no apparent reason, probably significantly reducing battery life that could have been 5yrs, for applications that really don't need instant response.

Something like msgpack might be able to compress ints to some degree since it can represent them as smaller data types.

Thank you for the msgpack reference.

This product is targeted at actual users of "IoT", think remote monitoring in remote locations. Not "smart microwave" style IoT.

Not sure there is any automatic solution for compression here. If you know your own use case you are in the best position to choose where to sacrifice accuracy with a lossy compression scheme. But this relies on understanding of the accuracy of the input data, possible ranges of values, importance of accuracy in different fields etc. Not sure how an automated algorithm could take these real world constraints into account.

A clever compression algo can't help you if you try compress 4 doubles, but in reality you only needed byte precision on some fixed 0-100% range.

> designed specifically for small data packets

Zstd dictionary mode? At small sizes using a dictionary is generally the way to go.

Yeah but once a minute is overkill for lots of things - remote river flood warning every 15 minutes, river water quality once an hour, temp or rain fall once and hour

I don't think the general use case is once per minute though, more like once per hour unless something goes wrong. Something like ensure pipeline pressure is still nominal for remote parts of the pipeline.

> Suppose that you can pack that info (sensor id, metric type and data) in 32 bytes

I suppose you haven't worked with such constraints before, so that's why you think this needs 32 bytes. In reality, 4 or 6 will likely do.

Send diff in metric value if applicable, use variable-length encoding to not waste bytes. Allocate bits and not whole ints to things (like metric type). ID is already part of lower level protocol - the address, so you do not need it

How do you use variable-length encoding in a limited-length package? You either allow a situation where this variable-length encoding would not fit in a package and then it's not really clear what to do. Or you just saving bytes for nothing, because you could use fixed-length encoding instead and satellite does not care whether you sent 65 or 69 bytes of data.

At work we’ve been experimenting with what amounts to a pre-calculated/pre-shared dictionary to allow for compression of small data, but just picking what data you send and it’s representation gets you very very far.

Because we need to put nearly arbritrary data into our NB-IoT UDP socket, Inmarsat or Swarm, saving bytes let’s us pack more data into a message so we can send fewer messages: saving money and increasing reliability (because if you miss a Swarm upload time you can have hours before the next one, for example, or our Simcom modem might hit a black spot and such)

For use cases of this product you will be packing much tighter than this.

You get terminal identification from the carrier, no need to waste precious packet space on identifiers. And it's probably also a waste to encode field names into your packet - just record fields at the same offset every time. If 16 bits of precision is enough for you (it likely is) you can squeeze 16 metrics (or 16 time series recordings of the same metric) into every packet.

Helium charges 0.00001$ per packet (about 24 bytes) using Lora protocol.

But it's associated with crypto so people hate it.

It also undercuts everything by a factor of 100x and is ubiquitous in US and western Europe.

The reason people are skeptical about its long-term longevity is that the vast majority of the income of the project is coming from the sale of hotspots, rather than from paying customers [1]. This means pricing for users is heavily subsidized by hotspot revenues right now. Eventually, the hotspot market will be saturated and prices will have to come up to keep the project afloat; it remains to be seen how competitive the pricing will be at that point.

PS: Saying LORA is "ubiquitous" in the US oversells it quite a bit. It is quite available in cities but most rural areas are completely devoid of coverage, see [2].

[1] https://cointelegraph.com/news/critique-on-helium-s-6-5k-mon...

[2] https://scwcontent.affino.com/AcuCustom/Sitename/DAM/031/Sen...

Oh boy, where to start with all the problems in this post.

* you can buy helium hotspots for $100 now, on par with other lora gateways

* " vast majority of the income of the project" is vague and incorrect. The hotspot OEMs (of which there are 50+) receive sales revenue, not "the project".

* You're referencing a Senet chart, which is actually a roaming partner on Helium. This isn't actually a helium coverage map.

I will correct myself on the coverage, about 98% of Americans are covered, not 98% of land in America. Oops. America is a big place.

Curious how it compares to Iridium, I've seen those with Sparkfun

I don't have numbers in front of me but it looks waaay cheaper. Seems they are trying to undercut Iridium SBD on price.

It's significantly cheaper but also in a frequency range that likely requires a bigger antenna.

Order of magnitude less expensive.

Inmarsat is surprisingly more competitive to Swarm than we expected, but man it is difficult to get passed all the bullshit resellers, which is where Swarm is way easier. But for our use cases Inmarsat has been a better fit right now, though we use both on our board depending on the deployment: if we don’t need as much bandwidth and are okay with the black out times where we’re not covered by Swarm, it’s great

Interesting, I didn’t know Inmarsat offers IoT connectivity, or are you using regular BGAN (SMS or data)?

They do have Elera:

https://www.inmarsat.com/en/solutions-services/enterprise/se...

But we’re using IDP as it’s not really out yet. Next year hopefully!

There's also a sparkfun dev board linked from that page which is around $150. It mentions that if ordering less than 25

Sparkfun has evaluated boards for $149 although they are SparkX experimental and could go away.

Could always do a groupgets anyway.

I love how they define "activation date" for the data plan!

> The hardware will be charged upfront but the subscription data plan will not be charged until the device is activated. The activation date (and $60/yr billing date) for a device is the first of the following month once ≥50 messages are sent. For example, if your device sends its ≥50th message on April 8th, your device would be charged on May 1st. And would renew annually on May 1st until you opt-out.

So 50 messages for testing and calibration before they charge you? I wish all wireless providers were this nice.

I have one of their dev kits (the neat one with a built in solar panel and Wi-Fi connection) - sadly I was unable to get it to send/receive more than 1 message in San Francisco over a 3 month period.

(I live on a side of a hill and had the entire dev kit standing on my roof that was tall enough to have a view over a large part of the city and out into the bay)

At the time, I heard from support that they were launching new satellites that would improve the connectivity in built up locations (such as a city) but I eventually gave up.

At the time, there were 2 satellites a day that were prospects to connect to based on my location, but we met an invisible issue that couldn't be debugged. Potentially it was a satellite visibility issue but could have also been that the available satellite connection was backlogged with other devices / the satellite didn't have a large enough window of connectivity to perform the upload / there was noise interfering with the connection / other mysterious reason that would need expensive test equipment to root cause.

I took it with me into the plains one day (Livermore) and it connected once in the 8.5 hours I was there so it was functional and the RF signal strength was good for the received packet.

My takeaway was that the tech is well suited for rural areas all over the globe where there is no other connectivity options and there exists applications that can work well with the high latency / low bandwidth payloads. Well monitoring jumps to mind as does CO2 gas monitoring in mines - slow moving signals, rural locations, expensive to send a truck to go and monitor.

Have you tried panel mounting the antenna on a big (at least 4 sq ft) ground plane? We use these at work and had a very hard time until we started building nice big ground planes into all mechanical designs (that and reading the design advice on the decoupling network with a microscope).

Huh, for some reason, I thought that the FCC would fine and regulate this company into oblivion after their unauthorized launch stunt. Looks like that didn't quite happen the way I expected.

https://spacenews.com/swarm-ceo-talks-past-mistakes-future-g...

Same, it's weird that this hasn't already been mentioned. They launched untrackable trash into space and got fined for it, nobody should trust this company.

https://www.reuters.com/article/us-usa-satellite-fine/fcc-fi...

It drives me up a wall that they got off basically consequence-free for that--I'm guess they probably made up the fine in reduced time cost spent on authorization proceedings anyway. "Move fast and break things" isn't a great motto anyway, but it's especially bad when those things are floating around in orbit.

I see some possible problems:

- IMHO, price could be prohibitive for many use cases (e.g. water meters)

- Signal strength could be an issue for urban environments (e.g. inside buildings)

- There aren't many small business use cases for IoT connectivity that does beyond WiFi (both wrt range and power budget). What I have seen is that there are usually several big players (e.g. utilities, municipality, factories) that at some point deploy LoRa based or similar connectivity. LoRa can actually cover a huge area with just a handful of self operated gateways (similar to Helium in this respect).

Why would you need a satellite connection in an urban environment?

There are some things that make urban IoT connectivity harder than in the open field.

- Wave paths are more complex (buildings, multi path propagation etc). Hence base stations need to be positioned more densely.

- Finding optimal locations for base stations and contracting the building/land owners is not straightforward.

- Using the less congested, licensed spectrum instead of the crowded ISM bands may be preferable in the city.

This reminds me of Hiber[1]. Hiber launched in 2016 with the mission of connecting the 90% of the world that didn’t have access to connectivity. They were working on HiberBand, a protocol to communicate ultra-low power for remote IoT solutions. It sounded very promising and futuristic, a bit like a mix between Starlink and Swarm of today. At some point Hiber pivoted away from a connectivity-only business model. You could no longer use them for general purpose IoT connectivity, instead they focused on building end-to-end solutions. Hiber today markets itself as an "off-grid asset monitoring" solution focusing heavily on the oil industry.

I suspect there is some underlying issue with these high-latency / low-bandwidth solutions that just makes them impractical in real life.

[1]: https://hiber.global/hiberhilo/

Ultimately what this means is that data transfer speeds will be faster, yes? (That's a question.) If most internet usage is direct to satellite and then satellite to endpoint this would mean that tier one hub networks would no longer be used. So laser replaces fiber for most of the distance, with a peer to peer satellite network that would make outages less common. Which would be both faster and more reliable. The only issue I see is for there to be an intentional or unintentional satellite debris event that would be self propagating, but that issue has been around for a while anyway.

What exactly is the plan here for when a satellite is obsolete or dies? Currently we either have the satellite re-enter earths atmosphere or we use the last bit of fuel to send them even fruther away. But these seems very static

If these truly are just 11x11x2.8cm - then in a few years we will have bunch of super fast and very destructive mini bullets orbiting earth and destroying very expensive space equipment.

Can we please consider the amount of trash we send out in space before it is too late?

I can't get to their documentation but it looks like it's using LoRa for the RF.

https://blog.semtech.com/satellite-iot-qa-with-swarm

Pretty cool. Just checked coverage over my house using their pass checker and it's already ~50% coverage through the course of the day. No good for realtime alerting but good enough for daily reporting.

Yep, it seems like you get a transmit window every ~2 hours max. All in all this service is much more reasonable(cost/antenna size/transmit rate/power) than I expected. You can basically send 2kb/day from a large chunk of the world(with a view of the sky) for $5/mo. Somehow even cheaper than m2m cell plans I've encountered, though granted I haven't looked at services in the past few years.

Helium has a competitive Lora network at 100x less cost, in western Europe and US.

Ducks

What does the data collection end of this look like? In a home IOT setup, for example, you'd typically have a single machine (Border Router in a Thread setup) collecting all the data from each node.

I understand the "field device" -> "satellite" uplink step. But, what's the next step? Another device just for downlink data would eat through the data cap in no time.

You get a REST API to to query to get your data.

Eg GET api.swarm.space/devices/$id/recent

Looks like they have a SaaS product called Hive to poll/push data from devices to your application. I don't see any indication of point to point message passing capability but it may exist.

You dont buy a sattelite terminal yourself , they company has their own terminal and will make it avaliable to you over an API

Downlink is internal to the sat provider, and will just send messages on to you over the ordinary internet.

> Swarm uses a narrow band carrier (about 40 kHz wide), and a chirped spread spectrum approach

Is this LoRa? I guess if it was they wouldn't need their own modems. There have been some hobbyist attempts to put LoRa gateways on satellites and from what I remember it worked reasonably well.

Looks like their documentation site is down. Any idea what kind of bandwidth we're talking about here?

1 kbps

are you sure?

And what latency. "Webhooks" and "REST API" isn't exactly screaming useful duplex communication channel.

High. At least if it's anything like Iridium, probably seconds to minutes depending on coverage. Really not designed to be used like general purpose networking.

Roughly 0.44 bits/s. Data is 750 packets of 192 bytes per month on the $5 plan.

Unrelated to Bruce Sterling's Space Swarm https://readerslibrary.org/wp-content/uploads/Swarm.pdf

You can purchase a single breakout board from Sparkfun: https://www.sparkfun.com/products/19236

Well this is terrifying. Pretty soon you'll have to crack open your appliances to snip the antennas to prevent them from calling home, rather than just not giving them the wifi password...!

Not likely with those data rates, for better or worse

Data will get cheaper, fast

You're already carrying around a device that monitors your location, what you are saying (and audio in the surrounding area), to some extent how you are moving (gyroscope), your data ingress and egress patterns and where you share virtually all your public thoughts (and most private thoughts).

Some people even connect their watches, which monitor (or will soon monitor) everything from their heart rate, sleep schedule, oxygen levels, BMI, etc.

I frankly don't think it can get much more intrusive.

While that's true, I have a little bit of faith in Apple and Google. But how about the $18 humidifier I bought on Amazon?

What data about you could they possibly make money off of, in less than a megabyte, that is worth more than the cost of putting the modules in?

This strikes me as defeatist. You can run GrapheneOS and pipe everything through VPNs for your phone. It's really easy, the phone does all the normal stuff except the constant listening ("assistant") thing, and it's way harder for anyone interested to spy on me. I don't run Google or Apple software.

If someone could get their hands on my data, it'd be a bunch of extra work to determine who it's coming from.

>>Some people even connect their watches, which monitor (or will soon monitor) everything from their heart rate, sleep schedule, oxygen levels, BMI, etc.

You can easily have a watch that does all these things but doesn't share any data, and you can be sure of that based on more than just promises. Open hardware, open software, local data storage and zero-knowledge for anything touching the internet. It's well within our grasp technically, we the people just need to demand it.

People's appliances are almost all within range of cellular connectivity, which is cheaper and higher bandwidth.

LoRa's much lower energy, maybe? I can see cell networks if it's plugged in. Or maybe cell radios are cheap enough that you could put one in and fire it up once a day to squirt out data on how to better sell your kids sugary cereals...

LoRaWAN on the Helium network could be a lot cheaper than a cell network, but coverage is spotty.

This is already possible with cell connectivity. Many cars have this now, for instance. Nothing is stopping the telcos from offering something to compete with this other than in remote tower-free locations.

And heavy equipment had this available years ago. One-time fee up front and good for the life of the machine. Can be installed in any type of machine. Worldwide GSM and can send commands to the machine. Was around USD$1500, IIRC. https://www.komatsu.eu/en/komtrax

I literally watched a large excavator from a dealer break sanctions and show up digging in Iran.

If anyone has a lead on which new cars don't call home lmk

Surprised dedicated ultra low bandwidth satellites are the way to go when developing a service like this versus piggybacking on something like StarLink.

It is SpaceX also, so presumably in some regard piggybacks on StarLink.

No. It was purchased by SpaceX last year, but uses their own hardware.

They will defiantly unify this stuff more over time.

Definitely, it’s on the roadmap I believe, but today they haven’t. We use Swarm in production for some of our remote industrial IoT deployments

No? SpaceX purchased it, so it's theirs now. They may for instance be using shared ground network hardware and software post acquisition. Do you have reason to think otherwise?

I can’t talk about some stuff unfortunately, we’re partnered with them, but they’re not using any of that yet as far as we have been told.

We’re eagerly waiting for them to, as it should increase what we can do with their product!

This was a separate company originally. In this case the RF protocol is completely different (~140MHz vs 12+GHz) as well so you'd have to add hardware to the StarLink satellite to support this (which obv. could be done moving forward)

Any idea what the speeds are going to be, especially when compared to connecting IOT devices to a Cellular network?

Under 150KB per month, $5.

Under 300KB per month, $10.

Under 450KB per month, $15.

Up to 600KB per month, $20.

Note, that's total number of bytes, not bytes per second.

With numbers like that, this seems completely useless except for all but the most wildly specific of applications

this is actually really great for applications like tracking, remote instrumentation, and other services that have a small amount of relatively important data in hard-to-service or critical applications which is not that wildly specific.

Critical applications that are only sending one data point per hour. It's limited to 750 packets per month.

Your hardware can record many datapoints, and bundle them up together into one packet for transmission in one big bundle. If you are reading a float off a sensor, you can get nearly 1 data point per minute, transmitted hourly, with a very naive implementation.

Much more with compression specialized to your use case and clever packet design.

You can also aggregate some stuff on the sensor and the send the aggregations.

What do you mean 1 data point? It transmits bytes, not data points. I could easily stuff 60 temperature and humidity readings into one hourly 192b packet(ie per-minute weather data). That's 120 data points, not 1. And you could easily double that with simple differential encoding or beyond other compression techniques

Yeah what I meant to underline was that for a critical application you won't be able to react faster than one hour in case the data reports a critical issue. So not suitable for security or any application where you need less than one hour latency. I guess if you need satellite coms maybe the sensor is not in a place you can go in one hour anyway.

Lots of sensors with a restricted pipe isn't uncommon - in the electrical transmission substation world they've been dealing with the problem of how to do a lot with a tiny amount of bandwidth for decades. An industrial fieldbus protocol even exists for it called DNP3.

The answer is your end devices only signal on meaningful change, whatever that may be, and stay quiet the rest of the time. How meaningful that change must be depends on how much bandwidth you have. When you get right down to it you can do a lot with a little in the industrial controls and remote monitoring space.

We get so used to overly verbose communications protocols and overly abstracted data structures and methods, it's refreshing once in a while to care about each bit!

You aren't going to be streaming video from the Atacama through this but it is no where near useless. 192 bytes is actually substantial amount of data depending on what you're interested in.

It's for global IoT connectivity, just as it says. That is a very specific application.

In the past this cost even more.

Our use case is industrial IoT sensing in extremely remote deployments.

So this is a division of spacex? Since that's where there "careers" page goes right.

SpaceX acquired them last year.

Wouldn’t ham radio works ? I thought by now those would be miniature enough to be used for IOT ?

Two main reasons:

1. Long distance communication require HF frequencies which requires very large antennas.

2. If you’re talking specifically about the HF amateur radio bands, you may not use those for commercial purposes.

This is not affiliated with SpaceX if I read the site correctly.

Super-interesting though.

Edit: Wrong, see below

"View Open Positions" link points to SpaceX page https://www.spacex.com/careers/?search=swarm

SpaceX purchased them last year.

https://www.space.com/spacex-to-acquire-swarm-technologies-s...

It is a wholly-owned subsidiary of SpaceX [1].

[1] https://en.wikipedia.org/wiki/Swarm_Technologies

Is this a new offering released today?

>SpaceX

Ignored completely. Musk was cool until he opened his mouth all the way.

Musk has literally been saying tons of stuff an talking a lot for 15 years.

Why isn’t this mandatory on ALL planes? A simple device like this (or similar ones) would make real-time tracking of all flights a reality.

We’ll see many new things and interesting technologies in the coming years, like this one and also the one in iPhones.

ADS-B is essentially mandatory nowadays, but it's just a radio that transmits location to others in the same local area. Get enough volunteers with receivers spread throughout the world and you do get real time location tracking of every aircraft. See https://www.adsbexchange.com/

ADS-B is mandatory these days, and Aireon [1] has a constellation of satellite receivers for 24/7 global coverage.

[1]https://aireon.com/

Yet another mandatory cost for what potential gain?

And engine makers and airframe makers already do it, much like heavy equipment makers already do it. https://www.boeing.com/commercial/aeromagazine/articles/qtr_...

This looks like a fantastic tool for ensuring the obsolescence of privacy.

This product is not for attaching to people, but things that you actually want to be tracked.

> This product is not for attaching to people, but things that you actually want to be tracked...

…like people’s cars.

Not that I think the sky is falling and this is the end of privacy or anything.

Can already be done for a few orders of magnitude cheaper (and with several orders of magnitude more bandwidth) with a cellular connection. The people who want to track you/your data aren't going to use satellites, they're going to use Verizon or T-Mobile or AT&T.

Reply on news.ycombinator.com