I mostly agree. But on second thought, you can pack a lot of encoded data into 192 bytes. Say you had like 20 cameras all hooked up to a local server doing local ML object detection and running Home Assistant monitoring a bunch of binary sensors (doors, flood, motion). You could then have 1 bit be a flag for your top 100 objects and 92 binary sensors and get 25 reports of what's going on per day. Could be useful for remote cottage monitoring or something.

I guess if you have enough power to run a ML server, cameras, and remote sensors, you probably have normal internet though, eh?

bits != bytes. There’s room for 8 times that ;)

OH DANG! Nice. Psh yeah you can run a whole facility operation off that kind of bandwidth.

You can probably do a lot better than your estimation with some clever coding schemes.

Your scheme is only optimal if each flag has an independent 50% chance of being set or unset.

See https://en.wikipedia.org/wiki/Succinct_data_structure for some examples.

Depends how many seconds per day you use them :)

I'm imagining a small, lone device mounted to a post somewhere in the wilderness. Probably solar powered.

This could work for alerting/notifications in such a case.

A larger sensor array installation probably has enough power and space available for better options.

Would this be a good fit for shipping containers?

I assume that most of the shipping process is already tracked, but if the error rate is anything like airports and checked luggage, I can see why someone might want this.

Although, I suppose if the actual ships have internet service, then airtags may be good enough even in this case.

Tracking of shipping containers does sound like a reasonable fit, yes. Oceanic buoys for scientific research, where satellite cover applies, would be another good example.

Having worked in that type of space. We tried for years to get some sort of thing to work like this for shipping containers. As the space is huge and a decent ROI. The problem we had was battery life and endurance. At best we were getting 2-3 years in really good conditions. For some items that works great. But for others not so much. Usually very high/low temps and stacked under 10 other containers and 0 signal and bashed around in a shipping yard. You can not put it in the container as it degrades the signal too much. On the outside it is a harsh env. Tough space to work in. There are companies out there that have other solutions that worked better than what we had.

> You can not put it in the container as it degrades the signal too much. On the outside it is a harsh env.

My first thought was to put the device itself inside but the antenna outside. But I suppose there are reasons why that won't work either?

    put the antenna outside

I don't work in this industry, but container ships carry from 1,000-14,000 containers at a time. Your container is likely to be in the middle of an absolutely massive stack of containers so even if you make your shipping container have great reception, it's still going to be buried under millions of tons of steel and god-knows-what-else.

https://en.wikipedia.org/wiki/Container_ship#Size_categories

I'm not entirely sure it's necessary, either. Presumably the shipping company knows which boatyard/truck/ship has your container, where that boatyard/truck/ship is, and there is some sort of liability/insurance in place and I don't know that entire containers get lost with the same frequency as individual packages. (Those in the know, please correct me!)

> I don't know that entire containers get lost with the same frequency as individual packages.

There are some that fall overboard[1]. I'm sure it's not huge compared to the global volume shipped, but the risk is there.

---

1. https://www.supplychainbrain.com/articles/33002-shipping-con...

While it's certainly not a negligible problem if your container goes overboard, seems nearly negligible as a whole! The linked article seems to suggest that it's a few thousand containers per year out of 226 million. Pretty darn good.

I would love to know more about the industry in general. When you ship a container, what kind of tracking information do you get? Etc. Theoretically it's relatively easy (not easy but... relatively easy) for shipping companies to provide that info (they know that container X is on ship Y and presumably know where ship Y is and when it's expected in port) but, I don't know how that industry works.

I do know that there are a lot of sensors you can attach to individual packages that let you know if your packages have exceeded g-force, temperature, or orientation thresholds during shipping. They're not necessarily wireless or satellite enabled, but when the package arrives, they make it possible to verify that the package has remained within acceptable limits during transit. I imagine at least some of that tech is used in the container shipping industry as well. Maybe.

> I would love to know more about the industry in general. When you ship a container, what kind of tracking information do you get? Etc. Theoretically it's relatively easy (not easy but... relatively easy) for shipping companies to provide that info (they know that container X is on ship Y and presumably know where ship Y is and when it's expected in port) but, I don't know how that industry works.

I haven't worked in the industry, but I have worked for FedEx for a short time. I have to imagine there are at least some similarities.

At FedEx, you're supposed to scan each package "received" at the warehouse. And then scan it again when it's loaded onto a truck.

But when you have so many packages to deal with, sometimes boxes go onto the wrong truck. Sometimes boxes go onto the right truck, but they don't get scanned.

Sometimes boxes are mislabeled. Or misrouted.

Sometimes the recipient goes out of business. Or moves.

Sometimes the label is unreadable.

Sometimes there are multiple different addresses on the box.

I'm sure all that stuff happens to containers as well. Especially when the largest ships are above 14000 TEU.

We went down that route. Your 'easy to install' is now drilling/cutting holes into the container. Then hope your cable does not act as a wick. Cost to install is now dramatically higher. Remember the shipping on these things is razor thin. The shipping company makes it up in volume. So if you go and add 500+ onto install they look for other solutions. The best one I saw was a giant bar code spray painted on the side of the thing and then make it up in the software back end. Low cost, easy to fix and fairly long lasting. Downside is no realtime tracking. Plus the stacking issue is a big issue.

Can the container not be the antenna?

Couldn't you package it up in a waterproof enclosure that is then glued inside the corrugations on the outside of a container? It might still be damaged, but it seems like it should usually survive. Power is definitely an issue, but you can pack a handful of D cells in there and run it in ultra-low power mode to stretch them out.

Of course if someone wants to do something nefarious they'll be able to just bash your device with a hammer to disable it, but that's true no matter where it is mounted.

The more batteries you add the more the cost you add. You also need a bigger enclosure. Someone to put it together, etc. Low power mode for the devices works to a point. But eventually the battery just gives out (age or usage). Which means you have to have someone go out and touch the thing. You need more power to go further with the radio so more batteries. The dudes working the docks probably are not allowed to do it. So you have to send a team out there to fix them up. Oh and that device is 'dead' so good luck finding it in the 20k+ containers in the current yard. You can prevent that to a point with scheduled maintenance. But again you are now pull containers out of service waiting for 'the guy' and your guy waiting for him to show up too, and that is costing money and time.

Malicious abuse is usually found fairly quickly in that industry and usually obvious it is happening. It is accidental and environmental that are the bigger problems. The more you harden for the environment the higher the cost on build, install, and replacement.

There are several key factors that go into making 10,000 of something. I can make 1-2 wickedly hardened items as prototypes. But make 10k of them takes a different set of skills and cost factors. Then you have to pay guys decent rates to go out and install/maintain them. Remember the guys running these companies are very cost sensitive. You can literally wipe out their whole margin on one container if you cost too much.

Endpoint tracking/scanning which they use the majority of the time is shockingly effective. Not perfect but very effective. Tracking the vehicle the item is on is usually more cost effective. For some loads it makes sense to track it that sensitively but most of the time it does not matter. Our cost budget for just materials (no manufacture) was 20-50 dollars a unit. We could not get anywhere near that. Just the radio was most of the budget let alone the cables, board, ics and case and battery. So the project was shelved. There are companies out there that make these sorts of things. But their costs are usually 500-3000 per unit plus install/ongoing costs.

As noted in other comments, shipping container stacks are a brutal RF environment. Have spent quite a bit of time with Iridium SBD in various applications and it is rather sensitive to the surrounding environment. The wider and less obstructed the view of the sky from the antenna, the better.

From a simplistic view - the satellite is flying over and dragging a beam a few kilometers wide over your antenna. The more time it has hitting the antenna, the higher chance you'll be able to get your messages up/down before it have moved on.

If your antenna is buried under a stack of metal and has a pencil thin view of space - you're not going to be getting a lot of data across.

Their twitter account sometimes talks about customers. https://twitter.com/SwarmInternet/

Here's the most recent one they posted: https://www.shapetechnology.co.nz/doc-deploys-remote-shape-e...

$60 / year for the connectivity, that is, $5 / mo. Compared to almost any other solution, it's peanuts, even in an urban environment where other options abound.

There definitely should be a lot of market for for such an extremely low-end kind of connectivity, with small hourly updates. Weather / water / fire monitoring stations, maybe even pipelines (as a fallback for the usual wires / fiber), a backup emergency link for trucks / cars / watercraft / small aircraft / drones. The built-in GPS is extra helpful.

No, this may be huge.

I wish that there was some kind of a lifetime connectivity plan. You buy a modem with a one-time fee, it has a unique ID, and you get to send/receive data at a low bit rate for the life of the modem (or until the satellite service is discontinued).

IMO that would make IoT devices a lot more attractive. The device works for years and you never even have to worry about setting up an account to pay a bill.

Checkout https://www.particle.io/, they have a pretty nice free plan that includes cellular data

If you're willing to be in the footprint of mobile phone operators, a mvno like this [1] can get you to pennies per MB, pay for use. If the T-Mobile/Starlink plan works out, you might even get LTE via satellite.

[1] https://www.thingsmobile.com/business

The limit of 60 bytes of down link is really what threw me off. For IoT sensors out in the wild that I've worked with we've still wanted the ability to remotely update their firmware. A 14kb firmware is do-able for a lot of systems, but if that update fails you don't get to retry for a month unless your firmware is incredibly tiny...

For systems that I've worked with that have firmware that small or even smaller have had power budgets in the <50mA/hour range... This modem uses up to a full amp at 3.3V (presumably only while transmitting but I'd bet its receive mode still busts that power budget in under a minute).

I can't think of the use cases for these modems that use that little data and have that much of a power draw...

I think they could potentially make this cheaper for larger deployments with multicast-like down link packets.

So if someone has 10,000 devices to update, this might be doable with one stream.

Wouldn't help a hobbyist or small batch IOT device.

> I can't think of the use cases for these modems that use that little data and have that much of a power draw...

Large asset tracking.

Why does it need always-on receive? Any number of data acquisition apps can just wake-up on a set schedule, power that modem on, send their data, and get any action/responses at the same time then power everything back down. So the W/h numbers could still be quite trivial.

And presumably much of the power is as you mention transmit, because its talking to a satellite, not a device 1/2 a mile away.

> Extremely niched use cases, e.g. rescue beacon or weather/radiation monitoring far out in nowhere.

That's if you're using it as the only channel (i.e. your data plane). It still works fine as a control plane or a signaling channel, e.g. so you know your remote sensor now has 100GiB of data ready to be collected.

That said, even as a data channel, you can pack a lot of information into a 192 byte packet; and 4800 bytes per day is plenty for some use cases - including one that's been rattling in my mind for years and the best solution I had come up with was a fixed-wing drone loitering over a site and exchanging data over a 2.4/5Ghz link

You could snapshot the internal state of an Atari 2600 25x a day! Such a wealth of bytes.

That's the worst data plan since SMS texts cost $0.10. At that data price they can give away the hardware for free.

Huh? Per month, sending 750 texts @$0.1 would cost $75, vs $5 for 750 Swarm messages ($60 p.a./12)