The uninvited Internet of things
lwn.net
lwn.net
However, sometimes it becomes unavoidable. After a recent diagnosis with sleep apnea, I got a cpap machine, which has a cellular network connection in it, which connects to a centralized service where the sleep technicians can tune the device to deliver the right pressure for my sleeping patterns. What it also did was allow my insurance company to know how much I was using the device, in order to decide whether it was going to pay for it or not.
So, I really don't see how I could have avoided this IOT device and received treatment.
On the other hand, no hacker is going to be able to compromise my network via this device, so maybe this truly is a better path forward for IOT devices?
But on the other hand, your network has already been compromised, in the sense that you have a connected device within your home that you cannot control at all, that connects to the internet as long as it has power and can do whatever it is programmed to do...
Maybe for a sleep apnea machine this isn't catastrophic, but consider this Ubiquiti hack that's in the news; it wouldn't surprise me if the security situation around medical stuff isn't any better than with a networking company, after all.
Me, I want to control what's connected in my house. At least a little.
The insurance company will also be a lot less antsy about monitoring if you buy the machine out-of-pocket, which brings up another issue with pervasive IOT: opting out is, if even possible, more expensive, making it potentially yet another modern method of class striation.
Do the insurance company's intercoms also send you recordings of their meetings so you can check if they're conspiring to illegally deny coverage, or is the spying strictly one-way?
Are the newer machines not using SD cards anymore?
Also I have always wondered if anyone has opened up and tinkered with the ResMed wifi attachment. Perhaps it could be used as general purpose hardware.
To really hammer home what I'm talking about, if tonight you extracted your data from that SD cards and you published the recordings from your last 10 nights of breathing with your machine you'd create one the largest open respiratory datasets online. It'd take 1 to 2 years for a medical researcher with industry or hospital backing to go through the process to have the right to gather this kind of data on a patient.
I too have such a device. I did a little (<5 minutes) research and was able to access the device's settings to disable the cellular modem.
The manufacturer complained about this to me and I told them to go fuck themselves. My insurance company never even asked about it.
If one is already building a new house from scratch, it would seem trivial to make the whole thing a Faraday cage while you're at it, no? Am I missing something?
Retrofitting an existing house would be harder, but still doable if one is committed to the cause, like a couple thousand bucks tops? (Guessing here.)
It doesn't seem that one would have to give up ANY actually desired functionality. Wired internet piercing the cage is not an issue, right? After that you can just have your own secured wifi inside. If one is okay with using VoIP for voice calls, the problem is now solved, yes?
Add a second AP outside if you also need Wifi in the backyard or whatever.
Still though, so far so good, I think?
Or is that just in my bug-infested homeland?
The windows pictured seem to have the cage wires inside the glass.
Wifi hotspot would have a password, so only trusted devices can connect.
Maybe Apple devices can already do that?
(don't actually do this).
I want a way to spoof GPS signals within my own home such that devices e.g. my phones cannot geolocate themselves using GPS, or get a reading with a pre-defined wrong location.
At the same time, I don't want to affect devices outside my home.
It doesn't need to fill my home either, I'd be happy with a PCB that be stuck to the back of my phone or in a backpack that does this for a 30cm radius.
Or maybe I should just desolder the GPS chip, if I can figure out which one it is.
Tencent does some horrible things with software beyond the laws of China. Apple does too -- that's why I don't use an iPhone as my normal phone. But I have to use an iPhone for certain test apps and I'd like to be able to spoof GPS for Apple and Tencent at the same time.
I'm much more concerned about these companies invading my privacy than the US or China government.
Also, I don't actually have an option if I want to maintain my social life.
But you really don’t want to do this in any way that could possibly leak out. You will make a bunch of connections you never wanted to make.
As for jamming it, I think most GPS receivers probably saturate at around -60 dBm in band, so you can probably just strap a noise diode with a suitable filter on the back of your phone or something...
Alternatively, I could just move my fake location a few blocks down the road and it would be enough to satisfy my privacy wishes while not being too much of an outlier.
Still though, I would like to at least /try/ it rather than just sitting here and doing nothing about it.
You might be able to set your bssid to all zeros or something to fight this tracking. Not sure if this would lead to your devices no longer associating though...
Some time back, I was helping prep a uBlox MAX-5Q receiver for a high-altitude balloon flight. At altitude and with no other noise around, we figured we could get away without an amplified antenna, so we had a little passive helix design. But before attaching that, for giggles, I just soldered a piece of plain wire to the module's antenna pad, trimmed to what I eyeballed as probably roughly something approximating a quarter-wavelength at 1.5GHz-ish maybe.
It got a 3D fix within moments. Indoors. On a workbench under a mountain of test equipment. With a ton of RFI around. In an industrial building with 2 layers of roof. With a paperclip for an antenna.
And that was a 5th-generation receiver. We're up to the 9th generation now and they've only continued to improve.
Yes, most cellphone locations are wifi geolocation, because it takes less power. But even if you can render that moot, the GPS signal itself is plenty strong to receive indoors. And modern chipsets are tracking at least 3 and maybe 4 constellations, giving them a lot more satellite options, so there's almost always a good set overhead at favorable angles.
Did your receiver work at high altitudes? I guess ITAR rules have been relaxed so GNSS receivers are allowed on balloons with no problems nowadays.
One of these days I'm gonna sit down with SatGen and figure out how to make a track that rises to high altitude, then simulate it with gps-sdr-sim and do some ITAR testing. I have a HackRF with a stable enough clock, and I have the shielded test enclosures, I just don't have a good way to generate a track that's not at sea level.
As for your results... Those are some pretty fancy receivers, do I know you as a fellow Galmon contributor? There's an F9P on my windowsill as station #41, and I had an M8U on my desk last week. Each has gotten a fix just fine sitting on the desk, by the way.
In your office I suspect the glass has a low-E coating which is metallic, and effectively forms a skin over the whole building. No explanation for the house though; I see moderately slower TTFF times and sometimes poor C/no stats for certain parts of the sky, but otherwise things work just fine in my house. Is there a way to get a background noise level rating out of these receivers?
Yeah, probably my office building has some sort of shielding... I barely even get cell reception in my office, even sitting by the window.
It worked better in my apartment, but I had to basically hang the antennas off my window. I have since moved from that apartment and have not tried in my new place, although I suspect it would be worse since there's another tall building across the street. I did (and still do) live very close to an elevated train line, which noticeably produces tons of noise (as in my TV antenna cuts out...) whenever a train goes by!
A while back, my family living in the countryside of Europe were upset about a cell tower going up right next door. My uncles are a physicist and a microcontroller chip designer respectively, so after trying to fight the build for a while, they worked out at the time that the 50cm distance in the rebar in the house construction would work as a somewhat effective faraday cage for the frequency bands of the tower. I'm not sure if that was actually that effective, and is likely no longer the case, but I do remember having pretty nasty reception indoors in the early days.
I can easily imagine that it will be harder for you to get the stuff repaired/supported, as in "I cannot see your dishwasher in our system and it's your problem".
I'm not saying you should not do that, I'm saying that it has little power compared to political action.
In any case, I suppose GDPR is a hopeful sign, but here in NorAm it actually seems more plausible to become an oligarch first and then make one's political wishes known then than to have any kind of political impact as a middle-class citizen.
At least anecdotally, when I talk to an average European vs US American, what e.g. their views about what restrictions for businesses are acceptable or even desirable, I get very different answers.
In addition to migration patterns that selected implicitly for such folk, this happened internally to the US as part of the movement to colonize ever further west.
Now, there hasn't been anywhere else left to go for a while, but society hasn't really internalised this yet.
I'm going to need you to expand on this point.
I'm gonna go with: I can choose not to work for my "owner". (you can argue if you want to be extreme that I'd have to leave Europe, but I could do that.)
Also, I get healthcare whether or not I'm employed, so that's nice.
Well, that makes sense, you're a serf to the State, not a corporation. Does the State give you freedom?
As for if the particular state I live in gives me freedom - yes, I think so. For context, I'm in Germany but not a German.
(Currently, due to being on a visa, I can't really, but:) I would be comfortable quitting my job without anything lined up if I were a permanent resident or citizen.
I feel comfortable being an atheist (which is not the case where I'm from), I'm happy speaking out, and I'm happy to have two way conversations with people who have different opinions. I do not need to toe any line to be kept safe in society or not be persecuted by the state (edit: extreme views aside)
I'd say I'm free, yes.
On to your point about being a serf to the state rather than a corporation: do you really prefer it being that way? Has the last year of unemployment, etc, not highlighted a weakness in having things so tightly coupled to having a job?
All said - I think there's probably a middle ground where things makes sense, and many "western" countries fall in there with the state providing the basics and private companies filling in the gaps in demand, the only problem is when that gap in demand is something that's basically a human right.
It's preferable because corporations die at a faster rate than States. Personally, I want to be a serf to no entity.
https://www.visualcapitalist.com/ranked-the-social-mobility-...
Both of these sites show the US as being behind those countries regarding social mobility.
So no, it's not possible to let some things through. Not even your phone will get a connection. (Besides the ones that originate inside the house, like wlan)
The big downside to my approach is heat dissipation and servicing.
Putting shielding foil in the walls is quite straightforward. Putting mesh in the windows and doors, or using solid metal doors, too.
Connecting the moving parts of the windows and doors, to the walls, in a continuous way that maintains a useful amount of RF shielding, is absolutely devilish.
Basically you end up with conductive doorjambs, which can't be aluminum because it makes a thin oxide layer over itself which is not conductive. They can't be stainless steel, which does the same. They can't be plain steel, because it rusts, or copper, because it tarnishes. You're left with brass, gold, or silver as your options. (Or, try aluminum and just use massive contact areas, so the joint functions as a capacitor across the oxide layer, which is effectively a conductor at RF even if not at DC.)
The corners of everything are murder. The soft flexible bits of "weatherstripping" are most subject to damage, and everyone who opens the door has to be trained on what parts not to touch, or how to re-fluff and repair whatever happens. You can make a notched overlapping sort of structure, but you end up with a 4-foot-wide door that opens on an articulated hinge to reveal a 3-foot-wide opening, or a bank-vault-style giant lever that operates a mechanism to engage the door seals. Or a fairly normal looking door with substandard performance and you put a few of 'em in series in an "airlock" sort of structure.
Doorjambs are a special challenge, since you want the jamb to be fairly low so you don't trip on it, and so you can roll equipment in and out. But that means all the sealing has to be part of the door, and you really want that vault-style compression mechanism. Or, you have a small high-jamb door that humans can step over, and a separate large equipment opening that you only open once in a while and just re-seal the perimeter with copper tape every time you close it.
Getting wires in and out is also a chore, though there are products for this, they're expensive. Without a proper passthrough, every wire that penetrates the envelope is basically an antenna that re-radiates signal on the other side. A single fire-alarm wire is enough to pass quite a lot of wifi signal! And good luck getting the fire AHJ to accept your notion of shoving a bunch of additional components on their wire to filter this RF stuff that they have no idea about. Got a safety cert for that?
AC likewise needs filtering, and the more power you're moving, the beefier the filters are. And the bigger they are in a physical-dimension sense, the harder it is to get them to perform at high (GHz-plus) frequencies, because the individual geometric features are resonant. You can't throw money at this problem, it's physics, so you just have to get clever. One approach is to put the breaker panel outside the enclosure and pass individual circuits through, since each individual one has lower current requirements and the filters are smaller.
Mind you, this is not a theoretical or absurdist "we want to block everything" exercise. Until the guy with the spectrum analyzer showed up, our quick test during assembly was just firing up a phone and seeing if we could link to the wifi AP on the other side of the enclosure wall, exactly the real-world scenario you propose. Every measure I've described here was necessary, and if we screwed up a single detail, we'd get wifi signal back.
It can be done. It's neither cheap nor easy. It helps if you have good test equipment and a lot of intuition for how RF behaves. And a massive, massive box of copper tape.
This an aspect of the increasing separation between “owner[ship]” and control that was predicted by James Burnham in his 1941 book The Managerial Revolution. (https://www.oxfordreference.com/view/10.1093/oi/authority.20...)
I feel like I’ve just accepted that we lost. I don’t see how we’ll ever get rid of mass surveillance. If we couldn’t get rid of the patriot act after Snowden’s whistleblowing, and we can get consumer privacy/protection now I’m not sure how we’ll ever get it. Maybe I’m just burnt out and cynical, but I don’t see why the government or corporations have any incentive to let people do anything without them knowing about it.
I have reasonable privacy online. None of my devices connect to google, except if I explicitly do so, which I do from time to time for a reason or other. None of my devices are even capable of connecting to Facebook owned domains, I have them DNS blocked.
JS off by default for all sites, no social media (except this site and one more that I run myself). I don't own a TV or any smart appliance whatsoever. No google assistant, siri, Alexa. None of that.
I buy things on Amazon. They know where I live and what I buy from them. They cannot show me ads because I let nobody show me ads.
All in all some people probably get some info about me, but the info collected on me is probably along the lines of what they were capable of getting in 2005. And good luck stitching it all together.
You can do a lot short of disconnecting. But you have to disconnect from the strip mall "internet". You can't wait on laws, because you're right, they have no incentive. You're the only one that does, so you have to do something about it, in your own life, for yourself.
When a lot of people say "it is just too hard" what they mean to say is "I want to use services that track me." For most people, only turning on JS for web apps and not using malware as a service is all it really takes.
Can't malicious application work around your DNS block? Is there a way to be 100% sure say all Facebook/Google is blocked? I am not a networking guy so maybe the future is to whitelist IPs/domains for each device and application.
Very quickly off the top of my head I could imagine:
- measuring out the appropriate volume for the teapot and delivering at the ideal brewing temperature for the particular tea
- knowing that one family member likes drinking water ice-cold while another prefers near-room-temperature
- recognising that it's about to pour onto hands and reducing the temperature to below scalding
- choosing from various sources (tap, filtered, recycled storm-water) depending on use-case, to optimise supply costs, filtration costs, environmental impact etc.
It's annoying (from a privacy and security perspective) how 'smart'-ening even the most innocuous things can be quite appealing to the end-user, and how nasty (again, from a privacy and security perspective) the implementations that make it to market end up being.
All of those tasks could be completed with robust and simple software, rather than ridiculous stacks of constantly updating libraries connected to the internet.
Don't forget "gadgets that are beholden to the manufacturer allowing them to keep functioning."
No iot involved, just a hand in front that is detected by a light sensor and 4 AA bateries.
Your smart faucet could do the same.
I agree. The benefits of "smart" devices could be a great boon to all of us, far beyond the examples you gave.
In fact, most IOT devices would be most welcome in my home if, and only if, they only pushed data to a local source directly under my control.
At that point, as long as I can manage the flow of data, I would be perfectly happy with a voice assistant (which connected to the Internet only to service specific requests and didn't pass my voice data to "the cloud") or an entertainment device which would connect to the service(s) of my choice and retrieve/stream voice/video/music when I direct it to do so.
Video door entry systems, room-by-room climate/lighting control, automated oven pre-heating and a raft of other conveniences would be fabulous too.
The problem isn't the technology, but rather how it's implemented. And these ideas aren't anything new either. Heinlein, The Jetsons and many, many other sources predicted home automation like this decades ago.
The real barriers to such locally managed systems is widespread adoption of protocols and software to locally manage this stuff.
But it's a chicken and egg problem. Why would home owners/builders/renovators include such systems when the IoT devices all phone home and specifically forbid reverse engineering and re-purposing?
And why would device manufacturers implement such open protocols and integration with locally controlled/managed systems when there's both no market for it, and they can increase revenues by selling all that tasty, tasty data?
There's a multi-billion dollar industry for locally controlled/managed home automation/IoT systems that could be created and cultivated, but the current barriers to entry are pretty high and (given that general-purpose computing devices for the home are on the decline) getting higher.
More's the pity.
- Mistakenly recognizing that it's about to pour onto hands and reducing the temperature
- Training users that it's safe to hold their hands under a faucet set to scalding, who subsequently burn their hands at someone else's place
- Training users that it's safe to hold their hands under a faucet set to scalding, then plain not working
I already have some rudimentary electronics skill, so it's mostly been metalworking, such as welding and machining. I have a 3D printer as well, I'm not against PLA for replacement plastic parts and not against using technology I have full control over as a means to my ends.
Don't need to buy a smart faucet if I can build my own very ordinary faucet.
Privacy is a side-effect of addressing other concerns. I use uBlock because the internet sucks without it. I delete old accounts to avoid data breeches. I avoid IoT because it's finicky and has a shorter lifespan. I compartmentalise my online life to keep the internet hate machine at bay. I don't use Facebook or LinkedIn because I don't enjoy it. I don't give my contact information because I don't want to be contacted.
All those privacy-oriented actions have other tangible benefits.
I see privacy like eating fruit and getting exercise. It's good hygiene, and even a half-assed effort is better than nothing.
Source: https://en.wikipedia.org/wiki/Patriot_Act#Reauthorizations
Connected remote controllable devices are useful, but letting them access the net is intolerable. X10 is a decent compromise as long as they promise not to internet their devices.
Disclaimer: Poster is not responsible for any damage caused by this approach. Do not violate FCC laws. Always use a Faraday cage for experimental transmissions.
You don't want to give it any longer than that though because it might cause physical evidence of damage (arcing on PCBs, ICs with holes in them, etc.).
Or so I've heard.
Many IoT-type radios use the same 2.4Ghz band as domestic microwaves and will be destroyed in a fraction of a second as the radio's AFE (analogue front end) is hit by many many orders of magnitude more energy than they're designed to tolerate.
I would wager that many devices that don't have other significant "energy-catching" features (antennas, radios, coils for NFC, etc.) would work just fine after experiencing a short stint in a microwave.
I did actually test this once on a very old 3G modem/radio interface and I found that every part of the interface seemed to function just fine, except that it could no longer receive any radio signals (and may have been unable to transmit but I couldn't confirm this with a spectrum analyser). The modem otherwise operated just fine and would talk to the SIM, operate all its LEDs and even seemed to pass an internal self-test feature that it had, with the obvious exception that it could no longer find any radio towers.
I obviously can't guarantee that your stuff won't blow up, so don't come blaming me when your device ends up being trashed by microwave energy going through unintended parts.
At this point I'm not aware of any TVs that come with cellular radios.
That said, I recently returned a soundbar after discovering it couldn't be used without app-based onboarding. Ridiculous.
I understand that those of us who would pay a premium for a dumb TV are a minority, but it seems like there’s enough of us to justify at least a few models.
The assumption that I would want to control everything from my phone is getting out of hand.
Here's a list of non-smart TVs on the local equivalent of Amazon[1], which includes this 75" 4K Samsung TV[2].
[1] https://www.emag.ro/televizoare/filter/tip-tv-f9181,non-smar...
[2] https://www.emag.ro/televizor-samsung-75-189-cm-75ru7099-4k-...
But you're right, apparently the TV makers aren't the issue, the stores are.
I love that "non-smart-tv" is a category on that Romanian site.
E.g. time to switch from one HDMI source to the next one. I have a projector where that clocks at an insane 15 seconds. Every single interruption of the HDMI signal results in at least 15 seconds of delay before the projector can show anything other than a black screen. It's just great when you are trying to setup the laptop for a presentation. The number of times I have cursed the entiry family of the designer of that projector's firmware, only to realize most commercial projectors are like that...
No antenna, no signal. Just make sure to get a TV that doesn't need online activation to function.
Step 1: Decide whether you actually need network connectivity for the main use case of whatever you are building. The answer will usually be no.
Step 2: If the answer to step 1 was YES, then now decide if you need Internet access for the device’s main use case. The answer will almost always be no.
Step 3: If the answer so both of the above was YES, finally decide if your device’s main use case requires you to build a server-side service, too.
If the answer is NO, plan on investing a lot in security professionals. If the answer is YES (I’d be shocked), plan on spending an enormous amount on security professionals. Then listen to those professionals.
Snark aside, there are lots of white papers on x509 provision of devices, including IoT devices. Go read them.
IoT security/privacy concerns in a nutshell.
https://blog.torproject.org/quick-simple-guide-tor-and-inter...
> Suppose I want to design an internet-of-things device with security in mind […]
Meanwhile, half of your answer is spent discussing whether or not OP might need internet access, when in reality this is already a given. In short, you are talking about general "smart devices" but smart devices ⊋ IoT devices.
Might be a case of the XY problem: https://en.wikipedia.org/wiki/XY_problem
OP's problem of "how do I secure my IoT device" might go away entirely if they don't actually need it to be an IoT device.
But without some knowledge of what they're building and why, all we can do is guess.
I foresee origami practice with aluminium foil over known radio-containing parts in my future.
Regarding Smart TV's, is there an OpenWRT-like project for TVs yet?
As I understand it. Nobody bothers since there are som many different versions of TVs ( even same model different region) that hacking and developing new os would just never catch up. So instead people opt out of smart TVs and instead buy "digital signage screens" basically a dum TV....
Install the firmware with specs as per TV panel. Swap out the control board altogether.
However, because you would have no data to sell, you would have to actually sell the item itself for a price that would be sufficient to make a profit.
Does anyone doubt that an unethical company could make well over $5/year from storing and analysing every second of audio spoken within earshot of its devices?
Lane keep: because you're on your phone
Collision detection/auto brake: because you're on your phone
High Beam Assist: because you're on your phone
Adaptive Cruise Control: so you can be on your phone
Auto Headlights: because turning on your headlights means putting down your phone/coffee cup.
I've only ever been annoyed by Collision Detection/High Beam Assist/Lane Keep/Auto headlights. They do the wrong thing at the wrong time.
The whole concept of owning a car outright and being allowed and able to maintain it if you choose has already been effectively ended by complexity, lack of shop manuals and legal 'don't remove the plastic engine cover or you'll void the warranty' restrictions was the first wave.
Now we have cars transmitting gigabytes of information back and forth to the vendor and who knows who.
IMO we should have consumer legislation protecting owners from this razor and blades model the vendors are trying to force through: you buy the car then we sell you endless 'features' and upgrades which will ultimately render the car inoperable unless you.
Cars used to be cheap and repairable. Now they are expensive, have a short lifecycle and are becoming infested with endless digital do dads $£ upgrade.
I disagree with the author because I think there are legitimate reasons for devices to connect to networks. For example Tesla cars can get better over time. TVs can do updates to fix bugs and new functionality.
But the down side is the opportunity of spying and monitoring the user.
For example what happens if your Washing Machine stopped working because a bad manufacturer decided to stop it over cellular?
Beating the bush around one's pet peeves or detailing what one deems an optimal policy is fine, except it often waters down that basic message for 99% readers, when only 1% will be somehow informed/enlightened.
There was a case in the news a few years or so ago where someone did that with the SIM that was in their electric meter, which was used once a day by the electric company to send a report of cumulative usage.
What they did not count on was that many cellular carriers have data plans specifically for devices like that electric meter that just need to send a very small amount of data every hour or day or so.
Those plans have a very small monthly fixed fee, a small data allowance, and astronomical overage fees. Typical data allowances for the cheapest plans are maybe 50 KB, which is plenty for sending "<meter-id> <timestamp> <cumulative kWh> <checksum>" once a day.
The person used the SIM for their calls and data, including downloading a bunch of movies. They ended up incurring around $150k overage charges.
The electric company cared very much, and the person ended up with a short jail sentence.
Depends if it's using a physical SIM card or an eSIM module like Apple does that's entirely provisioned from the device's userland.
> and use it to get free calls all over the world, including long-distance?
Depends on the plan assigned by the vendor / phone company. If it's a data only plan, then no.
> Would that be illegal? Would anyone care? :)
Depends on which jurisdiction you are and what you're doing. A couple of dollars worth of charges on a company that sells millions of "smart" devices? Probably won't even get flagged. A couple hundred dollars for calling a phone sex line or someone on a satellite phone however? That will cause someone to have a look.
There's an old (10 years?) case about smart traffic lights that included SIM cards for connectivity - which were taken out of the traffic lights and abused. See https://www.bbc.com/news/world-africa-12135841
Another case (for which I don't have a link) some years ago was that ornithologists were doing tracking of large migratory birds using ankle bracelets w. embedded radios - until people "underneath" the birds found out that they contain a SIM provisioned for international roaming, which resulted in birds with such bracelets being shot down to obtain these SIMs as they could be used to extract and sell a few thousand dollars in comms services to locals.
Once the kids get older, all that goes out the window. They'll become the security hole. Haha.
A router that analyses the packets and allow- or blocklists packets based on features would be useful.
Flashed TP-Link routers, hostapd on a NUC-style PC, Mikrotik, even Ubiquiti Unifi are what I would consider way beyond general consumer level. The userfriendliness just isn't there.