“The door refused to open. It said, “Five cents, please.”
He searched his pockets. No more coins; nothing. “I’ll pay you tomorrow,” he told the door. Again he tried the knob. Again it remained locked tight. “What I pay you,” he informed it, “is in the nature of a gratuity; I don’t have to pay you.”
“I think otherwise,” the door said. “Look in the purchase contract you signed when you bought this conapt.”
In his desk drawer he found the contract; since signing it he had found it necessary to refer to the document many times. Sure enough; payment to his door for opening and shutting constituted a mandatory fee. Not a tip.
“You discover I’m right,” the door said. It sounded smug.
From the drawer beside the sink Joe Chip got a stainless steel knife; with it he began systematically to unscrew the bolt assembly of his apt’s money-gulping door.
“I’ll sue you,” the door said as the first screw fell out.
Joe Chip said, “I’ve never been sued by a door. But I guess I can live through it.”
Harry Shearer's (of The Simpsons fame) LeShow [0] has a recurring segment titled "Smart World" where he catalogs recent news coverage pertaining to IoT/"Smart" device fails, it's amusing.
> Tech Enthusiasts: Everything in my house is wired to the Internet of Things! I control it all from my smartphone! My smart-house is bluetooth enabled and I can give it voice commands via alexa! I love the future!
> Programmers / Engineers: The most recent piece of technology I own is a printer from 2004 and I keep a loaded gun ready to shoot it if it ever makes an unexpected noise.
Thanks! I hadn't seen that, but yes, the magic phrasing is clearly derivative. I wonder why they bothered given that it was pretty clear what the metaphor was about.
I love his observations about the nature of magic and power. Terry is top of my list of people I wish I'd met while they were alive, he had a tremendously humane view of people and what makes for a good world.
For what it's worth I have a Roomba and you can still press the button on the vacuum itself and it works just fine. It's just the (back-end of) the app that's down, since they host it on AWS. Doesn't seem too unreasonable.
It's ridiculous. I don't understand why so many of these things need to connect to cloud services just to work.
I mean, surely computing power is cheap enough these days that you can build all the intelligence you need into the device for day to day operation and simply use internet connectivity to get updates, such as security patches, and the like?
Maybe the compute side of things could be a pluggable module that could be upgraded independent of the rest of the device if you ever do need more processing power. This would work particularly well for larger devices such as robot vacuum cleaners.
I find it indefensible that some data centre going down in Australia (or wherever) could render something as basic as a doorbell, or even a vacuum cleaner, inoperative.
>I mean, surely computing power is cheap enough these days that you can build all the intelligence you need into the device for day to day operation and simply use internet connectivity to get updates, such as security patches, and the like?
I work on embedded systems that can not fail when the internet connection fails, and it's bonkers just how much you can accomplish with a <$5 microcontroller or even a <$10 chip running linux. But the tide of the industry is pulling toward remotely hosting as much as possible.
Even if it doesn't strictly need networking, the desire for a companion app, firmware upgrade workflow and/or business desire for surveillance often justify it. Once you have network connectivity for any reason, well, then you're not not going to be phoning home to a surveillance system running in AWS anyway. If you have a companion app, making sure you can get a connection between the phone and device is going to have a fall-back path that runs through your servers because even self-styled techy people don't always understand networking.
Now that your networked device is already sending all the data home anyway, why not do more work serverside? Add in the prevalence of developers that can't be arsed to learn to develop for a microcontroller. Or those that can only get by in the very-handheld embedded environments. Now your design is starting to take shape, and you need to do some more number crunching. Do you invest in learning to leverage esoteric features of your microcontroller? Add a DSP to the BOM? Or just say it's easier to do on the serverside with comparatively infinite processing resources and the ability to rush development because features and bugfixes only require updating servers rather than pushing firmware to the fleet?
My theory is that consumers already have wifi networks and know how to enter a wifi password (or push a wps button). They don’t need to buy some new base station and then worry about if it speaks the right protocol as all their other devices. Historically there wasn’t a ubiquitous, push button, cross-vendor wireless protocol that fit the bill. I’m really hoping that everything actually moves to local comms over zigbee.
I've learned my lesson and no longer buy WiFi connected "smart" devices, unless I can flash them with custom firmwares.
ZigBee isn't without its flaws too but at least I can guarantee that the device doesn't require someone's online service to function. I just wish more devices offered it.
Pretty much anything with WiFi in it requires installing some dodgy looking app that requires you sign in with yet another account, and will probably stop working in a few years or when you eventually need to reconfigure your home's WiFi network.
My old dumb tv is failing. After looking online, it seems the only dumb TVs left are in the US (Spectre and very little else) and not available in Germany. At least the spying is removed by not connecting it to the internet, though I still get the great full OS experience :/
At least the spying is removed by not connecting it to the internet, though I still get the great full OS experience :/
I have a smart TV not connected to the internet, just a HDMI connection to another box. Then I learned that IP-over-HDMI is a thing!
In the very near future this will be even more futile - the TV will have it’s own built-in cellular connection just for telemetry. Anyone who objects will be shouted down as a 5G conspiracy nut, but this is exactly what 5G is designed for.
Anecdotally, the first Sony smart tv my grandma got (after her old tube TV died) managed to receive a firmware upgrade over the antenna cable (don't remember if it was cable TV or satellite). She certainly didn't have any internet in her home.
Of course that led to a fun episode of "My TV is broken" (actually just showing a prompt to start the firmware upgrade), in general she got very confused by the smart TVs UI often...
We bought a larger monitor for the computer and have it placed where we can watch movies from the sofa. It's not as nice as a real dedicated TV, but saves space and avoids the headache of modern TVs.
depending on your budget, you could look for "digital information displays" which are displays that are meant for advertising in shops/malls/storefronts, and most of them are literally 'giant monitors" with no smart features. Expect it to be pricier though.
Don't know about you, but when my 2017 Sony (not so smart, but starts fast and doesn't call home) will die, I'm not going to buy a regular TV set anymore but just a monitor, like I used to with my old PS3 and PlayTV. What's left of free-to-air TV (state broadcasters on DVB, if the signal is even strong enough) frankly feels like religious programs played in prisons. It's caustic and depressive social-pedagogical palliative crap, during coronavirus more than ever. Someone from Hollywood should come over and teach about the concept called "entertainment".
Nextdns is the best solution I found to this problem (they have a smart TVs blocklist) along with a firewall rule that blocks all port 53 traffic except to approved dns.
Shame this stops working with the next firmware update that switches your TV to DNS-over-HTTPS. Because then, just like google ads switching to the main google.com domain, you will have to choose to block either every service from your tv manufacturer (including firmware security updates) or allow it all through.
My next TV is going to be a dumb one with a bespoke HTPC...
to be fair, I'm actually too stupid to vacuum a carpet vs floor without it auto adjusting fo me, so just to play devil's advocate, my vacuum has worked wonders on sucking up the ight amount of floor baddies. I can also forgive the glass stuck on the floor on thanksgiving night because of all the help it's been 364 days of the year.
edit: oh yeah my letterr "R" key is broken on my 2015 macbook po fom all my refreshing, so sorry fo the typos
> I've learned my lesson and no longer buy WiFi connected "smart" devices
Not to gloat or enjoy too much schadenfreude, but many of us learnt the lesson in advance because we saw it coming. It may just be because we are old and cynical though.
HomeKit compatibility is also a reasonable proxy; that's because HomeKit requires that a device operate without a manufacturer's service, and expose local APIs that your iDev can use.
The fact that so many manufacturers are happy to certify with Alexa and Google, but not HomeKit, tells me a bit about their business models.
Yes it‘s 100% local, but i cant remember whether its using bluetooth or wifi. And if you want to control it from remote a homekit controller will male a secure connection from your apple device to your home network by proxying encrypted data through apple servers.
From a Reddit comment: "The home assistant tp-link integration already makes use of the local api connecting to the devices. I'd definitely recommend using a reserved dhcp address and specifying the devices in your ha config (not using auto discovery)."
They do if you want to use most of the functionality (e.g routines, scenes, connecting to a smart assistant, etc.) Using them without an account basically just lets you use your phone to turn them on and off.
Fortunately this is also true with many devices, but knowing which ones allow this before buying is the problem. Most shops won't advertise the chipset. This is also beyond the abilities of 99% of the population.
Yes (the Tradfri line). I just bought some to set up with home assistant and they all have the ZigBee logo.
I was a little surprised, but pleased, to see that Lidl are launching a smart home line this week, which is all ZigBee too - that is definitely not niche either. They even mention the interoperability and privacy as selling points!
It’s a home automation protocol, like X10. You buy devices that can talk ZigBee to a hub device (probably can use a computer on your local network for this). From the hub you can do all the usual automation tasks.
To extend that slightly: it's a low-throughput, low-power mesh network that doesn't need the smarts or power draw of Wifi. Because it needs an in-home hub anyway, it's unlikely to die without Internet access, and devices on the standard have at least some level of basic inter-operability.
So if I wanted to say put a bunch of sensors in my house to monitor temperature over time without necessarily tying that info to my thermostat or anything, is this ZigBee thing a promising thing to look into?
Yes, and there is even a great FLOSS hub software Home Assistant [0] (though to be fair, it requires some tinkering compared to the proprietary systems) and if you run it on a PI or small server, you might as well get a conbee II which is a USB stick that adds Zigbee support directly :)
The bigger hassle seems to be different zigbee hubs that have varying levels of compatibility, so while you can pair the devices you can't get the data you want.
I've just been down the research rabbit hole, and about to try using zigbee2mqtt to bypass this nonsense (this also then means the only internet connectivity will be through a device you control)
Also systems like Philips Hue use ZigBee protocol. We have Hue lights at home and I'm planning to plug a ConBee II to my router for the Home Assistant to see how well I can run the lights with another hub.
Philips Hue hub talks to diagnostics.meethue.com all the time, and it gets kind of annoying already in the pihole stats.
there's also a difference between IoT and "look ma I wired an ESP32 to a DHT22 and sent data to timstream on my NAS". The latter is entirely you. WiFI isn't the issue, it's running cloud based hardware that can't work without someone's servers.
Bought a WiFi light bulb for fun years back, it died and they replaced it then it recent died again and I just went back to a standard bulb. Even when it worked it would periodically lose connection the the WiFi and need a reset, assuming the app worked correctly and paired. Not worth the hassle.
Some of the colour changing WiFi bulbs seem pretty neat though. I used to have a Philips Hue (standalone, two of them) and mucking about with those remotes wasn't always that great either.
We are talking about a full computer that is small and cheap enough to control a single light bulb. We got here through unbelievable technological development.
Why would you want it to stop now, when the tech is both amazing _and_ rough around the edges? Better hope development continues and the end-user experience stabilizes to something much better.
I share your enthusiasm for how amazing our technology has gotten, but I'm also experienced enough to know that the technology will always be "rough around the edges" as you put it, because maximising profit doesn't require stable end-user experiences that are as good as they can be - they'll constantly be just barely good enough to make it into the shops, and then let the marketing take care of getting enough wallets to open up to make a profit on them.
Capitalism continuously fails to deliver on its promise of "may the best win, and may all be incentivised to be the best" because companies have found several loopholes in its assumptions - constantly introducing flashy new features so the basic faults don't get the airtime they should, advertising with the best psychological manipulation money can buy, and flooding the market with enough barely distinguishable models
that simple decision fatigue gets enough people giving up on research and buying the crappy models unwittingly.
I have 10 LIFX bulbs. Hardware wise they're really great. I don't mind having to connect them to WiFi. However, the app to control them is, bar none, the biggest piece of shit app I've ever used. I can barely control my lights.
The scheduling is pretty handy though. I have half a dozen outside and they turn on / off with dusk / dawn. In addition to that I have them set to go to bright white and 30-50% at night (just enough to keep security cameras in color mode).
I’ve had an unacceptably high AFR with all LIFX bulbs, although I’ll concede their support is fantastic at shipping replacements to me. We’ve had about 20-22% failure rate of bulbs within 2 years of ownership, seen across ~60 of them. In addition to that about another 20% have needed factory resetting once or twice in the same period, because they stopped interacting with WiFi and the first-party application in ways that just turning off and on again at the light switch didn’t fix.
In contrast to this we’ve had exactly one Philips Hue fail on us. Also cheerfully replaced for us. None of them have needed resetting and we have about the same # of light fixtures deployed. Probably the most annoying problem is Hue Play’s are prone to coil whine at 100% brightness on on certain scene colors, particularly those involving blue-white (6000K). Resolvable by going < 5000K or dropping to 90% brightness, I haven’t attempted to get warranty replacement for this issue but about 16/22 of my Hue Play’s are affected by the issue.
Integrated with Google Home it’s also common that some % of LIFX bulbs won’t respond to a “turn off all the lights” command, whereas Philips is very good at on/off even when the command involves large numbers of bulbs...
All the lights in my house are the philips hue that I bought back in 2013. So far I have replaced 2 only. The last one was coming back on by itself... wierd to replace a lightbulb because it doesn’t want to be turn off...
I don’t know if I will ever understand how someone can write out a comment clearly indicating a massive pain in the ass (20 lightbulbs failing and requiring calls to tech support, RMA, etc), and not be using it to talk about how this whole industry is a nightmare.
If you buy a car that keeps on breaking down, will you talk about how the auto industry is a nightmare or will you just pick a different manufacturer next time?
I am clearly bringing bias to this conversation, but I have dabbled with some smart items (hub, locks), and have a friend who is into them. It’s non stop fiddling and fixing from all my actual experiences.
You also have to remember the negative externalities of extra e-trash affect everyone, so to read about yet another iot failure, it’s easy for me to continue to trash the whole industry.
I’ll concede the point on a technicality, but will still stand by my motivation behind that comment - increasing the maintenance requirements for normal life sounds like a terrible way to spend money, environmental impact, and time...
"It works well" doesn't sound like "a technicality", but okay.
Humans remember negative experiences much stronger than positive ones, and "it just works" isn't really something you tend to notice in the first place. There's a lot of bias here.
Looks like zigbee is moving closer to critical mass. Echo’s now ship with zigbee radios, and the zigbee alliance has been a lot more active. Hoping in the near future everything just works on zigbee
Sonos has been a great example of this for me. For those that don't know Sonos products are WiFi speakers, sound bars and subwoofers. I've had them for around 4 years and whilst they're controlled through the Sonos app (and third parties) if our internet crashes there is no hiccup whatsoever.
I was a bit upset when I discovered that I needed to connect my new Bose speaker to internet just to configure it, but at least it is usable without internet, so I blacklisted it in my router.
Spotify sends home what I am listening to, why would I care if my speakers do as well. I can also control Sonos through the Sonos app or directly through Spotify.
Also these speakers have only improved over the years through software improvements which would have only been made possible having a connection to the internet.
A speaker's output transfer function is determined by the drivers, the chassis, bass reflex or not, frequency splitter, etc. Software has nothing to do with it. And even if you're only listening to Spotify or using fully-integrated equipment from Sonos or others, you will care if your speaker decides to play ads, in addition to Spotify ads.
Yep. As long as your LAN is functional and the hub (which uses the ZigBee protocol often referenced in the comments here) is appropriately routeable, internet access is entirely optional - unless of course you want to control the lights from outside of your home.
Then what prevents the app from being the path, in either/both directions? Reporting on the activities of the devices it can see, as well as sending firmware updates to them?
Your printer gets firmware updates from HP, whenever HP wants, right past all your firewalls and vlans and non-routable subnets, in the form of print jobs from your pc.
If you want to control your lights from outside your home you can either setup a WireGuard LAN or just enable the integration with apple home if you are already in the ecosystem.
Apple has less incentives to spy on you than whoever is running philips hue.
A friend of mine who lives in New York actually has an apartment where all building an unit access uses an RFID enabled lock system with an app on the phone. It’s very easy for the owner to lock someone out and the system went out twice and they couldn’t get in. She doesn’t even get a physical key for a back up.
Read the lease carefully. Two years ago some tenants in NYC won a lawsuit against their landlord for not giving them keys.[1]
Anyway, I assume Amazon is already considering what GP proposed. They have a program called Alexa for Residential[2] marketed to property managers who want to put IoT things in apartments.
Sinister? I’m not sure I understand. The doorbell doesn’t stop one from entering their home when Amazon servers are down. It also doesn’t stop people from knocking on the door.
You're missing the Forrest for the tree. What about smart door locks, smart boilers, smart heaters, smart AC thermostats. smart bathrooms, smart fire alarms..?
This reminds me of the other day when my power went out in the middle of washing laundry. I was going to take out the wash to put it on the line to dry but the lid was locked and because it is a solenoid switch, I wasn't able to open the washer (it requires power to open). This kind of device isn't even "smart" in the sense of internet connectivity, but it did really surprise me that it was not configured to release if not powered.
Earlier this year, after a few months of lockdown, I discovered that even car batteries that were in reasonable condition can go completely flat without much warning if the vehicle is only being driven rarely over an extended period. I was aware of the possibility of the small power draw draining the battery over time, but hadn’t expected such a strong effect so soon given I was still driving the car a few miles a couple of times a week.
That was annoying enough, but then it turned out that the alarm system in my car has a backup battery to continue powering the siren in the event that the main battery is disconnected, presumably someone’s idea to stop a thief from silencing the alarm quickly. That backup battery can run the siren for several hours, obviously without reference to where the vehicle is or what time of day or night it is.
Moreover, if you’re unlucky, when you try to unlock the car the failing main battery can have just enough juice in it to trigger whatever condition makes the alarm system think the main battery is being disconnected but without actually disarming the alarm system, thus setting off the siren.
Which you can’t stop without power from the main battery.
It's usually a voltage drop detection that sets of the alarm. If your battery is going flat, or is tired, it can drop below the voltage threshold set in the alarm it off. You may still be able to start the car.
You may have a bad alternator though, if it's not the battery. It really depends on the car configuration, but it is surprising that even that short trip doesn't recoup the cost of starting.
You can get alternators and batteries that are better at your use case too, work site utes and shuttles often get used like you describe.
Yes, the voltage drop was the likely trigger for actually setting the alarm off. When I had the battery replaced, they did a quick test on the old one, and it was performing very far below normal.
With the wisdom of hindsight, I should probably have left a trickle charger on the old battery now and then while it wasn’t getting much long distance driving to keep it charged up. I was just a bit surprised that there was so little warning given the dire state it had apparently reached before the alarm incident.
But of course, the real problem here isn’t that a battery with unusually low usage drained faster than I incorrectly anticipated, it’s that my own car had an alarm going off that I had no reliable way to stop even standing next to it with the key in my hand. Clearly this should never be possible. In fact, I read that a lot of vehicle security systems do have some sort of emergency override for this sort of situation, but I found nothing relevant in the documentation that came with the car, and unfortunately the various suggestions I found online for similar models didn’t work on mine.
My opinion and experience with car alarms is that they go off so frequently in error, that no one pays any attention to them. I have heard countless alarms go off, and not once seen a car broken into. I think it's better just to have a comprehensive immobiliser.
I have also had my car stolen. They had the alarm disabled in seconds, and then had all the time in the world to get it running. The only reason I still have that car is because they stalled it in the middle of the street, and the immobiliser kicked back on, so they fled.
That car is now mostly a track car, so it's actually got a pin-locked ECU that I pin in before driving it. No amount of hotwiring would get the ECU to work without that pin. Even still, it would just need to be dragged onto a trailer and I'd never see it again.
Btw, aside from the old battery, you might have something in the car that drains it excessively. IIRC I've had the audio system plugged in a wrong way and it put a small but considerable constant discharge on the battery, even when the car was off. Alarms themselves can do this too, I think.
Yes. I spent a month in lockdown not able to access the things I'd left in my car's boot (trunk) which can only be opened when the battery-powered central locking is available.
Anecdotally, once some modern car batteries die in this way, they fail to hold a charge at all, so classic jump-starting doesn't work. So I only replaced the battery when lockdown was over, to avoid the same problem happening to a new battery.
You say "lid", but is it a side-load washer? Those have a valid reason to not open if the power is cut --- because they'll empty their contents all over the floor.
I have an old top-loader, it doesn't care at all (or know) whether the lid is closed.
Meanwhile my Japanese side-loader sits at a slight angle and never fills up enough to go above the door line, so you're allowed to interrupt the cycle at any point and the door will unlock (after the drum has stopped spinning). When I first saw it I realized how poorly designed all the side loaders I saw while growing up in Europe were in comparison (nevermind not being able to throw something you forgot in after starting, also the risk of a problem leading to water not draining properly, and a flooded room when you go take the laundry out... been there).
I do wonder what it does if I pull the plug with the door locked though. I should try it.
> also the risk of a problem leading to water not draining properly, and a flooded room when you go take the laundry out...
My condo has two front-load washers, and I once saw someone start one up, the door lock, and the seal break - so it was spraying water everywhere. These don't have any way to interrupt them, so we got a maintenance guy to unplug it - but the door still didn't unlock, even after plugging it back in. The maintenance guy ended up jerking it open hard (might have broken the lock, don't remember), so the other resident could get their clothes.
I only ever use our top-load washers, which don't have any lock on them.
Yeah that was my thinking too, but it should still have a manual override of some kind, if for no other reason than the possibility of a child getting stuck inside.
It is a top loader. I don't understand why it has a lock. It has been very annoying actually, because I am used to letting the washer fill up while loading it.
It's a safety mechanism mainly intended for children. I think it's been around for a while. The washer Mom bought 10 or 15 years ago had it. I think there's an unlock button on hers, but I can't remember. I do know it's incredibly annoying.
Seems easier and cheaper just not to put your arm in there while it's spinning. Toploaders have existed for actual decades without solenoids locking them closed, is this actually a problem or just some manufacturer putting in a solenoid because their side loaders have one?
When she was five years old (late 80s/early 90s timeframe), my sister propped the lid up on the washing machine, and jammed a toothbrush into the safety switch so it would run thinking the lid was down. She thought it looked really cool and pretty. At one point she decided the vibrating was cool too. She got up and danced on the top of the washer. Being that soap gets slipper when wet, and kids don't always have the grace of a gymnast, she slipped and landed feet first in the washing machine. It's not every day the emergency room get to body cast a five year old kid with multiple femur fractures.
The grotesqueness aside, I do loath the pendulum swing we have toward safety these days. It's so bad now that it makes things worse in a lot of cases. I have to dick with my phone for navigation (which requires a lot more looking away and danger) because my car won't let press a couple of buttons while it is moving in the name of "safety."
You already live in this world, though. Or at least something close to it.
It's actually quite astonishing how, despite the growing chemical and mechanical complexity of household items, most household things are actually quite safe to have around children. It would be quite easy for these things to be actively dangerous, but fortunately a combination of thoughtful design, fear of lawsuits and regulation works to make them safe.
(Also if we're talking in context of white goods, you can't ignore children in the design process, as a good chunk of the target market has them, and the play around where such appliances are deployed.)
Example of thoughtful design: prescription drug containers that can't be easily opened by children, but can be opened readily by adults.
Example of mindless nanny-statism: washing machines that you are simply not permitted to open at certain times, period -- not even in case of power outages, component failures, or plumbing emergencies.
I hope you can see the difference. Don't defend shitty engineering. It makes the whole world a worse place to live.
Entirely my take, minus the irony. I hugely dislike driving with an armed boxing glove in front of me. A boxing glove originally put there for the benefit of people not inclined to wear a seatbelt. The vast majority of traffic accidents I see (and I drive 8 - 10 hours every day) are entirely self-inflicted. Too much speed, too little distance, no focus on the road ahead and the rear view mirror.
If there isn't a meteor on the hood, but instead the driver ran into something, then the number of times your assumption is wrong will be sufficiently small that the parents observation is valid.
Certainly it will be a far more useful way to guide policy than "You don't know for a fact that all those accidents weren't caused by external events outside the drivers control." however technically true.
There aren't evil driving-fairies nudging people's elbows, and we have manufacturing and inspection regulations to keep the percentage of mechanically unfit vehicles low. That just leaves the drivers themselves holding the primary responsibility in the majority of cases.
These are all things you already knew, which makes the question disingenuous and of questionable ultimate purpose.
I was in a major wreck last winter. Hit a tree head-on due to black ice on the road. I was not braking or turning at the time of the slide. My car had snow tires and was in 4-wheel drive.
How exactly is that my fault?
You make no men to on of environmental factors causing wrecks.
You also make no mention of multi-car wrecks.
Clearly you live in some part of the world with perfect weather?
Attempting to refute a point that no one else made is definitely your own fault.
The original commenter can be reasonably assumed to know which accidents they witnessed could have been influenced by things like weather, and thus are not part of his assertion.
Likewise all the follow-on vehicles in multi-vehicle events are irrelevant. No mention was made because they obviously have no bearing on the observation and point being asserted. Meaning, sure, in addition to a large number of a certain class of accidents he's talking about, there are also a other classes of accidents he's not talking about, so what?
Clearly you live in some part of the world with imperfect schools? Oh was that rude and childish?
Every top loader I've seen solves this with an interlock: opening the lid stops the motor and brakes the drum. Closing the lid resumes from where it stopped.
Yeah, I do the same too. Like if I find a loose sock I can throw it in during the first cycles. I don't think top loading washers make enough of a splash to even warrant a lock, I may be wrong though.
My car [Ford Falcon BA] battery completely died the other day (was reading 8 Volts) and I couldn't even unlock my car to pop the bonnet, since the locks are solenoid actuated. It also only has one external lock, on the driver's door.
I ended up having to wedge the door and open it from the inside with a coat hanger.
Apparently the only other way to open the car when the battery has completely failed (as opposed to having charge, but not enough to turn the starter) is to connect jumper cables to the positive lead on the starter motor and the chassis. Of course my jack was in the boot, which I couldn't open since the battery was dead!
I've now attached a wire to the bonnet latch, so if this happens again I can just pull the wire to open it.
Most cars seem to have a subtle trick to open doors when the battery is dead. Mine (an E90 3-series) has a physical key, but you have to know to pull the lock on the driver's door then turn the key to open it. Turning the key does nothing. Turning the key then pulling the lock does nothing. I found this information buried in the comments of a YouTube video that had elaborate instructions to get underneath the car, and look for some wires in the engine bay to pull so as to pop the bonnet.
Yeah, my car does actually have a power-fail-safe lock where you can pull off a cover and get a real keyhole. For some bizarre reason, it's on the passenger side.
I think most washing machines need power to release the lock. Water damage would be very common otherwise. Without power the machine cannot pump out the remaining water. At least that makes sense for front-loaders. Why this is also commonly implemented for top-loaders I cannot say.
There is probably a mechanical switch somewhere but it might not be documented.
All front loading wash machines I have owned had a small hatch at lower part of the front, hiding a sieve. That can be used to release the water and the emergency door release is usually in there, too.
I was working on a "solution" to automate stuff without relying on big players like Amazon or Google. A more DIY approach. I end up creating two projects (that really needs refactoring):
and I even record some videos of me using this two projects to control a raspberry car with voice controls. None of my friends were interested in it. They all got Alexa or some other assistant. So, eventually I end up putting my projects in the drawer.
But hey, who knows now they would appreciate a custom made solution that doesn't steal your data or infringe your privacy!
I started doing my home automation 15 years ago. Without much care, because surely, anything I'm able to do will be on mass market in a few years. Long story short even ignoring relying on 3rd party servers and privacy concerns there are still no comparable commercial solutions today I'm aware of. Latency, reliability, being able to create actually "smart" home instead of having 20 gimmicks.
This is not meant as a brag. I think the market is just too small and to truly have anything remotely "intelligent" you basically need to be a coder because there is no better way to tell your system how to behave. Even then it takes few years to iron out special conditions and however small ongoing maintenance.
On the bright side, micro-controllers, sensors and single board computers market blossomed so nicely that it's basically lego blocks if you have basic electronics understanding.
Coders can dive in easily, but I should warn you, an approach of "I like tinkering" will make you much happier than "I want to get this thing done, it will save me time".
> from hackaday gave me all sorts of sh*t because I used GET to activate and deactivate a PIN.
The biggest issue with this is that some common chat programs and text message providers will request URLs for virus checking or getting image previews.
Using a GET isn't a bad idea but if a random service using your GET url can cause trouble, it is a bad idea. You'd be surprised how often you need to move a URL from a device to another and end up emailing it / texing it / slacking it to myself :D
Home Assistant is such a solution. It's quite decently polished and supports a ton of devices. Where devices have local APIs, it uses those, so it doesn't rely on vendor cloud services unless there is no choice.
I use it with TP-Link Kasa smart plugs and Broadlink IR blasters, both of which have usable locale APIs, on an isolated WiFi network with no Internet access (it's harder to turn off the cloud stuff for these kinds of things, but as long as I can block it myself I don't care if they try and fail to connect to the cloud).
This was a big driver for me moving from Windows to Linux last month. Some stupid automatic update and the next thing you know my computer's hard disk gets corrupted, resulting in a disk cleanup that deleted all of my important stuff. Thankfully I had everything stored separately (cloud+hardware). Not to mention the new Windows update slowed down my other PC severely.
This is the big problem with most (all?) commercial "Internet of Things" devices. It's a real mess, and that is because this idea is unworkable from the start. What we should have instead are many, isolated "Intranets of Things".
Luckily, not everything uses WiFi. There are tons of Zigbee and Z-Wave devices which physically don't have a WiFi chip.
Of course, you should build your own bridge to connect them to (rather than using the bridge of a big tech company).
For Zigbee you can use the Conbee [0] and for Z-Wave you can use the Z-Stick [1] to build your own bridge.
Furthermore, some of the cheaper WiFi devices that are based on the ESP8266 chip can be flashed with a custom firmware called Tasmota [2] to make them cloud-free.
A ZigBee or Z-Wave stick + Home Assistant (or openHAB, Domoticz, others) gets you more reliable operation and no notifying $MEGACORP every time you flick a light switch (or something equally dumb).
Devices based on the popular ESP8266 can also frequently be freed from their cloud overlords with community-made firmware [0], [1] and used as proper Intra-o-T devices.
I decided a while back that I would have no part in this Inter-o-T madness and I think it was a good choice.
Trading cloud provided services with your own home grown systems just trades unreliability in one way with its own sets of maintenance and reliability problems, plus a lot of complexity that is out of reach for most people.
Smart home technology is a mess and someone that cares enough needs to design an open, reliable, future secure, and user friendly system that can work wired and wireless.
Do note that you can unintentionally expose data to the BigCo too -- Home-Assistant and I'm sure the other platforms can export all local devices to Google/Amazon to allow control via Assistant/Echo and will send entity state updates to facilitate that.
Everyone has their own threat model, but I'm more comfortable with Google knowing when my light switch is on than I am with their outage or service deprecation preventing me from turning it off.
For the technically-literate home automator, I think the sweet spot is something like Hubitat. It's not home-brew, has decent commercial support, integrated Z-Wave and ZigBee etc. and is designed to be completely self-contained on your LAN.
I dumped my Wink Hub for one after they did their bait-and-switch for subscription fees. Never been happier with home automation.
> What we should have instead are many, isolated "Intranets of Things".
Nah, Internet is great for updates.
The problem, while it doesn't fit a description as slogan-friendly as yours, is the reliance on centralized servers as essential for routine operations rather than just for optional (though potentially important) updates, and this isn’t an engineering problem but a deliberate business decision to enable X-as-a-Service business models (whether paid for via recurring subscription, ongoing user data, or both.)
Yes, I'm sure they would sell tens of vacuums that way. And then you would complain when you got hacked because you opened up your vacuum to the outside world.
Ah, hell, why not. You see, he made it his business to point out that the phone-vacuum communication has no business going through Amazon servers, there's simply no need.
You’re talking about an average consumer that can’t remember their WiFi password or simply pair devices with it. Of course the future is in devices shipping their own internet, and everyone who works in that space would tell you - whether by their professional experience or through user surveys - that WiFi / local internet is the biggest obstacle.
That is not universally true. There are also plans to use completely different radio stacks and offer localized control.
The amount of people that do not want any cloud controlled devices is quite significant.
It is also a huge waste of energy to have my device contact Amazon to set a value. Voice interfaces can be nice, but only very few customers seem to be interested. It also adds costs for customer support and you need maintenance on apps and services.
I work in that space, although it is true that IoT is not central to the products, so I don't know about the latest trends. Luckily it is just a feature we like to show of on expositions.
For me a line is crossed if people add security cameras directly connected to vendor networks. If that is allowed I should also be able to install my laser with 100w optical output power and face detection. I have a right to security after all.
If companies only listened to the CTO of the infra supplier they use: 'Everything breaks all the time, design your infra with that in mind'
AWS has multiple regions, and within a region multiple availability zones.
If you are as big as roomba, you should simply have your api's running in more than one region. Then, if us-east-1 goes down (which happens from time to time) your users will not notice (ok, the geeks will because latencies will change)
So yeah, while the underlying problem was an outage of a region of AWS, it's not AWS that failed in this case. People should simply design their applications like AWS is telling them.
Obviously @internetofshit is on it... Highly recommended account for those of us who struggle to see the flip side of automation https://twitter.com/internetofshit
Not a big fan of this account. It's gone political and completely anti-progress. They systematically denigrate any kind of IoT. I'm not anti-IoT per se, I think most IoT is bad but on the other hand I recognize that it can be pretty useful in many cases and once the tech become more stable it will get better.
What this account is doing is like systematically making fun of all the early flying machines that failed in grotesque ways, until they started working. We shouldn't systematically attack and ridicule any attempt to innovate.
In the early aughts I had a media player that you could plug into your TV, you would control it with a remote and play whatever movies you had on disk.
Nowadays (Apple TV, Chromecast, Fire TV etc), you have to create an account with Apple or Google or Amazon just in order to use the device at all, then another account with Netflix, Disney Plus and every other streaming service you are going to use. Did the OS receive an OTA update that you don't like? Tough luck, there's no easy way to avoid those. And don't even think about playing back files that reside on disk, even if you legally own a copy of the movie.
Sometimes "progress" is not the most desirable outcome, at least from the point of view of consumers.
That Twitter account, the way I see it, is anti-dark-patterns, not anti-progress.
If those early flying machines were mass produced and sold to consumers, they absolutely should be ridiculed. Products do not deserve the same protection as prototypes.
This is why I've kept my home automation offline with zwave. Zwave has its own barrel of shit, but at least it stays inside the barrel.
When I first started playing with HA a few years ago, I got an internet connected hub, and it eventually occurred to me that every time I walked into a room some server somewhere else was notified before my light came on. What an insane round trip to rely upon!! Some server about 800 miles away just told my lights that I was in my dining room.
Granted it was cool that it happened quickly enough to be useful, but what a dumb thing to rely upon. Add to that the unreliability of comcast and I felt like an idiot complaining that my lights weren't working because of my internet connection.
Now I rely upon zwave2mqtt and the occasional spare time to write my automations in javascript.
It's imperfect, but no company in the world will ever know . And no internet required.
On the other hand, it seems most companies want the exact opposite --- the more clueless lusers they can extract profits from, the better.
A lot of people are still surprised when they hear that I work in the tech industry, but don't have - and recommend against - "smart"[1] devices in my house. The majority of my appliances are simple and robust without any computers or electronics, have lasted several decades, and will likely continue functioning for many more.
[1] With the exception of a single smartphone which is rooted, not constantly within arms' reach, and kept off most of the time.
"A lot of people are still surprised when they hear that I work in the tech industry, but don't have - and recommend against - "smart"[1] devices in my house. The majority of my appliances are simple and robust without any computers or electronics, have lasted several decades, and will likely continue functioning for many more."
I have adopted the same strategy and have had very good luck with it.
"TV" is an industrial/commercial display (NEC P461). Microwave is a commercial panasonic (NE-12523). Coffemaker is a Technivorm Moccamaster.
In addition to being "dumb" these devices are very well built, are repairable, and last forever.
> have lasted several decades, and will likely continue functioning for many more
Well yes - Light switches made in 1900 will probably still work today, but you obviously don't have the features of IoT like toggling it from across the house or turning on multiple light [switches] at once. There's always a trade-off and people seem to want to pay for the convenience, even if it means replacing it in as short as 5-10 years.
It's optimizing for a different type of convenience.
If I get a smart toaster, such as the Revolution Cooking Smart Toaster, then I'll get a bunch of nifty features: touchscreen controls, notifications on my phone when the slice is toasted, etc. It'll be nice and a lot of fun for a while. Initial experience is great. A year passes. The touchscreen fades. The wi-fi chip doesn't work with my new router. The notifications app doesn't work with my phone's OS upgrades. I'm jealous of the cool new model at my friend's house. I'll be shopping again in the near future.
Instead, I could get a complicated dumb toaster, such as the Panasonic FlashXpress. It's got a ton of buttons that all do different things. Every morning for the next three weeks is going to be an experiment; I'll need to learn the ins-and-outs. Eventually, I can play it like a musical instrument. From there, I'll be able to use it until it dies, probably ten-plus years. The toaster will not be something I think about, really, at all. It's just a tool that sits there and does the job well.
So here's the question. Do you prefer the immediate conveniences of using the latest and greatest, next year be damned? Or do you prefer the long-term convenience of not ever having to think about it?
That makes sense to me. In software design terms, this is a generalization of the single responsibility principle.
That being said, I there’s still plenty of room for smart appliances that are fault tolerant and degrade gracefully. I agree that most “smart” things on the market today fail to implement either of these features particularly well, but I see no fundamental reason why they couldn’t.
Honestly, despite the claims of IoT devices, usually time saving too. Time saved in setup, time saved in maintenance, sometimes even time saved just in general use.
not having to deal with distractions, malfunctioning devices, and over time saving quite a substantial amount of money I suppose.
The purpose of technology is to enhance and improve human interaction. When technology is perfect it is invisible. These smart devices are in a sense anti-technology, they're gadgets that occupy people's time and attention and get in the way.
The only product in the last decade that actually in my opinion is smart is the Kindle (or equivalent ereaders). It's like having a thousand books I can carry around with the weight of one, it lasts forever, I don't need a light to read, and otherwise I can forget that it is technology and actually just read books.
Kindle's 2007, it's in the same decade as modern smartphones.
I'm struggling to think of any new tech product from this decade that's really improved general life. I'm not that cynical, there's gotta be something. Maybe modern dashcams?
I mostly agree with the article's premise, but I would add an exception for when you automate the solution yourself.
You personally no longer need to solve the problem, but in the process of automation you'll have to learn more about the problem then you are likely to learn just through solving it manually.
I'll go against the grain here but I think it's all fine.. it's part of the 99.9999%; we've just hit the 0.0001%. It'll get better over time and the pros outweigh the cons.
Convenience often trumps those two, unless it is horrifically bad. Like I won’t use an email provider, no matter how convenient, if there is no good authentication.
The risk of big company doing something evil though is typically one I’ll take.
Will it? AWS is the mort reliable service, and even it has outages. Then there are the thousands of other internal service dependencies. Any third party IoT device relies on those and then their own awful software stack and silly putty + paperclips code.
All of this ignores the added hardware complexity and likelihood of failure
AWS doesn't have some sort of magic reliability factor, in fact it's less reliable than GCP/Azure as it doesn't do live migration of EC2 instances.
Ultimately it's up to developers to create a system that can survive during outages.
The statement is not wrong! If GCP/Ature have the same reliability like AWS they will be more reliable than them as there will be no random restarts because of a planned migration of your vm.
Ring works perfectly fine offline if wired up to a traditional chime. Roomba can still clean your house without wifi, you just need to trigger it manually rather than through the app. Every smart plug I have seen has a physical button.
People here need to relax with the exaggerations. This thread is the perfect example of how far away "tech people" are from reality for the average user.
Well, there was the time that the "server down" icon for S3 wouldn't render because it was itself hosted on S3:
>In fact, the five-hour breakdown was so bad, Amazon couldn't even update its own AWS status dashboard: its red warning icons were stranded, hosted on the broken-down side of the cloud.
That's why I'll never buy any "smart" home stuff. With traditional locks, lights, etc they will last decades. No one is going to support 10yr or even 30yr old smart home devices, causing you to have to upgrade all that stuff later.
"What we can do now is probably not a good idea to rely on the same company for half of our Internet infrastructure."
Am I the only one who feels like this is kind of a ridiculous conclusion the article makes? I don't know the author's credentials but it feels like a knee-jerk reaction by someone who doesn't know what they are talking about. I'm not aware of any other cloud vendor that doesn't ever have outages, and I'd be even more worried if my vacuum relied on the manufacturers own data centers. I think a better conclusion is for vendors to use redundant cloud regions and availability zones, and for consumers to maybe not buy internet-enabled devices if %0.001 downtime is unacceptable to you.
Yes, the IoT of things can have some really absurd behavior.
But, it's pretty awesome as well. We are currently out of town for Thanksgiving, and some packages came that we weren't suspecting. We asked one of our friends in the neighborhood to come by our place, and I opened up my garage remotely, and they put it in for us. It super easy and overall awesome.
As opposed to your neighbors just hold your packages at their place, and you grab them when you get back home? Heck in pre-covid times, that would give you a nice little opportunity to say hi and have a chat with a neighbor.
The failure modes of most IoT stuff are just not worth the "inconvenience" most of them are solving for, especially when it has to do with safety or security things and we know how easily hackable most IoT software is.
That specific example was peak "solution in search of a problem".
Now instead of having your neighbour pick up the package, you can watch them from your garage camera, see them look around your garage for a bit (because of course why wouldn't you), leave, and close the door after them!
Services that actually matter don't use this model. Over an appropriate time period AWS availability is exemplary.
However do note. Avoid us-east-1 if at all possible. Understand that despite the documentation, service control planes do not operate in isolation between availability zones. If you need higher uptime you need to have cross regional deployments. When AWS says a service is global, it actually means us-east-1 :)
I made a comment a few days ago on hn where I went off on a tangent about our upcoming dystopian world where everyone lives under a corporate bubble where our company of choice dictates every aspect of our lives.
I didn't include the ability to clean ones home or the ability for someone to ring your doorbell in that rant because that seemed a little over the top...well...thanks world...you proved me wrong.
It hasn't helped that iOS and Android haven't always made it easy to connect to other local devices that aren't bluetooth. Cloud services are handy, cheap intermediaries for lots of device/user interaction. LAN connectivity is a lot more complicated than it seems on the surface, with devices constantly going offline depending on power consumption and what chipsets are in the device (which is a function of what price point the makers were targeting).
And yes, there are technical solutions to this (just add an IoT hub, just add this chipset, etc), but don't forget how cost-driven IoT still is. People want the features of smart devices, but won't pay 5x the cost for it...unless you're Apple making headphones.
IoT is still very much a hack-driven industry. I do 100% agree that the software side is substandard, but until some of these chips get cheaper it's surviving on small margins.
Playing devils advocate here. My power goes out probably three times per year. Water stopped working in my building twice this year. I lost my physical house key last year. My WiFi has issues probably once a quarter and needs to be restarted. My brand new vehicle (automobiles generally have the highest requirements for reliability AFAIK) wont shift into 1st gear (manual transmission) once every thousand miles.
I feel like this community especially, but we in tech tend to forget that everything fails. It is inevitable. We get all riled up about “iot this” and “I never use connected devices that” but honestly I feel like our digital reliability is just as high as the “analog” tooling so many of us proudly retreat to.
Shit fails. Once a month when it happens in a manner like this the HN community gets up in arms like we are the sole arbiters of reason when it comes to technology.
It's not a question of perfect, it's a question of __unnecessary__ fragility/vulnerability. We all rely on electricity because the alternative is far more difficult. We drive despite the costs and problems because most places in the US have been designed entirely around driving and it's very difficult to survive otherwise.
Does my fridge actually need to talk to the internet? Do my lightbulbs? Are the gains from having an addressable coffee maker worth the downsides when things break (or are broken due to security being an afterthought at best) over having a coffee maker that's just a coffee maker?
It's not any different than any other code all of us write on a weekly basis, when you start tacking on unnecessary features and increasing the complexity you will inevitably compromise the stability and simplicity of the features it must have to do its job. It's why the Unix philosophy exists to begin with.
It's a fair point, and you're right. But there's one difference: when your power goes out, it affects your area. Amazon's server outage covers a much larger area. I won't claim "the whole world," since server outages often aren't global, but sometimes they are.
For the sake of not sounding just like I’m complaining, and to actually look for more discussion: do you disagree with the regularity and nature of outages being a constant in our lives, and that a server farm somewhere is just another dimension to the scenario we all accept as normal?
Do you think that big tech and any iot “utility provider” should be held to a higher standard? (Probably IMO)
PS. Didn’t mean to go on an anti HN rant. Love this place. Just seems extra negative to me lately.
A) I find utility outages are rather rare - I can't remember the last time power or water failed at my house (it's been some number of years at least).
B) Devices _need_ power and water (depending on the device). There's no reason my Dyson fan has to talk to an external server to be controlled by an app on my phone, but it does. It's an unnecessary dependency.
Care to elaborate? My assertion is just that, in the end, I’ve had roughly the same experience due to different root causes more often this year due to old school problems than IOT and connectivity bugs. Just saying that random anomalies are the norm, not something (for me at least) to get worked up about
This seems reasonable to me. Disappointing thing is refrigerators and HVACs aren't smart enough.
Ideally I'd love to see a refrigerator which has VSCode built in. So I can keep my cool and finish my story when those agile daily standup meetings start getting hot.
The problem is that the whole architecture is broken. The only reason these devices are "smart" is because they are connected to the brain on the cloud. Even if the device included some smart features, which often ends up cost prohibitive at scale, connecting to the cloud solve tech-illiterate user configuration issues... "Download an app, create an account, register your device which 'phoned home' previously, and you're done. Essentially, the cloud brain acts as a as a hub in the "smart" devices topology.
I stopped vacuuming years ago because of my health issues. At some later date, I saw studies about how they can spread germs and throw particulates into the air.
I have a wood floor and no carpeting. Instead of cleaning a lot, we do all we can to just not let things get dirty/germy, like taking our shoes off near the door.
I spend all my time online but own zero IoT devices. I'm really enjoying some of the comments here, like this one:
My house only has carpet in the bedrooms, and wood flooring everywhere else. It's so much easier to clean than a fully-carpeted house, no need to worry about accidentally spilling drinks in the lounge or accidentally tracking dog shit through the hallway. Much easier to fix any damage as well.
Best of all, it's much less dusty than a carpeted house, since the floor isn't covered with a giant magnet for dust and crumbs.
If your condition is serious enough and incurable, not allowing dust to exist and wiping down wood floors instead of vacuuming is a better option than either of the two options you posit as the only possible outcomes.
How do you deal with skin waste etc.? Reducing dust is certainly possible, but I thought it is just impossible to bring it even remotely close to zero to make vacuuming unnecessary.
Maybe from your end it looks that way. It doesn't from my end.
I've done a lot of "surprising" things, healthwise, over the years and there seems to be no means whatsoever to make people on the internet happy with how I talk about it. No amount of bending over backwards to signal "this was something really, really, ordinary, I swear!!!!!" is ever enough to protect me from people reacting with animosity and I almost never know if it's because they do recognize me and know something about my back story or if it's because they don't recognize me.
For me, it's a no win situation and it's not possible for me to guess what the magic words are that will keep everyone happy and get me out of the corner with the dunce's cap for the crime of not dying years ago from my serious medical condition because I'm so incredibly rude like that.
For the sake of your own mental health, you should stop caring about what makes random internet strangers happy. There will be always a minority reacting with animosity to even the most milquetoast of content, and responding in frustration will only make it worse.
When animosity reaches a certain threshold, ignoring it is actually dangerous and not merely to one's mental health.
I didn't respond in frustration. That's one of the issues here. Other people act weird, tell me there's something wrong with how I phrased it, I engage them conversationally and now some other random stranger replies with a comment that can be interpreted to mean I'm merely neurotic and oversensitive.
That boils down to "When other people react weirdly to your prosaic observation, you cannot engage them. You are required to try to somehow rise above it and not speak to them. And when this becomes an entrenched pattern, it's somehow your fault for existing."
It seems really unlikely that vacuuming is worse for your health than not vacuuming. But maybe a good air filter and a roomba that runs while you sleep? I actually do that already but really because I’m lazy and don’t want to have to dust often.
“ More than just dirt, house dust is a mix of sloughed-off skin cells, hair, clothing fibers, bacteria, dust mites, bits of dead bugs, soil particles, pollen, and microscopic specks of plastic”
Most of those have nothing to do with carpet or furniture. And it’s hard to imagine that whatever a vacuum does to kick them up is worse for you than just breathing them in all the time and disturbing them when you walk.
But really, that’s a case for a robot vacuum because it can eliminate the stuff when you’re not there.
You are basically dismissing my firsthand experience. Maybe that's not your intent, but you start off It seems really unlikely that vacuuming is worse for your health than not vacuuming.
I'm telling you from firsthand experience that I don't have a lot of dust and my health is better these days and I don't own a vacuum. I'm not asking you to give up your vacuum.
If you want citations supporting the idea that vacuums can spread germs and particulates, here's one:
If you want to continue to imply that I'm wrong to choose to live without a vacuum cleaner, I'm done. That's out of bounds for debate. It's a personal choice.
There’s basically no chance that’s real but I don’t want to get into one of those weird HN arguments about whether or not vacuums can really make someone less healthy in response to an article about an AWS outage so you do you!
Is there anyway to route what alexa’s microphones pick up to another system (say a raspi) so I am certain it is not listening in when I don’t say the word Alexa. There could be primitive voice recognition on the raspi that would then pass on the full command to Alexa and then on to Amazon for normal processing if it recognizes the command starts with “Alexa”. Is this in the realm of possibility?
Didn’t this happen already a few years ago? massive S3 outage happened in 2017. Some connected devices were very confused. Mayhem..of a not-minor variety.. but also a tad hilarious:
[..] On Tuesday morning, members of the S3 team were debugging the billing system. As part of that, the team needed to take a small number of servers offline. “Unfortunately, one of the inputs to the command was entered incorrectly and a larger set of servers was removed than intended,” Amazon said. “The servers that were inadvertently removed supported two other S3 subsystems.”
The subsystems were important. One of them “manages the metadata and location information of all S3 objects in the region,” Amazon said. Without it, services that depend on it couldn’t perform basic data retrieval and storage tasks.[..]
[..] While S3 was down, a variety of other Amazon web services stopped functioning, including Amazon’s Elastic Compute Cloud (EC2), which is also popular with internet companies that need to rapidly expand their storage.
Amazon said S3 was designed to be able to handle losing a few servers. What it had more trouble handling was the massive restart. “S3 has experienced massive growth over the last several years and the process of restarting these services and running the necessary safety checks to validate the integrity of the metadata took longer than expected,” the company said.[..]
And this!!
[..] It’s also making a change to the AWS Service Health Dashboard. During the outage, the dashboard embarrassingly showed all services running green, because the dashboard itself was dependent on S3. The next time S3 goes down, the dashboard should function properly, the company said.[..]
The solution is geo-redundant API endpoints with Anycast. I imagine few IoT gimmicks are launched by vendors who both have the resources to implement this, maintain it and actually care. Keeping a geo-redunant data set in sync is pretty tricky, considering a lot of devs can't even get it right locally.
Reminds me Philip K. Dick’s short story “The Day Mr. Computer Fell out of Its Tree” (written in 1977).
"It had been a bad idea from the start, back in 1982, to operate every mechanism from a central source. Of course, the basic idea had sounded good: with the ozone layer burned off, too many people were behaving irrationally, and it had become necessary to solve the problem by some electronic means immune from the mind-slushing ultra violet radiation now flooding earth. Mr. Computer had, at the time, seemed to be the answer. But, sad to say, Mr. Computer had absorbed too much freaked-out input from its human builders and therefore, like them, Mr. Computer had its own psychotic episodes."
1. It's fine, when you buy a smart or connected device, you want the extra features, and you consciously accept that connection issues will impact your products. This is similar to the way electricity issues will impact products that require electricity, and products like a kettle can be offered in electrical and non-electrical versions.
2. It's not fine, it is not clear that internet/service issues will impact your product, and the more people buy these the more it'll become a thing, impacting everybody else. Products that can afford it (like a doorbell) should be made to work even when services are down.
I bought a large dumb 4K TV (hard to find!) from Haier for cheap at Costco. Love it.
It’s connected via Apple TV 4K, which I can easily upgrade whenever I want.
I have Lutron Caséta in-wall smart switches which communicate with their hub. Not a single delay or bug in 8 years. And they are intuitive for non-techies to use. It’s just a light switch.
I also have some Philips Hue for accent lighting. It frequently fails to respond.
I think HomeKit is better because it at least has to pass Apple’s certification process for some minimum bar of quality.
It’s also nice that all HomeKit stuff works through the ControlCentre app. You can 100% avoid ALL manufacturer-specific apps for daily use.
Does HomeKit rely on a cloud infrastructure? That is, if my Comcast internet goes out (but my local wifi is still up), can I turn on/off my lights via the Home app? Can sensors I have still trigger actions?
I'm thinking of getting various HomeKit compatible plug/switches/sensors and am trying to figure out what the communications path is.
Just bought a Ring device on a whim, so I can get notifications on my phone when I'm out of range of the doorbell. It arrived today. Spent quite some time wondering why it wouldn't finish the setup or work at all... it gets literally stuck in a loading state when you press the doorbell button.
I 1) am appalled that connection loss is not dealt with correctly, and 2) did not expect it to stop functioning completely without internet as it does not require a cloud subscription.
Single Point of Failure... is a phrase we'll be hearing and conversing about a lot more in the coming years.
As SaaS, Cloud Providers, and Other Tech companies solidify their necessity for the function of everything else... aka centralization around 4-5 tech companies which their infrastructure literally runs everything else.
It only stands to make us more and more vulnerable as a society.
It's like having one power station powering the entire electrical grid, monumentally stupid.
IoT wouldn’t be such a shit show if we designed in good manual failovers. There should always be functionality available at the next lowest complexity level.
Why would anyone buy a "smart" doorbell with batteries, lighting, and other useless crap only there to fail? I could see it's convenient when renovating and having forgotten to do the wiring, or to pimp a condo for sale, but otherwise you can get the same Honeywell ding-dongs for 10-15 bucks plus converter plus 8V panel that we've used for ages, and that last basically forever.
It’s just Amazon being a dick. Smart home stuff, especially with home assistant, works well and without internet, self hosted, many things even have fallbacks for when there is no connection to your home assistant server. But Amazon makes smart home the iot hellhole that we thought we left behind. Well then.
And no, smart home is convenient even for normal houses.
Perhaps next we will connect our toilets to the IoTs, so that the waste can be disposed of by smart routing of what's in the toilet. But then, oh shit, what if the network goes down? Have to hold it 'till Amazon servers come back up. :-)
When things like this go down, it makes me wonder what would happen if Twitter were built on the affected infrastructure. What would be the backup platform? I'd look here (HN), but where would others look? Facebook? LinkedIn?
Couldn't these companies have lambda's running on Azure, not costing them anything when not being used. After say three failures to connect to AWS the service would change to Azure for say 24 hours?
As another commenter is right to point out, a multi-region setup would suffice—no additional cloud provider(s) necessary. This outage says more about companies failing to architect adequate failover than it does about the risks of the cloud.
Hosting your own local server has it’s pains too, and most people doesn’t know how. The down rate of cloud services still likely have better up time than most can guarantee with their own servers.
We've got off easy this time. Wait a few years and a major outage like this will leave your "self-driving" car stranded in the middle of nowhere or crashed into a lane divider.
Aside from these same couple tweets every news story cites, is there any documented evidence of the exact impact on smart vacuums and doorbells? What models? What happened specifically?
I had one of those. And a $50 Walmart vacuum bests it in every way except proven longevity. Power tools have come an insanely long way since the 1970s.
If you need a robot vacuum because using a regular one isn't a realistic option for you, whether it should require internet access is irrelevant. It does, and making fun of people whose fault it isn't is mean.
It really isn't irrelevant, it should be top on their list of things to know about. As we're seeing here, it's a source of unreliability and if you want a robust home automation setup you should look for tools that don't require internet connectivity.
It is not consumers' fault that they have to go out of their way to find smart appliances that don't fail when servers go down. It's genuinely hard to tell by looking at something whether it can cope with only having access to the local wifi.
If my smart plug and my phone are on the same WiFi there's no good reason why they can't communicate without a go-between.
I work at an IoT company, not for home automation, but GPS vehicle tracking and temperature monitoring of commercial fridges, among others. I agree that requiring connectivity for essential home appliances and door locks is downright foolish.
My mum saw this news, and asked me "Did this affect your client's IoT system, Peter?" Here's my reply.
One of their computers is called US-east-1. That is the one that broke. The other Amazon computers (e.g. EU-west-1 in Ireland, AP-northeast-1 in Tokyo) were all fine.
Some of our programs run on AWS, but we are using the computer in Sydney (AP-southeast-2). So our services weren't affected.
Only the people using US-east-1 were affected. This is the default computer for many people in the US. Even big companies like Google and Facebook sometimes use AWS computers.
We also use DigitalOcean's computers for some other clients. That's like AWS but cheaper and slower. I helped copy files from one computer to another. That's called "DevOps".
Peter
• Technically it's a huge server farm with many computers, but they all have the same basic inputs and outputs: power [0], Internet access, air conditioning.
Anything in the same physical location could be affected by earthquakes, fires [1], power cuts, Internet connections going offline, etc. This is why daddy likes to keep an off-site backup, for safety.
The government spies on people by seeing what is happening on Amazon's computers. Wikileaks [2] has more detailed information about who is doing the spying in each place (usually a company around the corner).
> Anything in the same physical location could be affected by earthquakes, fires [1], power cuts, Internet connections going offline, etc.
It's worth noting different AZs within the same region are supposedly architected to have their own redundant power and Internet connectivity. AFAIK a single fire shouldn't take out more than one, either.
If a region-wide outage is caused by any of those things, it's because something else isn't working as expected. And this region has had region-wide issues before.
Still, boiling cloud computing down to "somebody else's computer" is an oversimplification I see too often from cynical techies. It's like calling a restaurant "somebody else's kitchen." Yeah, it is. But it's missing the point to frame it that way.
A remote server should not be an absolute prerequisite for a doorbell to function. That feels like complete nonsense. Like sure, the fancy cloud features should fail, but there is no reason that it shouldn't continue to operate in a degraded state like say, only making a noise indicating someone pressed it. If you have hardware sophisticated enough to interact with an API it should also be sufficiently powerful to do operations we could successfully do locally with much more primitive electronics.
I hope I never have a toaster that's too stupid to count to 4 minutes without connecting to a lambda.
Plus, you can currently solve a blackout with a generator. Or a backup internet connection in case of your ISP being down. Sounds like neither would solve an issue with these sorts of devices requiring remote servers to be online.
You're narrowing down my line of reasoning to a specific example and then disputing it. But you didn't answer to the overall fact that there are central points of failure in most electronic devices/infrastructure that are even more critical than a vacuum cleaner. Yet we still rely on them. That doesn't make the whole "electric" infrastructure "a mess".
And BTW not every blackout can be solved with a generator. For instance a city-wide blackout, in regard to streetlamps etc. Also, not everyone can install a generator. Not every landlord allows a generator etc etc. Even if you had a city-specific generator, that can just as well fail. Now, if your not concerned with a backup system failing you shouldn't need to be concerned with Amazon's system failing as they also have multiple backup systems in place.
I was talking to @jessaustin and not really interested in debating the entire topic. Just saying that an individual in a given situation can have some control over things like power and connectivity, but zero control over their IoT device working if the server goes down.
― Philip K. Dick, Ubik