Some Different Musings on the Future of the IoT, and 5G Especially
I don’t really get why, but thermostats on portable heaters, and most air conditioners, are awful. It seems like simple technology, but they are simply useless. To keep the greenhouse at a remotely reasonable temperature, we have to go out there three times a day, and change the settings.
But, the other week I got annoyed, googled, and found this:
It’s a thermostat! An inline one, so it just switches a line on and off, plug anything you want into it. Accurately. So far, been monitoring with a different thermometer and it is as specified, holds temperature properly, with about a 2°F droop.
But… not Internet connected. I got a proprietary one, with a remote and box with the switch. No internet, no service fees, no end-of-life because the VC ran out, no hacking unless you drive over with some signal analyzers. And what would you hack? It’s Just A Thermostat. No camera, no audio, nothing else.
5G
The real IoT boosters won’t understand why I didn’t get some connected thermostat for this. And if I complain about standards or BT/WiFi range and so on, might say: just wait for 5G, it’ll fix it.
Okay, I am not even going to talk about the health risks. And if anyone confuses radio waves with a virus, there’s just no talking to them. But operationally, how it will work for networks, is… interesting.
5G will be very fast, with very (for wireless) low latency, and in no small measure because it is very high density. There are lots more BTSs, so you are closer to one. Okay, this is actually because it’s very high frequency, so useful signal only goes say 500m. And, doesn’t go through anything, so will bounce about so I hope the multipath algos are well written.
All this is supposed to empower the IoT, but — even aside from fears, from poor security, from government privacy risks, because it has no backhaul, etc…) I think it may well help kill it.
Because that’s a hell of a network for a light bulb or thermostat. Just start with cost. It is inconceivable that any mobile operator (at least in the US) will charge less than say… $5 for a SIM to pop into your IoT devices.
But we want to IoT everything? And, to be clear, 5G is supposed to not aggregate, to eliminate the Home Hub concept, and everything directly connects.
That means my light bulb has a subscription. A $5 per month subscription. If every appliance and light and door and window is IoT connected, I am paying $1000 a month to connect the “smart” stuff in my house?
Better Networks Promote Worse Practices
Worse to me is that it will encourage the existing bad practices of code, data storage, privacy, and speed. Most IoT devices need a slow (and cheap!) network. Send a tiny signal, often “whenever.” A thermostat doesn’t need to be near real time, but can talk to the house or you on a delay up to minutes, and suffer no ill effects.
Long ago (in internet terms) the concept of IoT enabling networks gave legit credence to QoS (Quality of Service). Presumably, with some economic incentive so we don’t end up with the usual situation where everyone uses the special case to have top priority and no security. So the thermostat says “no worries, whenever, but I do need a delivery receipt.” The light bulbs say “pretty fast, within 2 seconds how about?” And the speakers insist they must match human perception speed for streaming, but they pay for it.
Oh, and encouraging better practices will also mesh up nicely with low power modes. Mobiles, wearables, and many IoT home devices are battery powered. The more we can get things that last for years instead of days, the best.
Smart is Stupid
Anyway, let’s get back to the way my connected thermostat works. Because it’s way, way too much like the way most IoT and wearable and other devices work.
The box that’s plugged into the wall is just a switch. And a thermometer, and radios. But, it has no intelligence. It regularly sends the temperature to the remote control, which then compares to the setting, and tells the switch nothing, or to toggle the settings.
Which. Is. In. Sane.
And before you make fun of my cheap German-market Chinese-made non-smart thing… this is how almost all IoT devices work.
Everything we build is offloading very simple processing to the phone, or to a cloud server you don’t own. The right way to do it is to make the website, mobile app, or actual remote control a Remote Control. It tells the box the settings to abide by, then it can work when the network drops, when the service dies, when the power is out, or so on.
Yes, I’ve worked on a few IoT / wearable systems, done some investigating, done a lot of detailed software design stuff. This works, and if you dig down, is really how a lot of the core networks were designed back in the day. BLE has some neat stuff to support very low-power, very intermittent connections.
Security and Privacy is Sales and Marketing
I’m old enough I remember when networks were so slow you could watch the screen paint. Every time we make the storage larger, the computer more powerful, the network faster, software gets worse.
High speed encourages poor practices, and 80% of the speed increase is lost as a whole javascript library comes down the line for each web page you load.
This is actively killing innovation.
We’re already seeing people say never again, every time something bad happens to their connected devices. Ring collaborating with police, Sonos bricking recycle mode, and more every day.
I think we’re on the cusp of ruining the IoT, and the entire concept of connected appliances, forever. For decades at least.
And… that’s it. This is why it’s a rant, not an article. I have no solution. It’s structural. It’s infrastructure worth billions driving it, and a million developers not paying attention to anything but the boss saying “faster” and also, “why not grab the users’ location and address book while we’re at it?”