X
Business

All your IoT devices are doomed

Until we look to emulate the open models of Wi-Fi and PCs, IoT will always be a source of never-ending e-waste.
Written by Jason Perlow, Senior Contributing Writer
wall-e-junkyard.jpg
Image: Pixar

First, it was Aether's smart speaker, the Cone. Then, it was the Google Revolv smart hub. Then NetGear's connected home wireless security cameras, VueZone was killed off. And in March of 2019, it was the Jibo cloud-connected robot. And while it's not rendering the products into e-waste, Google is cutting off Works With Nest functionality.

I'm sure I've left out more than a few others that have slipped under the radar. It seems like every month, an Internet of Things (IoT) device becomes abandonware or less useful after its cloud service is discontinued.

Many of these devices, once disconnected from the cloud, become useless. They can't be remotely managed, and some of them stop functioning as standalone (or were never capable of it in the first place). Are these products going end-of-life too soon? What are we to do about this endless pile of e-waste that seems to be the inevitable casualty of the connected-device age?

The problem is not as simple as IoTs and the companies supporting them going belly-up. It's more along the lines of the vendors themselves being willfully negligent in providing a path to community support for these devices and penalizing their early adopters.

It isn't just IoT either. It is anything that is dependent on cloud services -- which provide an update mechanism.

This includes very mainstream products like Apple's older generation iOS devices such as the iPhone 6, iPad Air, and the iPad Mini 3. All are popular models that were released in 2014 that will be unable to upgrade to the latest iOS 13 due next month

Without the iOS 13 update, there is the risk that those older iPhones and iPads might not be able to update third-party apps from the App Store at some point in the future. Developers may be forced to drop legacy device compatibility in their apps due to Apple's stringent OS version compatibility policies that have presented similar issues in the past.

They won't become "unusable," but a lot of functionality is in jeopardy. And there are a lot of these devices in existence. 

As an industry, we need to step back and think about the realistic lifetimes of IoT and smart devices, and what can be done to extend their lifetimes when they are at risk of abandonment.

The expected lifetime of an IoT device should probably be based on the type of device. I like to think of these devices as belonging to three, distinct groups: endpoints, hubs, and clients.

An endpoint is a device managed by something else. These are devices that if unmanaged, should still be able to function without a working cloud service. Examples include light switches, light bulbs, lighting fixtures, alarm system and sensors, smoke detectors, sockets, thermostats, fans, air conditioners, and video cameras, as well as major home appliances like refrigerators.

For the most part, endpoints are single-purpose devices. Because of their simplicity, there should be the expectation that they are also the longest-lived, with service lives of 10 years at least.

For that service life to be a reality, the management protocols need to be open. Many endpoints today use Wi-Fi as their communications mechanism, but others on the market are starting to embrace Bluetooth Low Energy.

While the Wi-Fi (as well as the overlaying TCP/IP protocol stack) and Bluetooth Low Energy specs are open in and of themselves, the management APIs and profiles used by these endpoints are not. You also have manufacturer-specific differences in the way they implement their management protocols over other wireless communication standards, such as Zigbee and Z-Wave.

While it may not behoove the endpoint manufacturers to make their devices interoperable with competing endpoints, as a consumer I would be extremely wary about giving any endpoint manufacturer my continued business unless these companies form a consortium and make a concerted effort to create a long-term interoperability specification.

This specification needs to be open and extensible and should be easily understood by the consumer from an industry-branding perspective, such as Wi-Fi or Bluetooth. Samsung has the right idea with SmartThings, but I think it needs to go much broader than that.

Any IoT communication and control standard has to be as accepted and interoperable as Wi-Fi is today.

How pervasive and open is Wi-Fi as a standard? Extremely. Most Wi-Fi devices over ten years old still function on modern networks even though security and functionality changes have occurred over successive generations.

Sure, a lot of that WEP-only stuff became junk. But WPA and WPA2 stuff still work fine. Even slower 802.11g and 802.11b equipment with WPA-TKIP again function provided those modes are turned on in new access points.

Arguably, these devices are well outside their expected useful lifetimes, and from a security standard, should be considered vulnerable and discarded. But this is an excellent example of how interoperability should work from a long-term perspective.

With the way things are now, you're not going to be able to say the same about your IP-controlled light switch or Bluetooth/Zigbee/Z-Wave lightbulb in 10 years. Probably even less time than that.

Hubs control endpoints. Typically, they act as the primary communication mechanism to a cloud service, have converged functionality, and may coordinate firmware upgrades. This includes things like Philips Hue, WeMo, and third-party hubs like Amazon Echo, Google Home, Apple Homepod, and Smartthings.

Some endpoints are not hub-managed; they may be entirely cloud-service dependent. This design weakness makes them especially vulnerable.

For simple endpoints, let us look into the need for an open-hub standard, and potentially, an open-hub operating system that can communicate with plug-in managing cloud services, which can be swapped out accordingly.

While I would not rule out Windows 10 for IoT, something based on Linux might be a good contender, because it supports so much hardware and could be easily virtualized or containerized inside many common computing infrastructure devices used in the home, such as a set-top box (like a game console or cable device), Wi-Fi router, residential gateway, or perhaps one of your PCs.

All you would need is a virtualization or container host and fire it up. Your host just needs to be able to communicate over TCP/IP and Z-Wave/Zigbee/Bluetooth Low Energy (and really -- why the hell do we need three standards for that?) for that virtual hub to manage those devices.

And because it would be virtual, you could back it up and move it to alternative hardware if needed, such as your smartphone, which has the benefit of being able to talk directly to a mobile network.

Hubs implemented in a virtual fashion running on an open specification should be able to live indefinitely. That means IoT abandonment along the lines of Revolv and NetGear's VueZone would not happen nearly as often as they do now.

And clients?

We already know how long clients should live -- and the more open and flexible they are, the better off we are as a whole.

I think we can say there's no hope of Apple opening its specifications on any iOS hardware so that they can be re-used or have their lives extended, which is a shame. Nobody should expect iOS devices to have a useful life of more than four years.

In theory, devices that use Android are somewhat better off, as there is a large community of open-source developers that can write alternative firmware and obtain the kernel sources needed to support the specific device drivers for the hardware.

But the reality is most Android devices become e-waste faster than iOS devices do, because Android OEMs do not release Android updates in a timely fashion, and in many cases, these devices live out their lives on a single Android version, never to be upgraded.

And it's no easy process for end-users to "root" their devices to bring it to a community-supported ROM like LineageOS.

The consolation is that Android devices are getting cheaper and cheaper due to heavy industry commoditization, so the "pain" to the end-user of having to ditch their devices more frequently than an iOS user is probably a wash.

That's fine when your only concern is how this affects your wallet, but if you have any concern for the environment, this is alarming when you think about how much e-waste is being generated from all of this.

Apple's policies that force obsolescence may be maddening, but unlike the collective of competing Android OEMs, it does have a strategy in place to recycle obsolete devices.

Now, I haven't seen any financial incentive for end-users to turn these things in so that they can act responsibly, but I suspect (or at least hope) this is something Apple is working on.

The only clients we know that have extended lifetimes are x86-based personal computers.

They are long-lived because the architecture is well-known, and despite their increasingly disposable nature, they have a rich third-party parts replacement industry -- and they run Windows or Mac OS, which have historical commitments of multiple generations of upgrades from their respective operating system software vendors.

But PCs are also usable beyond vendor service life because they have an open-source community not tied to those vendors at all. You have something your PC hardware vendor doesn't want to support anymore? You can always install Linux on it.

Sure, not everyone will go for this option. But at least you have a choice, which is not something you can say for anything represented in the current generation of IoT.

Until we look to emulate the open models of Wi-Fi and PCs, IoT will always be a source of never-ending e-waste.

Will the IoT e-waste ever end? Or are we doomed to repeat forced obsolescence endlessly with our smart devices? Talk Back and Let Me Know.

10 Linux distros: From different to dangerous

Editorial standards