News: 1608215707

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Stony-faced Google drags Android Things behind the cowshed. Two shots ring out

(2020/12/17)


Google is [1]discontinuing its Android Things IoT platform for non-commercial users. The Chocolate Factory will not allow the creation of new projects after 5 January and the entire platform will be nuked the following year.

This means that all project data – from device images to build configurations – will be permanently deleted. Non-commercial users will also be prevented from shipping OTA updates to devices.

Google first dipped its toe into the murky waters of IoT development during the 2015 edition of its I/O conference with [2]its Brillo operating system . The following year, this [3]was rebranded as Android Things .

At its heart was a slimmed-down version of Android that would happily run on low-power compute boards. The Raspberry Pi 3B, the NXP i.MX7D, and Intel's ill-fated Atom-based Edison platform, which was [4]quietly discontinued by Chipzilla in mid-2017 , were all touted as hobbyist options. Commercial users could take advantage of specialist platforms from the likes of MediaTek, Qualcomm, and NXP.

On the software front, developers would be able to use the standard Android development tools, languages, and frameworks they're familiar with, while also taking advantage of APIs designed to play nice with the wider IoT ecosystem.

In theory. It took almost two years for Android Things to hit its 1.0 release, and commercial appetite was tepid at best, with just a few vendors signing on. Nine months later, Google said it would refocus Android Things to target a smaller scope of products, namely smart speakers and displays. That's a huge leap from its original plan to create an all-encompassing IoT platform that would power everything, from art installations to more sophisticated consumer and industrial kit.

Parallel to this, Google largely refocused its IoT strategy away from offering an underlying operating system to using Google Assistant on other third-party devices. This manifested itself in January 2019, when Google launched Google Assistant Connect, which would allow manufacturers to bake Assistant into their products regardless of what underlying operating system they used.

But what about hobbyists? Well, it's not like they don't have much choice: from the myriad IoT-oriented Linux distributions to Windows 10 IoT Core, there are plenty of options available. And that's without mentioning the likes of Arduino. It's not as though the withdrawal of Android Things leaves a major gap in the market.

Still, it's annoying for those who actually signed on. Google doesn't say much about continuity, except that non-commercial users can download their images and manually flash them to their devices. Beyond that, it seems like the best option is to switch to another platform with long-term support.

If you're a hobbyist caught short by this news, we'd love to hear from you in the comments section below. ®

Get our [5]Tech Resources



[1] https://developer.android.com/things/faq

[2] https://www.theregister.com/2015/05/28/google_iot_platform/

[3] https://www.theregister.com/2016/12/22/brillo_android_things_google_nest/

[4] https://www.theregister.com/2017/06/20/intel_joules_edison_galileo/

[5] https://whitepapers.theregister.com/

Adrian 4

Has anybody produced a successful IoT platform ?

It seems to me that the description is too wide - no one platform fits all, and the solutions range from home-grown to Linux to some other proprietary platform.

A framework that needs a Pi 3 upwards is strongly into the top tier of products.

Anonymous Coward

>> to me that the description is too wide

Indeed, as defined (and abused) a successful IoT platform would replace the internet.

I mean, isn't everything on the internet a "thing"

A thought

Mage

For years people made Routers, TVs and other things like eReaders just with Linux. Sony strangely went from Linux to Android for the PRS eink readers, even though the main point of Android is the touch GUI and the Java like Android apps on a VM. Sony's PRS305 and PRS505 had better more reliable software than the Android versions, starting with PRS-T1, nor did it save the product. Problem was Sony's idea was to make money from the bookstore. That's really only ever worked for Amazon.

Some stuff used VxWorks instead of Linux.

TVs adopted Android TV, to allow Android apps, but otherwise give a worse user experience and of course lack of privacy.

1) Anything with Internet and Thing in the name or description is a bad idea. Adding Android too seems daft. Any arbitrary gadget with any OS and an comms API can be controlled from iOS, Android, MacOS, Linux and Windows pretty trivially. Having Android on it makes no difference.

2) Is there any point to Android other than on OLED/LCD touch screen phones and tablets?

I've wondered does Google pay TV makers and eink reader makers and auto systems money to use Android, or is it just some manager in a company thinks it's a better idea than vanilla Linux?

Well, my terminal's locked up, and I ain't got any Mail,
And I can't recall the last time that my program didn't fail;
I've got stacks in my structs, I've got arrays in my queues,
I've got the : Segmentation violation -- Core dumped blues.

If you think that it's nice that you get what you C,
Then go : illogical statement with your whole family,
'Cause the Supreme Court ain't the only place with : Bus error views.
I've got the : Segmentation violation -- Core dumped blues.

On a PDP-11, life should be a breeze,
But with VAXen in the house even magnetic tapes would freeze.
Now you might think that unlike VAXen I'd know who I abuse,
I've got the : Segmentation violation -- Core dumped blues.
-- Core Dumped Blues