Google Developing “Brillo” Internet of Things OS Based on Android | Hacker News Reader

Google Developing “Brillo” Internet of Things OS Based on Android

arstechnica.com

26 pointsingve22 comments

"aimed at ultra low-power devices with as little as 64 or 32MB of RAM."

I know Moore's Law and everything, but damn I must be getting old because that's not low-power by my standards.

My low-power standards are an MSP403 microcontroller running on a coin-cell battery.

Granted, it's going to be hard doing some level of interesting IP networking on that kind of hardware, especially with the increasing requirement of cryptography.

Maybe something slightly beefier running Contiki [1], but I can't resist the reactionary thought that anything Android- (or even Linux-)related is way too fat for many IoT applications.

[1] http://en.wikipedia.org/wiki/Contiki

Yeah, I spit my coffee at that line. That's cool for an IoT hub application, but not for a node.

Well, they do ship a mobile OS where 3GB of RAM aren't enough[1].

[1] http://www.gsmarena.com/googles_android_511_lollipop_in_work...

You must interpret low-power devices as low-end Linux boards. No one is aiming to target the microcontroller market with a Linux/arm board.

He is right. We run contiki with 128 bit AES on these systems. We need 30KB ROM to run a full ip stack.

I realized the other day that in this race to become the nexus (pun unintended) of home automation/IoT, the horse I think most likely to win the race is Amazon. The Echo just came out with updates that control wemo switches and hue lights, meaning you can now say out loud in your house 'alexa, turn on the kitchen lights', and your kitchen lights will turn on.

People have also made hacks to allow it to control a nest, which seems like something that would be an actual feature (since the same person also hacked in hue lights before amazon did it officially).

And the edge that Echo has is that people aren't buying it for Home Automation, they're buying it on its own merits.

A lot of that is simplified by IFTTT having channels for Nest, Hue, Echo (via Alexa), and Wemo.

I don't mean to be contradictory. It's exciting that we are getting glimpses of internet connected things! I wonder about IFTTT alternatives, especially self-hosted open options that can enable those interactions and more from a community.

I wish Google would focus on making regular Android more stable instead. Given the memory management issues I've had on my Nexus 5 and other devices since Lollipop, there's no way I'd choose to inflict Android on a low-memory IoT device...

Google is not a person. Google is a transhuman organization with over 60,000 individual consciousnesses taking part in it. Google can do more than one thing at a time (if their management infrastructure doesn't require a single person to approve every action).

Java for low-power devices. This is some high-level trolling by Google!

There is not a single mention of Java at the Brillo announcement.

Irrelevant but flush your biases and assumptions down to the toilet. Java can statically compile to microcontroller architectures with minimal runtime footprint. Stop blaming the languages, blame the runtime implementations.

Well. Even a single person can work on more than one task in a day. But much like a single person, the Android team has limited time and energy to allocate.

In my perfect world, the whole Android team spends the next three months doing a bug bash and releasing an Android 6 where I reliably get Hangouts messages and don't have to reboot 2-3 times a week. If some of the Android team is working on stripping down the OS for IoT devices, less time is being spent on that hypothetical bug bash.

Of course this is all moot because fixing bugs doesn't seem to be a priority for them regardless; adding stickers is far more important. :)

I'm amazed Android 5 wasn't a trainwreck considering the scope of changes and speed of development.

5.0 was pretty bad on some devices (I got lucky with the Moto X, but if I was a Nexus owner I'd have been really annoyed). 5.1 is massively improved, but it's a pity it took them an extra 6+ months to roll it out.

I'd love it if Google took more control off the carriers, and took responsibility into their own hands. If Apple can roll out updates in a couple of days, Google should be able to do much better than they have been doing.

Here is the dirty secret: Carriers are sold all kinds of creeping horror to put in their handsets. This badware collects all kinds of information on you. As with a lot of badware, the quality is crap. The engineering is optimized for cost, not quality. So updates take forever.

This is why carriers are highly resistant to allowing Google to manage the software on phones. This is why Nexus handsets are mostly sold direct. This is why Google was tempted into exploring being an MVNO.

There are a variety of interesting small-ish RTOSs out there for IoT SoCs and devices. Most of these are are significantly smaller-footprint than a full Android implementation.

BUT the 64MB RAM target they are aiming for could be reached by a "headless" variant of Android. The Android runtime does a lot of interesting things to minimize memory use and power consumption. Android's installable service components could make updatable, highly flexible IoT nodes realistic.

In the context of the (scant) facts in the article, it does make sense.

It doesn't make sense if you think of an IoT node having only a few kB of program memory, a sensor io link, and a Zigbee radio. These ain't going to be $2 nodes.

>> Android's installable service components could make updatable, highly flexible IoT nodes realistic.

In concrete terms , what are the benefits of those to IOT nodes ? Aren't IOT nodes mostly fixed function, only updated for security - and there are solutions for that.

Brillo contains a very minimal part of Android for HAL and the network stack. 64MB is our medium-end target. The ART is not in the scope of Brillo.

I feel like this is a reactionary measure after Microsoft announced Windows for IoT, and Facebook announced Parse IoT.

Google started talking about Android @ Home at I/O 2011.

Google is building their own hardware, an embedded Linux distro is primarily a need at Google to support our hardware projects for a long time. But, for the first time, the industry is capable of building hw projects at scale. It's not surprising we reinforce our interest to get involved by opening what we're already doing.

And it will be called .. wait for it.. the Internet of SLOW Things !

Reply on news.ycombinator.com