Arm to IoT devs: Go faster with our pre-made chip subsystems

The belief that IoT would become Arm's main growth engine never played out as owner SoftBank Group thought it would, but the British chip designer is still doing what it can to keep IoT developers hooked with a bevy of new offerings meant to significantly speed up development. On Tuesday, Arm announced its first major …

[martinusher]

Too much of the wrong sort of technology

IoT seems like it should be a major growth area but the technology is focused entirely on turning 'things' into peripherals of some proprietary cloud based ecosystem. This may seem to make sense as a business plan but its not that welcome to customers because it takes everyday household objects and subjects them to the security and reliability (and latency) issues of the public Internet and the ebb and flow of corporate business decisions. ("...and all I wanted to do was make a cup of coffee")

To illustrate how absurd this is take the example of the thermostats in my house. These are "WiFi" enabled and with appropriate configuration I can control their setting from a voice assistant like an Echo. When I ask the Echo to change the thermostat setting the echo parses my speech in the cloud -- fair enough since we're not quite at the state where we can do this locally -- but then Amazon's cloud has to open a conversation with Honeywell's cloud. Once they've exchanged credentials and mentioned the thermostat is to be changed the Honeywell cloud waits for an opportunity to push the request t the thermostat. This takes about a second and relies on all the relevant Internet and cloud components being functional. I could, of course, just step over and press the button on the thermostat but it would mean cutting out corporate players -- I'm pretty sure Amazon has no interest in my thermostats but judging from feedback through my email (mandatory) Honeywell is very interested in my domestic arrangements.

I tolerate the thermostats because I can just get up and "give them a re-programming they'll never forget" if they don't behave themselves but the prospect of mission critical devices being beholden to corporate whims -- that's just a bit too much. We should have learned our lesson from ink-jet printers.

[Mike 125]

Re: Too much of the wrong sort of technology

>that's just a bit too much.

Yea but....

"Because they can leverage the scale of the cloud and don't need to build hardware farms, they can take advantage of all the benefits of modern development flows, things like continuous integration and streamlined ML DevOps and simplified security,"

Jeeeeeeeeez.

Ok, he's talking about the development process. But inevitably that cloud model will be forced on end users of the product, no doubt justified by 'securidddy'.

They've slowly accustomed us to endless updates for OSs. Why not for thermostats?

Brave new world.

[pimppetgaeghsr]

Re: Too much of the wrong sort of technology

They don't want to make it easier for you to change your thermostat pal, they want every invasive bit of data from you, your family and your home so they can see it, feed back all of it to their machine learning algorithms, then find a model for what will make you buy more products, change your political leanings (or predict them earlier), or something more sinister in the future once they realise the power that they have.

We humans like to think we are adopting free will and all the millenials I know buying houses, having kids have smart devices all over them (and have no explanation when I ask as to why they have them, other than to distiguish themselves form the neighbours next door in the same cookie cutter house, by buying more expensive home features or cars). This is going to be the biggest generation in 10 years and imagine being able to predict their vote years before they make it, then only recommending them certain TV shows, adverts or news stories months ahead of time in order to skew them one way or another, we already see it all over social media.

[Richard 12]

Trouble is, we can't buy the chips

These "top end" MCUs and application processors are all very well, but we all actually need lots of the little M4 class MCUs for the real IoT devices.

The thermostats, presence sensors etc, washing machines, bus-connected button units, touch screens, keyboards, mice....

They all have a baby MCU in them, and we just can't buy them.

[martinusher]

Re: Trouble is, we can't buy the chips

Those baby chips are common parts and, yes, you can buy them in sample quantities. Whether you'd want to or not is debatable -- the development, programming and debugging environments are easy enough to work with but working with these parts can be an acquired taste.

What we need are documented interfaces and a straightforward, robust security mechanism to protect the system. This should not include any Web protocols - POST and GET methods running on top of HTTP have no place in this type of system, they're a sign that the designers/programmers haven't much of a grasp of machine control and are an ongoing security headache.