Show HN: Industrial I/O solutions for developers, AI and robots | Hacker News Reader

Show HN: Industrial I/O solutions for developers, AI and robots

gpio.online

35 pointsgpioMaris18 comments

Great looking site, really catches your attention.

My input from a "currently evaluating different controllers" perspective is that the page somewhat lack information about interfaces, possibility for expansion (in our case we have requirements on interfaces not commonly existing out of the box).

I also believe that of your target audience is industrial solutions skip the "BS" (i.e. "Our controller allows you to develop voice controlled devices that are truly beneficial in the real world.") and instead go for specs. Even though your controller is not ready for the world i guess you have a goal or conceptual model?

Looking forward to see where you end up, I signed up for the newsletter :)

No information about the controllers (interfaces etc), no information about pricing.

I would be interested, but that site only leaves me shrugging my shoulders.

The single most important point of something like this would be: Update strategy. How do you do security updates? I feel like this should be literally point #1 on your hero page, because this is the biggest problem you have today if you do the "slap Linux/BSD on a SBC and put it on the internet" dance.

A quick look at shodan will illustrate why this is the #1 concern.

I agree that security is a most important point. That's why we choose to connect AWS, IBM and Google IOT servers and encrypt all traffic. We not going to implement our security features.

This is surprisingly lacking in most IoT equipment / deployments.

Hey, everyone! We are developing an industrial grade controller that allows developers to build cloud controllable machines to measure analog signals and control output pins. I am interested in what you think about this product. What do you think about this idea? What can I improve? Who do you think will buy this controller?

Thank you for the post, I couldn't find enough information on what would differentiate these devices from e.g. these https://catalog.azureiotsuite.com/ ?

What diffirentiates your product from the thousands of other similar products? Is it the ruggedness? The temperature range the hardware can operate in? Is it all open source code?

The IoT board/industrial computer market is pretty saturated. I work in this space, solving some problems professionally. It was tempting to turn our hardware/software stack into a product in addition to the actual problem we were solving, but after doing some market research, it was clear that without significant differentiation from products already on the market we would be Yet Another IoT Industrial Grade Controller.

Especially with AWS GreenGrass/IoT and similar efforts from Microsoft, Intel, Samsung, Dell and many others. There are already a ton of IoT/Industrial PCs on the market.

There is also Phoenix Contact, AdvancedTec, Siemens, ABB, Moxa, MPL...

Hey, nice to see some robotics stuff here.

First the page seems pretty vague on actual HW, which is important for potential customers, as that is what actually does the work for them (otherwise I did not understand the product), so you might give more information there, in particular the "Products" category is sparse, and the "Register" button does nothing.

Secondly, people hesitate (rightly so) to connect their robot (or IoT, etc.) to the cloud, as security in this area has a catastrophic record. Why not sell the platform/products as offline stuff, i.e. HW that includes the OS (e.g. ROS might a good start), etc. and control the stuff inside the LAN. Obviously the drawback is that you now do "on-premise", but I think in this area it doesn't make sense at the moment to go cloud, as too many questions about security do not have a stable solution, however it could be a specific selling point.

I see 'Industrial' in the description. Is security your number 1 and pervasive concern in the design of this thing? If not, no one serious will buy.

Even small shop industrial settings are pretty stringent about keeping automation networks air gapped from the rest, there is no way they would get a cloud plc unless you are willing to guarantee their security at cost and take on liability for damages.

> cloud controllable machines

Never mind security, what about latency and reliability?

I think there's actually quite a lot of market for this based on the number of people doing it adhoc with ESP2866/Raspberry Pi devices. But that would depend on price.

Hi. What makes it industrial grade out of interest? Under the hood are you using microcontrollers or..?

Could you give some more information on the controller.

Cheers

Controller inputs and outputs features rugged protection circuits to operate the in the harsh environment. All interfaces work up to 30VDC and has overvoltage, overcurrent protection. Output pins can control relays and stepmotors directly. Currently, we use a Cortex-M3 ethernet microcontroller, and running bare metal firmware in it. But we have a plan to migrate ARMv7 processor and run an embedded Linux, to improve security.

> But we have a plan to migrate ARMv7 processor and run an embedded Linux, to improve security.

What will the update strategy be?

Because it is meant to control machinery we provide two different update ways. Manually uploading firmware and second option when customers distribute from cloud and schedule restart procedure. Do automatic updates are not a option.

This is great stuff. Any GUI in the works?

One could probably connect that to our Flowhub IDE (https://flowhub.io), either by wrapping it with NoFlo, or implementing the protocol:

https://flowbased.github.io/fbp-protocol/

Flowhub is open source, but there is also a supported and hosted version

Reply on news.ycombinator.com