$10-20 Industrial Compute Module

The Raspberry Pi compute module has not been updated in a long time. What will really accelerate industrial IoT would be a $10 to $20 industrial ready compute module. Ideally the module is a combination of an ARMv7 or v8, minimum 512mb memory, minimum 4GB eMMC on the board, support for Ethernet, WiFi, Bluetooth via a breakout IO board. But the key feature would be to also house a micro-controller. The micro-controller can be used for all real-time zero latency tasks related to machine control, robotics, drone control, and so on. Via an Arduino-like bridge, snaps on the Linux partition would then be able to define the life cycle management of the micro-controller logic but would not run the actual real-time code. This would be best of both world: real-time zero latency guarantee with flexibility, speed of deployment, security, industrial app stores, etc.

$20-50 Industrial Breakout IO Boards

To accompany the industrial compute module, you need breakout IO boards. High on the list are PLC (read ALC) breakout IO boards that only would need an industrial housing to be put around them. Just put snaps on the Linux side of the board and securely control the micro-controller side via snaps and you have an ALC. A concrete example could be an industrial actuator pump in which the real-time control is handled by the micro-controller but the snap chooses the programme to load. The Linux side would also be able to run as snaps: monitoring solutions, analytics, autonomous machine learning solutions, control system integrations, all type of industrial standards for remote control, cloud integrations and much more. Other markets for IO boards would be for specific appliances, e.g. vending machines, home appliances, drones, robots, etc.

$30-75 Industrial FPGA-Enabled Compute Module

The industrial micro-controller + Linux board should also have a more powerful version. High-end ARM 64 bit or Intel combined with an FPGA instead of a micro-controller. This version would be ideal for very complex processing at the edge. Think about industrial drones, autonomous robots, real-time image recognition for security cameras, digital signage and much more. Via snaps you should be able to reprogramme the FPGA so each customer has their own solution.

$30 ARMv7/8 broadband modem

Lots of broadband modems are MIPS and have 32MB of memory and 8MB of storage. You can't do any really interesting stuff on these specifications. What is needed is more powerful ARM v7/8, 512mb memory, >2GB eMMC, 5 x Gigaethernet, WiFi AC and either ADSL, DOCSIS Cable or Fiber modem. Getting these boards close to the $30 mark would allow telecom operators to substitute their current broadband modems and generate new revenue via app stores while each customer personalises their broadband experience.

$75 or less ARMv7/8 network equipment with FPGA

The lower end of the network equipment market is being dominated by ASICs and MIPS. What if in 2017 we could see FPGAs and ARMv7/8 instead? From a software perspective software defined networking and virtualisation of network functions would become real. Performance could be similar but flexibility and revenue generation would be a lot higher while cost of ownership a lot lower.

$99 software defined radio

Let's make sure the maker community can create new wireless innovations and sell them to lots of others. Having a $99 software defined radio that can both receive and transmit enables lots of new innovations in drones/robotics control, low-powered long range, ham radio, WiFi, low-bandwidth LTE, bluetooth, garage door control, HVAC control and most importantly new protocols.

$150 3D printer

The affordable 3D printer with apps permits manufacturers to have a Gillette model in which replenishments, 3D models and so on can be automatically bought from an app store. An app can rent your printer to your neighbours.

$500 or less 4K screens with plug-in compute modules and sensors

Digital signage wouldn't need all the items a regular television needs. Ideally it is a bare bone 4k screen model and you can just plug-in a compute module of your choice. Afterwards you plugin lots of other sensors, e.g. camera, microphone, temperature, software defined radio, etc. Apps create new revenue streams on top of just showing advertisement. The software defined radio converts digital signage into a mobile base station inside cities.

$5000 and less for high-end Intel mobile base stations

Putting up thousands more would be possible if the cost goes down substantially. Connecting the unconnected. Via app-enablement of the base stations, new services can be launched that generate new revenues for operators. Making deployments economically viable in rural regions. Think about combining base stations with content delivery networks and streaming local content as one example. Costs of management can be dramatically reduced if local communities would be able to do basic maintenance and apps can guide them.

This article was originally posted on LinkedIn.