The internet of (no)things is what the Economist magazine calls the wild west of standards that is IoT (internet of things) technology presently.
A quick web search brings up the following (at least) 6 competing IoT standards.
- Intel/Dell/Samsung: IOC
- Microsoft/Cisco/LG/Panasonic: AllSeen Alliance
- GE/Cisco/Intel/IBM: Industrial Internet Consortium
- IBM: MQTT
- Apple: HomeKit
- Qualcomm: AllJoyn
- BT/Intel/ARM/IBM: Hypercat
Google haven’t quite shown their hand yet. With all this jockeying for position, venture capitalists would be wise to let the different groups duke it out and let a winner emerge before committing to a consumer IoT start-up. (If a winner ever emerges).
Consumer IoT needs a cheap, super low power, bluetooth equivalent, wireless protocol device with a range of a few KM with an integrated CPU that can run a stack as powerful as OPC UA on TCP. As of 2014, that’s still a pipe dream.
With all the talk of IoT companies taking over the Industrial Automation sphere, perhaps these companies could take a step back and instead look at how Industry has already solved some of these problems. Sure, Industrial Automation doesn’t have device auto discovery and geo-location standards (though RFID is in use for tracking mobile equipment). But maybe the industry figured out long ago that it didn’t need this tech and didn’t want any surprises on its device network. Every device is on the factory floor for a reason: it has a purpose that was defined during detailed engineering design. It didn’t just appear organically. Every device is polled for specific data for a reason. By contrast, in consumer IoT, all these devices chirp events like the programmer discovered the popular observer event software design pattern for the first time.
There are 2 types of data; useful and useless.
Most consumer IoT data is useless and just expensive to store. Do I really need my printer to tell my smartphone I’m running low on ink, or my smoke detector to tell me its battery is low? What’s wrong with beeping like it always did?! I don’t need another app for that.
In this regard, IoT consumer tech could perhaps take a short break from the IoT gold-rush and learn something from industry before they become a busted flush.
The key is to understand what is valuable to the business. That data appraisal task takes upfront engineering design and data modelling.
It takes effort to figure out what data you really need. The right tool can reduce the effort required.
While consumer IoT consortia are off reinventing the wheel, Industry has used OPC technology for decades to enable synchronous data communication between systems. OPC DA (Data Access) has its limitations, but most of these have now been addressed in the next generation OPC UA, which is poised for mass industrial automation acceptance. Sure, the OPC UA stack may be too heavy on CPU cycles for cheap consumer devices, but it has public key encryption and the payload can be tunneled and routed over the internet end to end.
Its all about interoperability & security, and OPC UA supplies this. The industrial automation industry and its clients need to get behind a single technology if we are to avoid a mess akin to that which is the burgeoning consumer IoT standard.
HAL Software Spike Prototype is an OPC UA-centric industry specific data modelling and prototyping package that provides a framework for data FEED (front end engineering design). It will be available for software demo download from February 2015.