May 19, 2015

Industrial Internet of Things

On behalf of my employer, Owens-Illinois, I attended a conference titled The Automation Conference in Chicago, IL. While automation is a broad topic, this conference seemed centered on the Industrial Internet of Things, or IIoT. I must admit, this is what peaked my interest as I've been playing with the consumer side of IoT for a while with products like Spark Core (Particle now), ESP8266 boards, various sensors, etc. My largest curiosity was how this ideology applies in an industrial sense and what types of problems exist in the industry that this technology could solve.

As with most industry, at least my experiences with my employer and what I've gathered from talking with people, manufacturers are remarkably slow at adopting new technologies into their workflow. This is due in large part to the associated risk as well as the capital involved in pursuing a venture that might ultimately fail once, if not several time before it's done right.

The biggest advantage IIoT presents is not adding more data points for engineers or end-users to analyze, but to enable a system of connected devices to store data, analyze it, and ultimately make decisions based on it. All of this with little-to-no human interaction. The automobile analogy is brought up quite a bit to better explain this. Imagine an ecosystem of cars that are self-driving and able to communicate with each other. Routes could be determined based on any number of factors whether they be real-time congestion, weather, construction, etc. While en-route, a series of events could render the current path less optimal than another and the car would automatically re-route based on that data that it has gathered. You can imagine how this technology could be adapted to a manufacturing setting and have a tremendous impact.

From what I've gathered and come up with, the current roadblocks that exist to implementing the IIoT methodology is as follows:

  • Leadership buy-in
  • Communications protocol unity
  • Accept that there will be failures

In order for a paradigm shift to occur in a company, the leadership must approve of it and be on-board. If this doesn't take place, there is no hope for the idea to come to fruition, regardless of how much sense it might make to an engineer. Sometimes this buy-in is driven internally by engineers and other thought-leaders, and sometimes it's driven by customer demand.

No factory or manufacturing establishment seems to have the same concoction of control systems. This means that everything is talking a different 'language' and nothing is able to easily talk with it's neighbor. Clunky protocol converters exist, but add failure points and are an added cost. If a service or platform could sit above all the hardware in a plant, do the conversions and aggregate the data to make it more easy to act upon, this problem would be solved.

Being a technology that is just now breaking into this space, anyone willing to try it out first is bound to have a few failures. That just comes with the Research & Development space. Not all companies have the capital or the manpower to invest in something that is not a sure-fire win. This fact alone knocks the majority of manufacturers out of the game until it stabilizes and becomes more robust.

Suffice it to say, the speakers have been correct in saying that it's not a matter of if the IIoT will happen, but when. Being the trillion dollar business that it is or will become, I am excited to take this knowledge and continue to try and identify new gaps in the technology so that I may be lucky enough to have the skills to develop a solution for it.