I was making a routine 30-minute drive to raid a local vegetable stand. As I processed all the information coming in, I realized there were several dimensions to it. The same principle is at work with Big Data/IoT projects, no matter what their scale.
One of the classic charts/concepts widely used in information management is determining urgency vs. importance. Urgent & important things are represented in the upper-right quadrant of a square, urgent but not important in the lower right, important but not urgent in the upper left, and that phone call from your old college friend asking for money in the not urgent and not important lower left.
Riding Along in My Automobile
In my car, my speedometer sends me a reading that is urgent and important. My gas gauge reading is not urgent but important, and becomes urgent and important once I start running low. My gearshift is important but not urgent.
My car has a tachometer but also an automatic transmission, so the tach reading is not urgent and not important. I can go down through the checklist of everything from the odometer, water temp, battery, and oil pressure, to the heating/AC settings, the radio; to the stream of other info and alerts kept by the car's computer.
It's Not That Simple
But this simple classification misses a lot of important stuff. For one thing, as I note above, much of the information changes its nature depending on conditions. Not only am I in trouble if I run out of gas, I'm in big trouble if any of the water, battery, or oil gauges changes. They seldom do, but if they do, they become red-flag urgent and important.
So an IoT monitoring & analytical architecture must take into account the changing, sometimes fluid nature of the dataflows.
It must also take into account frequency, consequences, remedies, and cost. Returning to the gas gauge, I need to look at it infrequently, the consequence is embarrassing but usually not life threatening, the remedy is generally easy, and the cost is not high. The speedometer, on the other hand, must be monitored in real time, whether I ignore it or not.
Should the engine throw a belt and cut my power while driving at night, the consequence is more severe, the remedy more difficult, and the cost higher. Should I ignore a low oil pressure warning, then I'm potentially getting into a dangerous, difficult, very costly situation.
I've had air conditioning systems conk out on me, something for which there is neither a gauge nor an alert. This problem is not urgent or important (as uncomfortable as driving without AC in the heat can be), has little consequence (unless you need to drive people somewhere who may be less tolerant of the heat than you), but can be very difficult and expensive to fix.
I've also had wheels fall of my car--twice. The first time was when I was 19 years old, driving a decrepit used car, and speeding down the road in the manner of many teenagers who think they're going to live forever. I missed slamming into a bridge by inches, but ended up merely on the side of the road with nobody hurt. I also slowed my driving down forever.
The second time I was going 5 miles per hour, and thought I had a flat tire. Even at this slow speed, the incident could have had very bad consequences. Losing control of the car is a terrifying thing. This car was not particularly old, but had these supremely annoying aluminum wheels that raise all sorts of Cain with tires and linkages. Again, no gauge here to help me out, but both times I was put in a potentially fatal situation.
The "gaugeless IoT" of AC & wheel failures exists in many other aspects of driving my car. For example, even though a thermometer courteously tells me how hot or cold it is outside--something I had already nailed down as I was walking out to my car--it doesn't have an anemometer.
I have to know and gauge the effect of wind myself. It doesn't let me know about traffic jams ahead, flooding, or perhaps a tornado coming my way. It doesn't warn me of kids on bicycles or dogs or deer darting out in front of me, or big trucks suddenly crowding me. It doesn't predict any of the other thousands of examples of atrocious driving I've encountered with my fellow Americans.
But wait a minute, that's why I have eyes and ears. And memory, my own organic expert system.
So, So Early
This brings me to my overall point: Even as the IoT will generate trillions of dollars within the global economy very soon, we're still in the very early days. I'm tempted to say that we're only now in Act I of the Information Age. All previous progress has been mere prologue.
I'm speaking here of enterprise IoT applications: smart manufacturing, smart grids, smart cities, states, and nations. If personal IoT such as smart watches need to be corralled into a grid, then I'm speaking of them as well.
There's been a golden quest for decades to smarten computers up by attaching sensors to them. Sure, our wondrous machines have been able to crunch numbers better than us forever, and can perform all manner of brute-force tasks. Today they can even play a nice song or movie at a very high quality.
But the IoT will bring sensory to our systems in a big way for the first time. They will need to measure all the new incoming data in several dimensions, with built-in decision points that assess remedies and consequences. Even so, it will be decades before IoT applications can "see" and "hear" with anywhere near the ability of people. As always, let's not get too arrogant about the power of our machines, and remember that human beings are sometimes the weak link but sometimes the salvation of the systems we build.