The Missing Part Of The Internet Of Things Conversation

The Internet of Things (IoT) promises all new ways to use sensor and machine data to create efficiency, new features and foster business model competition. But the conversations too often lack an essential element that no IoT project can do without: the ability to make systems “think” in ways similar to what humans provide in the analogue world. Consider not just the rapidly increasingly amount and types of data that are created and derived automatically, but also the exploding number of ways that data needs to be filtered, mashed up, compared, contrasted, interpolated and extrapolated in the IoT scenario.

If digital information caused humans suffer what Clay Shirky calls “filter failure,” imagine what happens when human-engineered machines are asked to manage exponentially higher amounts of data in significantly shorter time frames, both with and without the possibility of human oversight.

Perfect memory and absolute patience

This isn’t a warning about the weaknesses of machines, but is instead about their powers of tireless, perfect memory and absolute patience. The benefits that computerization brings can quickly become a rabbit hole if we don’t mitigate the weakness – the lack of human-equivalent reasoning – that comes along with those benefits. The way clever companies have been addressing this gap is through the use of multiple types of event processing technology.

This area of tech is based on the construct that everything that can be observed and understood is broken down into discreet elements known as events. Events can be as simple as a change in data value, like a rise in temperature, or something that doesn’t actually happen, like a patient failing to pick up a prescribed medication. Events in combination form higher-level events like financial transactions that consist of an offer, match and settlement. Events are temporal as well, meaning their occurrence can be monitored for timeline and frequency.

Blending historical and immediate

Event processing is at its most powerful when analytics are used to understand historical patterns that can be added to the mix of what’s being observed and anticipated. Marketing has been moving forward rapidly on this front as historical buying patterns are teased out of large data sets and then used to anticipate future consumer engagement success.

Event processing is fascinating because it offers a way to describe and understand the environment in a very brain-like way. It also allows a way to spot patterns that represent risks and opportunities no differently than a human does, and to see complex patterns in ways that are increasingly similar to our brain.  This is a critical capability for event processing to fill as IoT-led automation by necessity takes humans out of many points of decision making and exposes new risks unless we very carefully study the changes automation brings.