The end of this decade is drawing nearer. Rather than simply look ahead to the rest of 2016, let’s get more ambitious and focus on 2020.
The uber-trends for the next several years will be the deepening impact of IoT on all facets of our lives and the embedding of machine learning into every niche of that ecosystem. We’re already seeing that trend pick up speed in the intelligent products unveiled at the recent Consumer Electronics Show.
At the heart of this revolution is the smartphone. Every day we see more innovations that use the smartphone as a combination universal remote control, secure access key, and personalized dashboard that works across myriad IoT-enabled applications. Analytics of various sorts -- both on the smartphone and in the cloud -- are a pervasive feature of most of this new wave of consumer IoT applications.
The embedding of cognitive computing technology directly into smartphones will accelerate in the next several years. This trend will reduce the need for round trips between the devices and the cloud data centers they’ve traditionally relied on for intelligent services such as personalized advisers (such as Siri). These embedded algorithms will drill at memory speed into the growing streams of sensor data being captured and cached locally on the smartphone themselves.
As the price of flash storage continues to drop and their density and reliability improve, we’ll see embedded analytics secure a footprint on every smart device, material, or other artifact that comes to market. A new generation of data scientists will build and optimize these analytics to drive intelligence into every product we use.
By 2020, new products in every sector of the economy will have been rearchitected as “cognitive IoT” endpoints. As such, they’ll embed local sensors, actuators, smart materials, and algorithms enabling them to adapt continuously to their environments. And all of them will handle most cognitive IoT processing locally, more rapidly and flexibly than any cloud service.
From what I see, the next generation of app developers and smartphone designers are moving in this direction. For example, I recently came across a headline that would have been inconceivable only a few years ago: "How to Exploit All Sensors of Your Mobile Phone as an IoT Device." In the piece, author Abdellatif Bouchama calls out the range of sensors on a typical Android phone. These include an accelerometer, linear acceleration sensor, gravity sensor, gyroscope, light sensor, and orientation sensor, though he could have easily included the ever-present camera (still and video) and microphone. Not only that, he discussed open source cloud-based tools for acquiring, storing, analyzing, and visualizing all that sensor data.
As the material world shifts toward endpoint-embedded cognitive IoT, cloud services will remain important. After all, cloud-based storage, processing, and analysis will still be essential for many applications that require massively parallel processing and storage capabilities in the petabytes and beyond. But true real-time responsiveness and agility at the device level will require the local processing and native intelligence of endpoint-embedded cognitive IoT.
As 2020 rolls around, cognitive IoT will be so embedded in everyday reality that life without it will become increasingly unthinkable.