Monday, October 12, 2015

Internet of Thinking: Computing Races to the Edge


I have written previously about the five core requirements that must be addressed for any IoT application to work in reality.

By the way, most people call IoT the Internet of Things, but I call it the Internet of Thinking, because it opens up tremendous new possibilities for human thought.

As I talk to people around the industry about how companies are seeking to IoT-enable their businesses, these five topics stand firm as the recipe for enabling remote collection and activation of otherwise “dumb” devices.

When you closely examine the interdependencies of these elements, you realize that they often invert conventional thinking and logic around managing complex systems.

Most organizations have evolved over time with the fundamental understanding that centralization drives higher levels of control and predictability. The notion of fewer variables led to the conclusion that risks could be managed more effectively, and efficiently.

In the world of technology-enabled business processes, that’s what drove the creation of centralized “data centers” to house the computers and the storage arrays that processed data and produced essential information and services. It’s also what drove organizational design that placed accountability for performance of processes into the hands of designated experts.

Simplistically, the model was to leverage the precious computing power, and talent, that was organized and managed centrally.

In an IoT world we can anticipate a massive explosion in the volume of information collected and/or generated by myriad devices. The act of transporting that information to a point of central processing will be arduous and cost-prohibitive. In fact, it’s likely that not all the information being generated needs to be processed; some of it will be noted and discarded.

This will give rise to some exciting new information models and services that enable the IoT model of operation. 

Imagine that a device (i.e. a refrigerator, dog house, or fence camera) not only has the smarts to detect, recognize, and collect information, but also has the computing capacity to process that information. And, make decisions or take actions as a result. This is known as edge computing, where the first-line processing of the information occurs in a distributed manner. No longer will the information be transported to some central computational utility; it will be analyzed and evaluated by the edge devices themselves.  With resulting actions potentially initiated autonomously.

This capability will rely on a material reduction in the costs, power requirements, and physical profile of the edge devices. If Rover is wearing an IoT collar, it will need to be small, lightweight, and not reliant on power cords. 

Venture funding is already supporting such innovations. According to a recent report from research firm International Data Corp., the market for IoT devices and services will nearly triple to $1.7 trillion by 2020.  Plus, a few weeks ago Freescale Semiconductor announced the world’s smallest single chip module for IoT applications. Follow the money.

Taken together with the shift that many companies are making towards off-premise cloud-based computing and storage services, the IoT rush to the edge has the potential to turn traditional data center and IT service management models upside–down.


Peter Allen has many years of operating experience as a top executive and strategic advisor for companies of all shapes and sizes, with focus on technology-enabled business services. He is now a Boston-based Managing Director at Alvarez & Marsal.


Image: Andreas-photography/Flickr








No comments:

Post a Comment