Computing at the Edge: The Internet of Everything


Since the advent of the internet, technological advances – facilitated by ever-faster data-generation, computation, processing and intelligence – have enabled a new, sophisticated connectivity, seamlessly merging people, computers, machines, robots, sensors and devices through the Internet of ThingsOpens a new window . And as computation increasingly shifts to the edge, the Internet of EverythingOpens a new window will encompass it all.

The Fourth Industrial Revolution

Following the first, second and third industrial revolutions of steam, electricity, production lines and automation, we are now entering a fourth, digitally-powered revolution that not only extends and deepens the automation of factories, but also involves a proliferation of “cyber-physical” systems in the industrial environment.

A cyber-physical system is a mechanism controlled by computer algorithms that gathers inputs from the environment and conducts analytical processing to determine a course of action. The system then feeds instructions back to the mechanism, which incorporates and uses the information to interact with the environment. Such cyber-physical systems will have a steadily increasing impact as machines and devices become even more embedded in our everyday lives.

Already computers, robots, sensors and peripheral devices are controlling progressively more of our physical environment, continuously capturing data, creating digital images, monitoring and interacting with all aspects of the world around them. Machines and other devices are increasingly expected to understand and respond instantly and appropriately to events.

In short: they must exhibit intelligence.

Far from being “revolutionary,” these systems have underpinned technologies in many industries for a considerable period. However, now that they are being connected through the Internet of Things, they can also interact with other machines and humans, in real time, both within businesses and externally, and crossing organizational and national boundaries to connect globally.

From the IoT to the IoE

This explosion in the number of devices and their interactions via the internet has created new technological capabilities that are fast gaining traction as business-value creators. According to Accenture’s 2018 trends report, IoT technology will add $14.2 trillion to the world economy by 2030.

The IoT consists of a network of physical objects connected via the internet that communicate with their processing centers and other objects on the network, sending status updates and accepting instructions to conduct specific actions within their immediate environment.

This machine-to-machine connectivity enables businesses to streamline and optimize operations such as preventive maintenance, where systems detect anomalies through measurement of their own performance, and create alerts or fix problems before they occur – saving businesses from breakdowns, delays and damage. With a huge body of IoT cases documenting return-on-investment gains, more businesses are seeking to utilize this technology.

But the IoT is just the beginning. According to networking giant Cisco, IoT devices are only part of a much larger phenomenon: the Internet of Everything – which brings together people, process, data and objects, expanding the applications of IoT and making networked connections more valuable by turning information into actions. These create new capabilities and opportunities at the level of businesses, individuals and countries.

The IoE has four pillars – people, process, data and things (as in the IoT) – that together form an intelligent, interconnected environment, allowing each component to communicate and interact over public and private networks. The IoT is a subset of the IoE consisting of machine-to-machine connections, but the IoE is the technology, philosophy and environment encompassing everything that can connect to a network, together with the value created by the information, products, services, capabilities and experiences such total connectivity enables.

So, if the IoT alone can add $14.2 trillion to the global economy by 2030, how much can the IoE add?

From the Center to the Edge

Because connectivity resources are not infinite, the connectivity of everything can also threaten our very ability to connect. As more and more elements of our networks become “smart” – generating and harvesting data in huge quantities, continuously and at high speed – the transmission of data to and from remote data centers can slow up or stop as traffic congestion creates network bottlenecks.

In this scenario, bandwidth depletion and latency are major issues, as is the time to transmit data from the point of generation at the edge of a network (e.g. from a sensor) to the center, to analyze it to determine the actions required, and transmit these decisions back to the data’s source. If data collected locally, at the network’s edge, is needed to drive local decisions urgently, this does not allow time for the data to travel to and from the remote processing center.

Enter edge computingOpens a new window : a type of infrastructure architecture where small-scale data center facilities sit at the edge of the network where the data are gathered. Often referred to as “the fog,” these data-processing resources enable analytics to be performed locally, speeding up intervention where needed and reducing network data traffic.

Self-Healing Environment

For example, a patient at risk of a seizure has his brain activity monitored by an implanted neuro-stimulation device. The IoT device looks out for unusual brain-wave patterns, conducts analytics locally, and within milliseconds of detecting a data pattern suggesting the onset of a seizure, delivers electric pulses designed to stop it – without connecting to any external system. If the device needed to consult “the center” before administering the intervention, it could affect the patient’s outcome adversely. Computing at the edge is the only viable scenario in such a case.

Another example: a semi-autonomous tractor is used to optimize planting decisions based on soil conditions determined in real time by the local processing unit as the machine travels through the field, assessing the best seeding options. The tractor’s computational autonomy operating at the edge enables it to act intelligently, making smart, data-based decisions that improve crop yields.

Combining the sensors monitoring the performance and environment of a system with a local analytical algorithm that understands what normal operational parameters look like, and can intervene when abnormal patterns are detected, creates an intelligent, self-healing environment that does not need control from the center. In both of these cases, computing at the edge delivers local system intelligence quickly where breakdowns could have serious, expensive or catastrophic results.

Computing at the edge can enable businesses to design smarter environments to operate in,Opens a new window creating a winning combination of real-time sensing of environmental factors and systems’ ability to act locally in an autonomous fashion, thus delivering intelligence at exactly the right time and place for businesses and their customers.

Between the Cloud and the Fog

As the fourth industrial revolution advances – the industrial Internet of Things – businesses will need to adopt strategies that utilize both central processing in the cloud as well as in the fog at the network’s edge to realise the IoE’s potential. They must navigate the new reality of extreme connectivity, transforming their technological and business models to stay relevant in a world that demands local, actionable real-time intelligence.

Businesses will need to extend their computing facilities beyond processing centers to the edge of their networks, exploring customized hardware solutions to combat latency and connectivity limitations. Non-time-critical, high-end analytical processing will remain the preserve of the IoT cloud, but the real-time processing driving local interventions will need to occur closer to the triggering events, otherwise the delay will destroy its value.

Each business will need to find its own mix of cloud and edge processingOpens a new window to determine where its most valuable interactions occur on the IoT, and how best to deliver business-critical decisions there. They require new processes, roles, strategies and operating models to connect to the IoE in order to ensure they remain competitive.