Page 26 - EngineerIT June 2022

ICT – INTERNET OF THINGS

How edge computing will boost

the expansion of IoT services

By Morné de Villiers, Integration Architect and Project Manager at TechSoft International

ith the rapid expansion of digital solutions to all sectors, some confusion has
crept in regarding edge computing and the IoT device. It’s an easy mistake to
Wmake, as the devices themselves sit on the edge, and when those in the know
start talking about adding intelligence to the edge, the waters get even murkier.

Spot the difference
To truly understand how edge computing will boost IoT devices’ use and prevalence, we
need to look at the actual difference between the two. The main difference between
true edge computing and IoT devices is based on compute and storage capabilities. IoT
devices are internet-enabled, allowing communication to be established with the device
remotely.
Although communication is provided, the IoT device is typically limited in its ability
to process large volumes of data and perform complex analytics and data processing by
itself. This is where edge computing comes into the picture. Morné de Villiers
Edge computing allows the architecture of the IoT solution to put processing power
closer to the source and allows the heavy lifting to be performed closer to the IoT devices
themselves. With more processing capabilities closer to the IoT device that is connected application of the user and whether edge
to the edge, near real-time AI and analytics become a reality. computing is required, but IoT devices
For example, in an industrial setting, let’s say a factory, there could be several IoT- won’t be able to do away with the edge
enabled devices throughout the factory floor connected to an edge computer, through entirely.
various communication protocols such as BLE, Wi-Fi, LAN, etc. In this model, the IoT
enabled devices can provide telemetry data to the edge but won’t have any knowledge of The edge supporting IoT app
the other devices around them. This is where the edge computing infrastructure comes growth
into play as it collates data from all the devices, complex computations, decision making, Remember we discussed the distance
analytics, and data aggregation and then pushes this data to a central cloud back-end for making a difference here. The big game
processing. changer for IoT devices and services
In the case of the intelligent edge, the edge computing device could receive doesn’t just depend on edge
the information from the IoT device; note that there is an anomaly that it computing; it is the impending
understands as an alert, and will send an instruction back to the IoT device The big introduction and
immediately. The time in which the IoT device receives its instruction is game changer for implementation of 5G
significantly reduced, as the distance between the device and the edge IoT devices and services networks. In the past,
is vastly less than between the device and the cloud. large volumes of data
doesn’t just depend on could be processed
Adding intelligence to the device edge computing; it is the by the edge and
While many believe that placing intelligence on the IoT device itself impending introduction and connected to the IoT
will do away with the need for the “edge”, I would have to disagree. implementation of 5G devices via numerous
The value edge computing delivers goes beyond just supplying currently networks. communication protocols.
“dumb” devices with instructions. It adds the additional compute and In remote computing
storage capabilities these devices themselves don’t have. environments where there is
A device on the edge that creates the edge computing network could be as small no fibre for example, the data needs to
as a micro-processor; think Raspberry PI or Arduino. Similarly, in the case of industrial- be transmitted over traditional mobile
strength devices, these could be the size of a lunchbox. Therefore, if you want to connect networks such as GSM, GPRS, 2G, 3G and
smaller electronics into an environment where size capabilities are limited, you will find 4G to a cloud-based processing back-
that their processing capabilities are also limited, which is where the power of edge end. The problem here is that latency
computing becomes important. limits the back-end system’s ability
Again, will it do away with the edge? This will always be based on the use case and to propagate near real time decision-

EngineerIT | June 2022 | 24