Johannesburg, 19 Aug 2021
Edge computing is a rapidly growing industry driven by a leaning towards a decentralised form of compute. Both Forrester and Gartner predicted that edge computing would become mainstream this year, taking advantage of 5G, the internet of things (IOT) and cloud native software.
“Edge computing refers to what it is - a location outside of the core where the data needs to be processed,” explains Jim Holland, Lenovo’s Regional Director for Africa, “moving data production to the edge, closer to the end-customer and end-user.”
From agriculture to mining, local governance, security, healthcare and retail, Holland is seeing a lot of interest in edge computing in South Africa. Both 5G and the availability of data - something Holland refers to as the ‘ramp up of exponential data at the edge’ - are driving edge computing discussions. In addition to this, artificial intelligence (AI) has a role to play in understanding how quickly we can process data.
“A lot of those industries are looking at the edge… we’ve had a few successes on a smaller scale but more opportunities are opening up all the time. The benefits of edge are very clear,” says Holland. “There are no disadvantages if used in the right environment. Edge computing has to be part of an end-to-end hybrid approach - it’s not a solution that works in isolation.”
In the era of cloud computing, edge is far more than geography. There are a number of factors driving edge computing. Governance, for example, has a role to play when it comes to privacy laws - the where and how of storing and processing data in the cloud isn’t that simple.
“The first one to mention is latency. The closer compute happens to the edge, the less time you’ll spend waiting for responses,” explains Holland. “The second thing is economics - the cost of moving data to and from the edge of the core. Data volumes can be large so from an edge perspective, you may only choose to move key data instead of old data to optimise your costs.”
Being independent from the core is clearly beneficial - in the case of communication link failure, edge computing provides redundancy in terms of additional edge sensor devices that don't necessarily rely on a central infrastructure to operate. While core data centres do not generally experience a lot of downtime, edge computing could eliminate that risk.
While edge computing is not new, the idea of a decentralised network of endpoint devices placed at the edge of computing networks has created a lot of misconceptions, especially around security and encryption.
“One misconception is that devices are insecure and that isn’t the case,” adds Holland. “Edge is not a cloud replacement, it’s not all about endpoints, sensors and IOT. It's another tier of distributed computing architecture where the compute analysis of data is required at the point of collection.”
Holland brings up the example of SANParks. Game reserves are rugged environments - dusty and faced with continuous power issues. Here, edge computing would be a good fit in terms of processing data, easy-to-manage rugged devices and getting information to guests entering the reserve at the endpoint. Today’s edge devices are designed with the understanding that the edge is not a sterile environment.
“There’s no cooling requirement so if the aircons go down, these devices can operate up to 75 degrees. They’re not insecure because if they get tampered with, they get bricked automatically. They’ve also got dust filters - a lot technology today is catering for the edge,” says Holland.
According to Holland, edge computing relies on simply moving the infrastructure to the point of collection as opposed to keeping the infrastructure at the core and waiting for the data collection. “The edge doesn’t have to be a data centre. It also doesn’t limit you in terms of connectivity. The perception is that edge is quite limiting but in fact, it’s probably broader. When WiFi connections are not available, 5G enables edge computing to operate at higher speeds, reducing latency, making it more of a viable option.”
Increasingly, more corporate data is being created outside of data centres than internally. From telehealth to mining and data-driven sports like Formula One racing, edge computing is often required to provide compute and storage at these locations, and to control and manage devices.
“Getting the data from the edge in real-time is what we’re all aiming towards,” he says. “Communication information technology is really maturing to support data centres at the edge. Time to respond is critical in today’s world and edge computing is really a key enabler.”
Share