Trent Odgers, Cloud and Hosting Manager at Veeam Software, explains: "True edge computing, which combines the intelligence of AI and the cloud, is still being developed locally. The use of edge to overcome connectivity challenges is gaining in popularity."
Odgers cites an example of true edge computing that was presented at a conference he attended recently, where it was used to process data from Formula One cars so that real-time decisions could be made. "If that data had to be sent to the cloud before being processed, even a few seconds' delay could have cost the team several laps before they could make the required adjustments to the vehicle. By using the edge, decisions could be made on the fly around the car's performance and team strategy."
The ideal is to have the data processing power as close to where you need it as it possible to avoid unnecessary delay while the data flows to and from the cloud, explains Odgers. The use cases for edge are limitless. You get the answers that you need when you need them. The ability to instantly send and receive data is critical.
However, as exciting as the instantaneous nature of edge is, in South Africa, edge is particularly relevant where connectivity is an issue. Odgers says: "When you consider the farming sector, for example, one of their biggest challenges is connectivity on their farms. How do they get their data to the cloud and back when they live in the middle of nowhere? In this space, edge is very relevant as it enables them to analyse and react to data where they are, then send a copy of their data to a cloud platform as an offsite copy, or perform more in-depth analysis of the data on a bigger platform. But that application is for data that doesn't require instant reaction; it's more for proactive planning purposes."
This type of application is also common in the mining sector, where connectivity can be a challenge, yet the ability to respond quickly to data is key.
However, edge has other applications, such as helping remote or branch offices stay connected. He says: "Imagine a dental operation that's part of a large national group, but based in an outlying region where connectivity is a challenge. We're talking about a very small operation with very little data collection outside of point of sale. This type of hub and spoke architecture would work with a centralised data centre for 80% to 90% of the company's production workloads, but with severs at branch level servicing several types of applications, such as file and print and certain network-dependent applications.
"A scheduled backup ships all of the remote site's accumulated data to the cloud and a central location. This enables businesses to service areas where connectivity is more of an issue, while still bringing everything back to one point of control so if there's a data loss or failure, there's an offsite copy of that data."
Currently, Odgers says the focus for all businesses is to ensure they are relevant in this new world. "While things are moving more to the cloud with platform as a service and doing away with operating systems, at the same time we're also moving away from the cloud with edge computing, where everything needs to be self-contained and secure."
Odgers talks about the relevant edge. "End-users in outlying areas have data locally on their physical devices, while at the same time are able to use Office 365. Edge is also about making data available close to the network for instantaneous analysis. South Africa is still getting there and building up those platforms."
While there are instances of people deploying their own, low-cost IOT devices, such as farmers placing sensors on water troughs, for example, the overall drive is towards having onsite access to localised data that can ultimately be centralised for the greater good.
Finally, some advice from Odgers for companies looking towards the edge: "We see companies readily moving their workloads to, from and within the cloud. It's important to find the right fit for each specific workload before they migrate in any direction. Companies also need to consider the total cost of ownership to get in, and out, of the cloud. Another important thing to consider is whether the operating system that's generating the data is supported as well as how well protected that data is."
Read more about edge computing by downloading a white paper on the topic here.