Edge coming to the rescue of cloud
Cloud computing is at risk of being throttled by its own success, overwhelmed by the rising tsunami of data. However, edge computing is riding to the rescue, particularly in regions like Africa, which are most vulnerable to cloud’s three major weaknesses: bandwidth limitations, excess latency and network congestion.
That’s the view of Pramod Venkatesh, Group CTO at Inq. While acknowledging that the concept of edge computing isn’t new – its roots go back to the earliest days of remote or distributed computing – he maintains that edge computing is the next evolution of cloud computing.
Cloud computing itself is an evolution of traditional enterprise, client-server computing where data is moved from a user’s computer across a WAN – or the internet – to a centralised computer, where it is either stored or worked on, and the results sent back to the user.
“With the rise of 5G networks, more companies than ever can harness comprehensive data analysis without the IT infrastructure needed in previous generations. That’s the power of the cloud,” Venkatesh says.
But the quantity of data moving across the internet is enormous and getting more so by the minute. The World Economic Forum estimated that at the start of 2020, there was about 44 zettabytes of data in the world – or 40 times more data in the digital realm than observable stars in the universe. By 2025, that unfathomable number would be added to by another 463 exabytes of newly generated data produced by an ever-increasing number of connected devices, every single day.
Gartner has predicted that by 2025, three-quarters of all enterprise-generated data will be created outside centralised data centres on or by these devices. The internet would buckle under the load.
Traditional cloud platforms, including those set up and operated by the world’s largest providers, which are struggling to cope now, could be overwhelmed. The impact on time- and disruption-sensitive data could be catastrophic. It doesn’t take much imagination to appreciate the chaos that would result if data sent from a self-drive car for analysis at some distant data centre were delayed, disrupted or distorted – by the time the confirmation came back that the car was approaching a hazard, lives could be lost.
The further the data centre – where the analysis is to happen, from the end-point, where the analysis is needed – the greater the risk of delay. And in Africa, those distances are not only large, they’re also – with the exception of South Africa – across borders.
“From a regulatory perspective, cross-border data transfer can be problematic in some African countries, requiring certain types of data to be processed in-country,” Venkatesh says.
“Another major issue with traditional cloud computing in Africa is the cost of connectivity, which is still significantly higher than in the rest of the world. Add to that the fact that the locations from which data is being generated may be in hostile environments – down a mine, for example – with limited or intermittent connectivity.”
According to Venkatesh, edge computing effectively addresses all these issues as the data is processed as close as possible to where it is generated, even possibly on the device that collects or generates the data in the first place.
“The beauty of edge computing is that it has endless potential applications, particularly when those applications require some form of AI. This can range from security and medical monitoring to self-driving vehicles, video conferencing and enhanced customer experiences,” he says.
“Many users today are not even aware that they are using some form of edge computing. For example, it’s already widely used in entertainment and gaming: streaming music and video platforms often cache information to reduce latency, thus offering more network flexibility when it comes to user traffic demands,” he says.
An ever-increasing number of devices are needed to communicate and process data in a localised environment, such as devices like voice assistants. Without the help of decentralised processing power, devices like Amazon Alexa and Google Assistant would take far more time to find requested answers for users.
Manufacturers use edge computing to keep a closer eye on their operations. Edge computing enables companies to monitor equipment and production lines for efficiency and even detect failures before they happen, helping to avoid costly downtime delays.
Edge computing is even being used in a mine in Zambia to detect dangerous snakes and warn miners of their location in real-time. Edge computing is also being used to detect cars and goods arriving and leaving company premises, thus preventing unauthorised use as well as theft.
“As all or most of the computing work is done at the edge, only data that requires deeper analysis, review or other human interaction need be sent back to the main data centre. The amount of data to be sent is thus vastly reduced, requiring less bandwidth or connectivity time than would otherwise be the case. Edge computing is thus reshaping IT and business computing,” he adds.
However, Venkatesh warns that edge computing comes with challenges of its own, not least of which is security, both physical and cyber, as well as the management and control of edge devices.
Nevertheless, he believes we are only just beginning to scratch the surface of edge computing’s potential and predicts that the uptake of edge computing, particularly in Africa, will continue to accelerate over the next decade.