How edge computing completes the cloud equation
Now Reading
How edge computing completes the cloud equation

How edge computing completes the cloud equation

Cloud data centres are only slightly more distributed than on-premises data centres


As internet of things (IoT) devices proliferate and incorporate more processing power, vast amounts of data are being generated at the edge of computer networks. IDC predicts that by 2025, there will be 55.7 billion connected devices worldwide, 75 per cent of which will be connected to an IoT platform.

Traditionally, the data produced by IoT devices was relayed back to the cloud, processed and further instructions sent back to edge devices. This setup is however unproductive as it creates inefficiencies with speed and latency.

Edge computing filters and processes data closer to the source, sending only relevant data to the cloud. This minimises bandwidth and cloud storage costs associated with data derived from IoT devices, observes Walid Yehia, senior director, Presales for MERAT, Dell Technologies.

Additionally, as many industrial applications for IoT also require critical real-time sensor responses, network disruptions cannot be risked. This applies particularly in remote locations, where network connectivity is not always available. “As edge computing capabilities are becoming a critical component of IoT platforms, they are making a stronger case for deployments. IoT is spreading across many industries and generating a lot of data from connected devices, with the presence of edge in IoT starting to create new opportunities and business value with reduced costs and real-time decision making,” says Yehia.

While supporting ubiquitous access, cloud data centres are only slightly more distributed than on-premises data centres. By contrast, the edge enables organisations to deliver applications closer to users.

“In many ways, the edge is just the next step outward in an expanding universe of distributed applications, with benefits – and drawbacks – aligned with those of multi-cloud strategies,” says Lori MacVittie, principal technical evangelist, Office of the CTO at F5.

“Data analytics represents a key edge computing use case, enabling the insights required for digital transformation initiatives,” MacVittie adds.

AI has traditionally resided in data centres, where there’s sufficient compute power to perform processor-demanding cognitive tasks. This works fine when immediacy is not paramount – the issue is that more and more applications require instant or near-instant reactions to the information they are delivering.

“Moving that front-end information-gathering part of the app to the edge, and then applying AI intelligence at the same point, allows AI systems to use inference (how AI uses observation and background to reach a logical conclusion) for faster decision-making,” says Joe Baguley, vice president and CTO, VMware EMEA.

Widespread 5G rollout and edge computing deployments will go hand in hand, as they both drive and benefit each other. With almost 10 times the speed of 4G, 5G is set to unlock numerous potentials in many industries. “With its ability and bandwidth to support billions of connected devices, new applications for sensors and connected devices will emerge, raising the demand for edge devices that can process, analyse and transmit data in real-time,” says Yehia.

Likewise, edge computing is essential for helping 5G reach its full potential by solving the latency problem. “Quick network performance is a necessity for 5G when connecting numerous devices, especially where AI applications are present, such as in smart cities or for autonomous vehicles that require feedback in milliseconds,” he observes.

Gartner predicts that around 75 per cent of enterprise-generated data will be created and processed outside a traditional centralised data centre or cloud by 2025. “As we continue to see more edge deployments, the combination of the two technologies [edge and 5G] will be a game-changer,” says Yehia.

With the decentralisation of computing technology, moving workloads from the cloud to the edge exposes a larger surface to cyber threats. “For edge computing, every device can be seen as a point of entry. This calls for the need to build in protection for data at the edge, with a plan that includes maintaining business and service continuity despite one or more edge sites being compromised,” warns Yehia.

Measures should be put in place beyond network and endpoint security that enterprises may rely on from providers. Designs, standards, processes and best practices geared toward minimising the risk of data loss should be baked into the process from the beginning, he recommends. “Additionally, protecting data at the edge can entail building a separate network fabric for data assurance operations, including backup, restore, archive and snapshot. With security done right, edge
computing can reap more benefits than pose risks.”

Edge computing can also mitigate some of the security shortcomings inherent in cloud infrastructures. With public cloud, ensuring security falls on the provider, and organisations don’t have much control over how their data is managed since the cloud is shared with other users, Yehia observes.

“The privacy and compliance problem regarding sensitive data (especially in the finance and healthcare fields), is better solved with edge since organisations have more control over their data, access, and security by filtering data at the source,”
he says. “Additionally, since data is processed onsite with edge computing, this minimises its risks for distributed denial-of-service (DDoS) attacks and other vulnerabilities, such as network disruptions and power outages,” Yehia adds.

Edge computing has the potential to transform healthcare, retail, transportation and logistics, gaming, and surveillance and monitoring industries. These sectors are increasingly moving towards AI and machine learning applications, while generating tonnes of data through devices and sensors, making real-time feedback and insights necessary. “Increased use-cases for moving processing closer to the data source are especially favourable in various industries due to the nature and volume of the data created. Examples include industrial sensors, autonomous vehicles, augmented reality/virtual reality use-cases, connected healthcare devices, smart logistics, real-time surveillance, etc,” says Yehia.

Though edge solutions can be leveraged to solve some of the limitations of cloud computing, the debate should not be framed as ‘edge vs. cloud’, rather, how the two should work in tandem. They both fall under the wider umbrella of employing
a hybrid approach that best suits business needs, says Yehia.

“The question has to do more with which computing workloads need to be placed where and why. If we assume that all workloads are deemed to be placed on the cloud, then edge can come in as a competitor. But that was never the case and the
deployment of cloud and edge should be seen as complementing each other rather than opposing.”

Edge isn’t going to replace cloud-based apps; it’s going to sit alongside it, as a necessary complement to allow organisations to get the most from their applications and data, agrees Baguley of VMware.

IT environments are becoming more decentralised, and organisations must be forward-looking to identify their unique needs to develop a robust hybrid approach that includes a mix of cloud, edge and core.

You might also like


Scroll To Top