Companies are pursuing digital transformation at an unprecedented rate, propelled often by the innovation enabled by edge computing. Aware of the limitations of legacy systems and cloud computing, businesses are looking to leverage edge devices to harness data, streamline operations, and run complex workloads in a flexible, fast and resilient way.
The data reflects this story. Recent estimates forecast that there will be 27 billion connected IoT devices by 2025 (IoT Analytics 2022). Demand is being driven by strong appetite from particular sectors, such as automotive, telco, manufacturing and retail, where optimized supply chains and advanced automation have outsized impact. Industry 4.0 is no longer just a buzzword; it is happening right now.
SEE: Hiring Kit: IoT developer (TechRepublic Premium)
But while there is a shared vision of edge and the role it will play in overhauling traditional infrastructure, this consensus doesn’t change the fact there are significant challenges to overcome.
At a foundational level, businesses are often at very different levels of maturity in the transition to edge solutions. At our annual SUSECON Digital 2022 conference, we saw first-hand that edge cases are diverse and no one solution fits all. Instead, solutions must be tailored to where a business is on its journey, resulting in a pressing need to create optimal solutions in mixed environments involving legacy hardware while adopting Kubernetes.
The challenges of running applications at the edge
As enterprises seek to realize their Edge vision in these environments, there are three central hurdles which need to be addressed beyond the inherent complexity of containerized workloads.
Scale is the first significant challenge for most customers. Edge environments are far larger than those introduced in traditional enterprises. For example, looking at the automotive sector, a modern car is a computer itself: From the engine to the braking system, even a reversing camera — whenever a compute happens, there is an analytical event that takes place in the car, instead of that data needing to be shared to a central location.
Your car represents a series of edge devices. Consider the number of edge devices managed by BMW — not just the cars, but also retail and manufacturing locations — the scale is significant. The same remains true for big box retailers such Home Depot, which uses edge to manage thousands of its retail storefronts to streamline operations and applications.
Security is the second challenge. Edge deployments introduce new threat vectors and significantly expand the attack surface area. With systems sitting outside of a classical data center, they need to be secured from foundation applications all the way through to operating environments and the workloads themselves.
Management is the third challenge. If you have thousands of edge devices, how can you quickly configure them all? This can be practically managed from a central location, having one device that is connected to a cluster of edge devices, enabling a batch update without the need for physical intervention onsite. By having a common platform centered on K3s, one can deploy and update the underlying container platform quickly and effectively. With this, resiliency is frequently front of mind as consistent communication is never guaranteed in edge deployments.
How can enterprises navigate security risks?
Security must be the invisible thread that runs throughout the entire environment. As such, it must be baked in across the full lifecycle management.
Resilience must be weaved into edge infrastructure from the outset, which means that the onboarding process must be secure. For it to be secure, it must also be simple to deploy. Practically, if you have three edge nodes at a site, and need to add a fourth, you don’t need a team of IT people involved. It’s about getting the node (a box) shipped to the location, having someone who can physically plug it in onsite, for the node to then be updated from a central location without any need for onsite intervention.
This is the full lifecycle management, with any updates being made remotely to that node, connected to the other nodes as a cluster. It’s an integral part of navigating security risks that involves having a zero trust security approach. This is a practical approach to mitigate against risk and handle the new threat vectors introduced by edge deployments. With the proliferation of devices in edge environments and the insufficiencies of traditional verification approaches, models that do not trust by default are becoming mandatory.
Unlocking the potential of edge
If these challenges are met, enterprises will be able to truly deploy IoT devices at a scale that far outstrips traditional infrastructure. This scale can be secure.
We need only look to the industrial IoT field to see that the opportunities are immense, from predictive maintenance of machinery to remote, rapid monitoring of equipment.
For example, if you’re at a lumber factory, there can be a device installed that can predict how long a saw blade will last, therefore reducing the number of people who are required to manage it as well as increase the lifespan of the saw blade. Timely analytics inputs using edge devices here save costs of tools and improve productivity.
It’s easy to see how this essential part of maintenance can translate to other industries. Whether it’s the management of robots in a warehouse, monitoring of heavy goods vehicles or reading the analytics from medical devices. Pursuit of omnichannel strategies and walkout technology are two edge-enabled innovations driving change in the retail sector.
Access to real-time analysis and harnessing previously lost data will further produce analytics value for a business, especially the ability to identify and remove inefficiencies from the supply chain. Edge will be able to bring applications closer to the end-user for a vastly enhanced experience. The opportunities presented by edge computing are clear for enterprise and customer alike, but so too are the risks if security doesn’t keep up with the pace of digital transformation.
Addressing the challenges of scale and carefully deploying solutions that weave both security and resilience into the core infrastructure are unavoidable if enterprises are to succeed in what is fast becoming an inevitable journey to the edge.
With over 21 years in cloud and related industries, Basil leads product management, marketing, engineering, and communications alignment for SUSE’s Edge business unit. Prior to this, he was Vice President of Cloud Native Infrastructure, where he drove strategy and management of SUSE Rancher cloud-native products. Before Rancher, Basil led product management, positioning and business strategy for security within Red Hat’s Cloud Platforms business unit.