Decoding Kubernetes (k8s)

23rd June, 2025

Kubernetes has revolutionized the way organizations deploy, manage, and scale applications. Originally designed for large-scale, cloud-native environments, Kubernetes is now finding its way into new frontiers, including edge computing and industrial automation. This blog post explores the fundamentals of Kubernetes, its traditional use cases, and how its strengths and limitations play out in edge deployments—especially in industrial settings.

Introduction: The Need for Kubernetes in Modern Infrastructure

Kubernetes (often abbreviated as K8s) is an open-source platform for orchestrating containerized applications across clusters of machines, whether physical or virtual. Developed by Google and now maintained by the Cloud Native Computing Foundation, Kubernetes automates deployment, scaling, and management of application containers, ensuring reliability and efficiency at scale.

Understanding Kubernetes Architecture

Kubernetes operates as a distributed system composed of two main layers:

Control Plane: The centralized layer responsible for orchestrating the cluster.
Worker Nodes: The decentralized layer where containers (applications) are executed.

Industrial automation controller

Image source: Wikimedia

Core concepts include:

Clusters: Groups of nodes (servers) running containerized workloads.
Nodes: Individual machines (physical or virtual) within a cluster.
Pods: The smallest deployable units, typically encapsulating one or more containers.
Services: Abstractions for exposing and load-balancing applications.
Namespaces: Logical partitions for organizing resources.

Kubernetes provides features such as automated rollouts and rollbacks, self-healing (restarting failed containers), load balancing, declarative configuration, and extensibility.

TRADITIONAL USE CASES OF KUBERNETES

Kubernetes excels in large, centralized computing environments. Its key strengths lie in:

Datacenter Deployments: Kubernetes was built for server farms and cloud infrastructure where nodes are virtual or physical servers housed in controlled, centralized locations.
Abstracted Physical Location: Since cloud-based applications don’t depend on hardware positioning, Kubernetes can freely move workloads to any available node.
Scalability & Load Balancing: Kubernetes can automatically scale applications and distribute workloads across multiple pods and nodes to meet changing demands.
Redundancy & High Availability: It keeps services online by maintaining multiple replicas and restarting failed containers on other available nodes.

These features power the world’s largest web platforms, SaaS products, and internal enterprise infrastructure.

Our Use Case: Kubernetes in Industrial Edge Deployments

At SALZ Automation, we are exploring the limits of Kubernetes beyond its traditional comfort zone. We want to use it in edge deployments — for example, on industrial PCs embedded directly into factory equipment, production lines, or control cabinets.

Our requirements diverge significantly from cloud-native assumptions:

Physical Device Relevance & Location-Conscious Deployments: Unlike the cloud, each edge device has a specific physical purpose tied to its location in the factory. A machine controller on one end of a conveyor system cannot be arbitrarily replaced by one at the other end. Deployments must consider physical device capabilities, connected sensors, and I/O hardware — not just compute resources.
Redundancy via Quick Spin-Up: We rely on fast recovery. If a device fails, Kubernetes should spin up a new instance — but ideally on a pre-positioned edge device nearby.
Centralized Remote Management & Updates: Ideally, all edge infrastructure—across a single plant or multiple factories—can be centrally monitored, managed, and updated from a web interface, reducing the need for on-site technical intervention.

Advantages of Kubernetes

ADVATNAGES OF KUBERNETES FOR INDUSTRIAL EDGE

Despite the friction, Kubernetes still brings major advantages for our vision of future-proof, scalable industrial automation:

Web-Based Factory Management: With proper setup, a factory’s entire fleet of automation nodes could be orchestrated and managed remotely from a browser-based control plane.
Less Need for Tiny Controllers: Instead of distributing hundreds of microcontrollers, we can centralize more intelligence in fewer, more powerful edge servers.
Cheaper & Scalable Hardware: Off-the-shelf rack servers or ruggedized industrial PCs can be reused across different tasks, reducing hardware variation and cost.
Seamless Remote Updates: Rolling out software updates or deploying new logic can happen without physical technician intervention — even during ongoing operations.

These strengths are especially valuable in a global manufacturing environment where uptime is critical and skilled technicians are not always on-site.

CHALLENGES FOR EDGE DEPLOYMENT

Not Built for Hardware Integration

Kubernetes abstracts hardware and is optimized for virtualized, cloud-native environments, not for direct hardware management

Mismatched Update Cycles

Kubernetes updates frequently, while edge devices are updated far less often. This mismatch can cause compatibility issues, increase maintenance effort, and introduce potential security risks.

Resource Overhead

Edge devices often have limited CPU and memory, and the Kubernetes stack can be resource-hungry for such environments.

Slow Boot & Download Times

When a controller fails, spinning up a replacement instance might involve large image downloads and slow start times — not acceptable in critical industrial processes

Connectivity Constraints

Redundancy and failover are only possible for devices connected via reliable networks (typically Ethernet), limiting applicability in some industrial scenarios.

The Way Forward

To optimally use Kubernetes at the edge, we need a container-first mindset specific to industrial realities. This includes:

Lightweight Kubernetes distributions like K3s or MicroK8s, optimized for edge performance.
Smart scheduling with location-aware metadata to ensure applications only run on physically suitable hardware.
Preloading critical container images on local devices to avoid long startup delays.
Integration layers that expose hardware I/O to containers in a secure and real-time-safe way.
A unified platform where IT (Kubernetes) and OT (automation engineering) meet — without fighting each other.

Conclusion

Kubernetes remains a powerful tool for orchestrating containerized applications, offering automation, scalability, and resilience. While its roots are in datacenter and cloud environments, its principles are increasingly being adapted to edge and industrial use cases. The promise of centralized management, rapid updates, and hardware abstraction is appealing for factories and industrial sites. However, practical challenges—especially regarding hardware overhead and network constraints—must be carefully considered.

As Kubernetes continues to evolve, expect to see further innovation aimed at making it more lightweight and edge-friendly, helping bridge the gap between cloud-native architectures and the realities of industrial automation.

For beginner insights, read our previous blog on CODESYS Virtual Safe Control: Simplifying Functional Safety with Containerized Automation.

For hands-on labs, check out Kubernetes’ official tutorials.

Are you working on a proof of concept, pilot project, or considering Kubernetes for your industrial automation needs? We’d love to hear from you! Reach out to us to discuss your ideas, challenges, or experiences with Kubernetes at the edge!