Cost is always a factor in technology decisions, but every decision is multiplied when designing edge infrastructure with hundreds or thousands of nodes. Total cost of ownership is critical — especially operations and deployment on-site at remote locations — and the long-term duration of deployments has to be factored in. Part of the solution is designing a very compact and flexible system, but the system also needs to work with nearly any configuration, from virtual machines to Kubernetes.
Watch the episode
Exploring the economic considerations of edge
The episode opens with hosts Stephen Foskett and Alastair Cooke noting that edge computing requires a different approach — financial resources at edge locations are often constrained, especially in industries like retail that operate on thin margins. Carlo highlights that the economics of edge are too often overlooked, with focus placed on the technology itself or on specific use cases. Deploying solutions at scale across multiple locations with varying conditions and constraints presents unique challenges. The economics have to account for not only what works in a lab, but also real-world deployment and long-term operation.
Every decision made at the edge has a multiplier effect. Across hundreds or thousands of sites, even small inefficiencies compound — and so do small wins.
The conversation explores how cost-effective solutions deliver value in edge locations that previously wouldn't have been considered for compute. Carlo shares examples of NodeWeaver customers running cost-effective edge solutions in industrial automation, video recording, and AI processing. He emphasises the importance of considering the entire lifecycle of an application — hardware replacement, maintenance, and support — and the ability to handle failures autonomously, which is crucial in edge environments where failures are expected and on-site troubleshooting may be costly.
VMs, Kubernetes, and what real customers run
The conversation touches on the choice between virtual machines and container-based solutions like Kubernetes at the edge. Carlo notes that while there is a strong trend toward containers, many customers still run VMs for their applications. The level of reliability and manageability required depends on the specific application and the customer's needs — and a good edge platform has to be flexible enough to host both.
The episode closes with the hosts highlighting the gap between the idealised vision of edge computing — Kubernetes-everywhere, containers-only — and the reality of legacy systems and production environments at the edge. They appreciate NodeWeaver's focus on autonomous management and the minimal infrastructure needed to deliver value. DNSOps, where DNS entries are used to drive configuration, is highlighted as a good example of lightweight infrastructure that fits the edge.