Treat Your Virtual Machines Like Physical Hardware

New tools allow teams to manage virtual machines with the same workflows used for physical bare-metal servers, all within Kubernetes.

A powerful new integration is changing how teams manage virtual machines (VMs) within Kubernetes. According to a recent demonstration on the CNCF Blog, the Metal3 project can now provision and manage KubeVirt VMs just as if they were physical, bare-metal servers. This breakthrough is made possible by a component called KubeVirtBMC, which effectively tricks bare-metal automation tools into seeing VMs as physical hardware. The development marks a significant step toward a truly unified infrastructure management plane, where the distinction between physical and virtual resources begins to fade. For organizations running mixed environments, this offers a path to simplify complex provisioning workflows under a single, consistent, Kubernetes-native approach, eliminating the need for separate toolchains.

This capability hinges on emulating a key piece of physical server hardware. Bare-metal servers include a Baseboard Management Controller (BMC), a small, dedicated computer that allows administrators to perform out-of-band management tasks like powering the server on or off, regardless of the operating system's state. Provisioning tools like Metal3 are designed to communicate with these BMCs using standard protocols like IPMI and Redfish. The innovation here is KubeVirtBMC, which creates a virtual BMC endpoint for every VM running under KubeVirt. When Metal3 attempts to connect to a machine to provision it, KubeVirtBMC intercepts the request and presents an interface that looks and acts exactly like a physical BMC. This allows Metal3 to execute its entire bare-metal provisioning lifecycle—power control, boot device selection, and OS installation—on a VM, without realizing it isn't a physical machine.

This matters immensely for platform engineering and DevOps teams who struggle with fragmented toolchains. Historically, managing bare-metal servers required a completely different set of tools and processes compared to managing cloud or on-premise VMs. This operational divide creates complexity, slows down automation, and requires engineers to be proficient in multiple, often proprietary, systems. By enabling a single tool like Metal3 to handle both, this integration breaks down those silos. Teams can now define and manage the lifecycle of all their compute resources—physical or virtual—using the same declarative, API-driven, and version-controlled methods they already use for their Kubernetes applications. This brings the full power of GitOps and infrastructure-as-code principles to a domain that was often difficult to automate consistently.

The business impact of this unification is direct and substantial. Standardizing on a single provisioning workflow significantly reduces operational overhead, minimizes the potential for human error, and lowers training costs. It accelerates the process of deploying new infrastructure, allowing companies to respond more quickly to changing business demands. For organizations heavily invested in the Kubernetes ecosystem, this integration deepens its value proposition, extending its control plane beyond containers to become the central nervous system for their entire data center. By treating VMs as first-class, hardware-like citizens within Kubernetes, companies can maximize their existing investment and talent pool, driving greater efficiency and building a more resilient, agile, and cost-effective technology stack.