The original release of Anthos required you to deploy your clusters on a vSphere infrastructure and didn’t offer the option to deploy on a different hypervisor or to physical servers. For the initial release, this made sense because vSphere is used by numerous enterprises, and it allowed businesses to use their existing infrastructure and skill sets. As the use cases for containers and Kubernetes grew, however, it became clear that organizations wanted, and needed, more flexible deployment options.
To address these additional use cases, Google expanded Anthos to include a bare metal deployment model. One point to highlight is that you do not have to deploy Anthos on bare metal to actual physical servers. The bare metal model allows you to deploy to any supported operating system, whether a physical server or virtual machine, or even VMs running on Hyper-V or KVM.
You can think of the bare metal option as a “bring your own Linux” deployment model. Rather than having an appliance to deploy your nodes, like the vSphere deployment model, you need to provide ready-to-use servers before you can deploy Anthos on bare metal. Now let’s introduce you to Anthos on bare metal.