Each node in an AKE workload cluster corresponds to a virtual machine or physical machine. Cluster nodes are divided by node group, including one control plane node group (used to run the control plane of the K8s cluster) and one or more worker node groups (used to run the worker nodes of the K8s cluster). Based on differences in cluster node types, AKE workload clusters can be divided into VM-based workload clusters (VM-based clusters for short) and physical-machine-based workload clusters (physical-machine-based clusters for short).
A VM-based cluster is a workload cluster in which both control plane nodes and worker nodes are virtual machines.
A physical-machine-based cluster is a workload cluster that uses virtual machines as control plane nodes and physical machines as worker nodes. Users can enter SSH and other information through the AKE UI to conveniently register physical machines with the AKE service, and then select to create a physical-machine-based cluster, that is, select multiple physical machines as worker nodes to build a K8s cluster.
AKE also installs kube-vip, the AKE package manager, and some preset AKE cluster addons (listed below) in workload clusters to provide infrastructure services such as high availability of the cluster control plane, CNI, CSI, monitoring, logging, Ingress Controller, LoadBalancer, node configuration, and identity authentication.
The components of an AKE workload cluster are shown in the following figure.
