AKS Edge Essentials requirements and support matrix

This article describes the requirements for the host machine that runs AKS Edge Essentials:

Hardware requirements

Specs Local cluster Arc-connected cluster and GitOps
Host OS Windows 10/11 IoT Enterprise/Enterprise/Pro and Windows Server 2019, 2022
Total physical memory 4 GB with at least 2.5 GB free 8 GB with at least 4.5 GB free
CPU 2 vCPUs, clock speed at least 1.8 GHz 4 vCPUs, clock speed at least 1.8 GHz
Disk space At least 14 GB free At least 14 GB free
Disk type SSD SSD

To better understand the concept of vCPUs, see this article.

To better understand the etcd hardware recommendations, see this article.

You can run AKS Edge Essentials in an Azure VM. You can create a Windows VM with either Windows 10/11 IoT Enterprise/Enterprise/Pro and Windows Server 2019, 2022 SKU, on a VM image that supports nested virtualization such as the Dv5 and Dsv5-series virtual machines. When using an Azure VM, a premium SSD is required, per the Azure Virtual Machine documentation.

You can also run AKS EE in a virtual machine on VMware and Hyper-V as described here.

OS requirements

Install Windows 10/11 IoT Enterprise/Enterprise/Pro on your machine and activate Windows. We recommend using the latest client version 22H2 (OS build 19045) or Server 2022 (OS build 20348). You can download a version of Windows 10 here or Windows 11 here.

Note

Windows 10 IoT Enterprise, Windows 11 IoT Enterprise, and Windows Server 2022 IoT Enterprise are binary-equivalent to the non-IoT releases, differing in licensing and distribution.

Maximum hardware specifications supported

Parameter Permissible limit
Maximum number of VMs per machine 1 Linux VM + 1 Windows VM (optional)
Maximum number of vCPUs assigned to virtual machines 16 vCPUs
Maximum number of machines per cluster 15 machines

Note

The vCPU limit is per-host, not per-node.

GA feature support matrix

  • Kubernetes Distribution supported: Kubernetes (K8S) - Version: 1.24.3 and on Kubernetes (K3S) - Version: 1.24.3
  • Deployment options: Single-machine clusters and full Kubernetes deployment on single machines only. Full deployment across multiple machines isn't supported in GA.
  • Workloads: Only Linux worker nodes.
  • Network plugins: Calico on K8S and Flannel on K3S.

Experimental or prerelease features

  • Deployment options: Full Kubernetes deployment on multiple machines.
  • Workloads: Windows worker nodes.
  • Network plugins: Flannel on K8S and Calico on K3S.

Next steps