Edge Kubernetes reference architecture

Log in to AWS
  • In this application, we deploy Amazon’s managed Elastic Kubernetes Service (EKS) to the network edge using AWS EKS in 5G Edge Wavelength Zones. To do so, simply extend your virtual private cloud (VPC) to a Wavelength Zone as part of a new subnet. Follow the steps below to deploy an EKS cluster at the network edge while understanding the key differences with EKS in AWS Wavelength. Visit the AWS Compute Blog to learn more about EKS support on AWS Wavelength.

    Infrastructure deployment plan

    1. Launch EKS cluster control plane in parent region (multi-AZ)

    2. Configure EC2 and Amazon Elastic Container Registry (ECR) interface endpoints to communicate with control plane

    3. Add worker nodes to the EKS cluster by launching EC2 instances with specialized Amazon Machine Images (AMI), called Amazon EKS-optimized AMIs, in the Wavelength Zone subnet(s) using self-managed nodes CloudFormation template 

    4. For self-managed nodes, download kubectl to communicate with the cluster API server

    5. Deploy pods to EKS cluster