IBM Cloud Platform Architecture and Compute Options
Domain 2 - IBM Cloud Platform Architecture and Compute Options. This domain tests how IBM Cloud organizes resources globally and how to select compute for workload shape, control, and operating model.
Platform structure
IBM Cloud resources live in an account and are organized into resource groups for lifecycle and access management. Regions contain multizone regions with three zones for high availability; services may be regional or global. A VPC is regional and spans zones; subnets are zonal. Resource groups are not network boundaries.
Compute choices
- Virtual Server for VPC: a VM with instance profiles, images, SSH keys, and security groups; good for general workloads.
- Bare metal: a dedicated physical host for performance, licensing, or strict isolation.
- Power Systems: Power virtual servers for AIX, IBM i, and Linux workloads needing Power architecture.
- Red Hat OpenShift on IBM Cloud or IBM Cloud Kubernetes Service: container orchestration with worker nodes.
- Code Engine: serverless containers, jobs, and functions; scales to zero and reduces cluster operations.
- VMware Solutions: extend VMware estates with dedicated IBM Cloud infrastructure.
Worked example
A team has a stateless API with unpredictable traffic and no Kubernetes skills. A VPC virtual server requires patching and idle capacity; Code Engine fits because it scales on requests and bills for use. If the API needs custom networking and persistent daemons, OpenShift is a better fit.
Traps
Do not place a zonal subnet across zones; do not treat a resource group as an IAM boundary by itself; do not choose bare metal for spiky stateless traffic; and do not assume every service is available in every region or zone.