Compute design for Red Hat OpenShift virtualization

IBM Cloud® provides compute options that support diverse workloads, from traditional applications to modern cloud-native solutions. These offerings deliver flexibility, scalability, and enterprise-grade security, that help you deploy diverse workloads.

IBM Cloud VPC offers two compute options:

The key compute architecture elements are shown in the following diagram.

Red Hat OpenShift Virtualization on IBM Cloud Compute
Red Hat OpenShift Virtualization on IBM Cloud Compute

Red Hat OpenShift worker nodes

Red Hat® OpenShift® Kubernetes Service is a managed service that uses Red Hat OpenShift clusters where you can deploy and manage virtualized and containerized applications. A Red Hat OpenShift Kubernetes Service cluster has a managed control plane and one or more worker pools. A worker pool consists of two or more compute hosts that are called worker nodes. Worker pools contain worker nodes of the same profile of CPU, memory, operating system, attached storage, and other properties. The worker nodes are managed by the Kubernetes control plane, which controls and monitors all Kubernetes resources that are in the cluster. The Kubernetes scheduler decides which worker node to deploy resources on and accounts for deployment requirements and available capacity in the cluster.

After a cluster is created, you can add more worker nodes to a pool by resizing it or by adding extra worker pools. Clusters that have a worker pool in only one zone are called single zone clusters. For high availability, you can create multizone clusters with worker pools that span multiple availability zones. You can create multizone clusters, but they are not recommended for virtualization workloads because of storage latency.

IBM Cloud Bare Metal Servers for VPC are recommended in the worker pool to run your production virtualized workloads because Red Hat supports only bare metal worker nodes for production virtualized workloads. The bare metal nodes must be provisioned with local NVMe drives. So they can be used by Red Hat OpenShift Data Foundation (ODF) as backing storage.

For more information, see Red Hat OpenShift overview.

IBM Cloud Virtual Servers for VPC

IBM Cloud Virtual Servers for VPC provide secure, isolated virtual machines that are deployed within a Virtual Private Cloud environment. You can use these instances for production workloads, and development and test environments that require flexible resource allocation and comprehensive infrastructure control.

For more information about virtual servers, see About virtual server instances for VPC.

The following table lists the key features for IBM Cloud Virtual Servers for VPC.

IBM Cloud Virtual Servers for VPC key features
Feature Description
Customizable profiles Balanced, compute-optimized, memory-optimized, GPU, and high memory configurations
Flexible tenancy Shared tenancy infrastructure with optional dedicated host placement for compliance requirements
Advanced networking Integration with VPC Security Groups, Network ACLs, Load Balancers, and VPN connectivity
Persistent storage IBM Cloud Block Storage and File Storage for VPC with configurable IOPS and encryption
Operating system flexibility IBM-provided stock images or bring-your-own custom images
Scalability Vertical scaling through profile changes and horizontal scaling through instance groups with auto-scaling

For more information about virtual server instance profiles that are supported by Red Hat OpenShift Kubernetes Service, see IBM Cloud Docs - Worker Nodes VPC flavors.

IBM Cloud Bare Metal Servers for VPC

IBM Cloud Bare Metal Servers for VPC provide single-tenant, dedicated physical servers that deliver maximum performance, security, and control. These servers are required for production deployments of Red Hat OpenShift Virtualization.

The following table lists the key features for IBM Cloud Bare Metal Servers for VPC.

IBM Cloud Bare Metal Servers for VPC key features
Feature Description
Single-tenant isolation Dedicated physical hardware with no resource sharing
High-performance NVMe storage Local NVMe drives for Red Hat OpenShift Data Foundation deployments
Consistent performance Predictable performance without virtualization resource overuse
Large memory configurations Support for memory-intensive virtualization workloads
Network performance High-bandwidth, low-latency networking for VM traffic

Bare Metal worker nodes that are supported by Red Hat OpenShift Kubernetes Service varies by region and Availability Zone. For more information, see worker node VPC flavors.

The following table shows the typical specifications for a bare metal profile.

Typical bare metal profile specifications
Instance profile Cores GiB RAM Storage
cx2.metal.96x192 48 192
cx2d.metal.96x192 48 192 8x3.2TB SSD NVMe drives
bx2.metal.96x384 48 384
bx2d.metal.96x384 48 384 8x3.2TB SSD NVMe drives
mx2.metal.96x768 48 768
mx2d.metal.96x768 48 768 8x3.2TB SSD NVMe drives

For more information about bare metal servers, see About Bare Metal Servers for VPC.

Next steps

Now that you understand the compute design for Red Hat OpenShift Virtualization, explore these related topics: