Simplifying and optimizing Red Hat OpenShift on OpenStack with hosted control planes

Cloud-native infrastructure is still evolving, and one thing that's clear is the need for scalability and efficiency in managing your Red Hat OpenShift clusters. Hosted control planes (HCP) is a paradigm-shifting approach in OpenShift designed to optimize the way control planes are deployed and operated. For cluster service providers operating Red Hat OpenStack Services on OpenShift, the HCP model aligns perfectly with the demands of modern infrastructure. By separating concerns of management and workloads, providers can achieve better resource utilization with high density control, help ensure high availability and simplify lifecycle management—critical aspects for environments where uptime is paramount. This article provides an overview of HCP features in OpenShift.

Benefits of hosted control planes

With the release of Red Hat OpenShift 4.17, HCP for OpenShift on OpenStack is available as a developer preview. Red Hat OpenShift 4.18 will bring significant improvements, integrating OpenShift innovations with robust infrastructure like OpenStack Services on OpenShift by simplifying and optimizing the clusters lifecycles.

Resource and cost efficiency

In traditional OpenShift deployments on cloud platforms, control plane components run on three dedicated virtual machines (VM), called masters. This model often leads to underutilized resources due to static allocation. With HCP running on an existing OpenShift cluster (the management cluster), control planes are decoupled and hosted independently in pods in isolated namespaces, allowing dynamic scaling and more granular resource allocation. OpenStack Services on OpenShift’s flexible compute, network, and storage services provide a foundation that enables providers to optimize resource usage across their infrastructure and naturally reduce costs.

Improved multi-tenancy

For providers hosting clusters for multiple internal or external customers, multi-tenancy is critical. HCP’s separation of control planes aligns with OpenStack’s tenant model, allowing providers to manage project resources (such as quotas) independently for each hosted cluster. Because HCP makes multi-cluster easier, it's easier to implement the strongest form of tenancy (cluster per tenant).

Hybrid cloud enabler

HCP enables a hybrid cloud architecture where the management cluster responsible for hosting the HCP can run on bare metal, support public clouds and on-premises. With the OpenShift 4.17 release, the hosted cluster infrastructure (which hosts the user workloads) can now run on OpenStack clouds. This flexibility allows providers to balance performance, cost and compliance requirements while taking full advantage of OpenStack Service on OpenShift’s capabilities for workload hosting.

Use case: HCP on OpenStack (hub and spoke)

Suppose you deploy a management cluster as a hub cluster for other clusters within an OpenStack environment. In this cluster, you can install the Hypershift Operator to handle the management of control plane pods required for managing hosted clusters in a dedicated namespace for each cluster.

Each HCP can map to distinct OpenStack projects (tenants) to provision and manage dedicated worker node machines called nodepools. This design allows for a master hub cluster to manage multiple bespoke hosted control planes with each mapped to a unique project and quota in OpenStack. Dedicated sets of credentials ensure that nodepool machines are provisioned securely and isolated for each cluster.

This architecture is compatible with OpenStack environments running Red Hat OpenStack Platform 17.1 and Red Hat OpenStack Services on OpenShift. It provides a streamlined and scalable solution for managing multiple clusters without the need for complex prerequisites, offering flexibility for cluster service providers to operate efficiently across diverse projects and different clouds..

Improved control planes

OpenStack’s integration to HCP marks a significant step forward in the Shift-On-Stack evolution, offering cluster service providers a modern and efficient way to manage Kubernetes control planes. While hosting control planes on OpenStack is new, the concept of hosted control planes has matured over years of iterative development for public clouds including AWS (with the Red Hat OpenShift on AWS offering) and Azure (with Azure Red Hat OpenShift), but also on-prem with Kubevirt and Agent platforms. It has consistently proven its value by simplifying cluster management, enhancing scalability and optimizing resource utilization.

For providers looking to elevate their services, the combination of Hypershift and OpenStack provides an innovative solution that is both robust and future-ready. This is just the beginning. Stay tuned as we continue to explore how these tools and strategies empower platform service providers to build, manage, and scale their offerings in an increasingly complex cloud-native landscape.

Read next: How to build hosted clusters on the OpenStack platform