Hybrid Cloud Apps with OpenShift and Kubernetes

Chapter 6. Multicluster Fleets: Provision and Upgrade Life Cycles

The terms multicluster and multicloud have become common in today’s landscape. For the purposes of this discussion, we will define these terms as follows:

Multicluster

Refers to scenarios where more than a single cluster is under management or an application is made up of parts that are hosted on more than one cluster

Multicloud

Refers to scenarios where the multiple clusters in use also span infrastructure substrates, which might include a private datacenter and a single public cloud provider or multiple public cloud providers

The differences here are more academic; the reality is that you are more likely than not to have to manage many clusters just as your organization has had to manage multiple VMware ESXi hosts that run VMs.

The differences will matter when your container orchestration platform has variation across infrastructure substrates. We’ll talk about some of the places where this currently comes up and may affect some of your management techniques or application architectures.

Why Multicluster?

Let’s discuss the use cases that lead to multiple clusters under management.

Use Case: Using Multiple Clusters to Provide Regional Availability for Your Applications

As discussed in Chapter 4, a single cluster can span multiple availability zones. Each availability zone has independent failure characteristics. A failure in the power supply, network provider, and even physical space (e.g., the datacenter building) ...