Running on Kubernetes
Using the ODH operator, the ML platform truly unlocks the full potential of Kubernetes as the infrastructure layer of your ML platform. The Operator Lifecycle Management (OLM) framework enables the ODH operator to simplify the operation and maintenance of the ML platform. Almost all operational work is done in a Kubernetes-native way, and you can even spin up multiple ML platforms with a few clicks. Kubernetes and the OLM also allow you to implement the Platform as Code (PaC) approach, enabling you to implement GitOps practices.
The ML platform you've seen in this book works well with vanilla Kubernetes instances or any other flavors of Kubernetes or even a Kubernetes-based platform. In fact, the original ODH repository was mainly designed and built for Red Hat OpenShift.
Avoiding vendor lock-ins
Kubernetes protects you from vendor lock-ins. Because of the extra layer of containerization and container orchestration, all your workloads do not run...