Perhaps a silly question after I just enrolled in the course to learn it, but could someone give me a layman’s explanation of when you would use OpenShift vs using one of the cloud provider’s Kubernetes or Docker Swarm offerings? At one point, I thought OpenShift would be what you’d deploy if you wanted to set up your own internal cloud.
Let’s ignore docker swarm for now because it is a different thing altogether
Openshift is kubernetes. However it is a curated version of kubernetes built by Red Hat which has a few important differences from standard kube.
- Firstly, it is not free. It has an annual licence cost which can run into many thousands of dollars depending on the size of the cluster and number of users. Regular kube has no license requirement - you only pay for the servers you run it on.
- It is strongly opinionated. There are things you just cannot do in Openshift due to its built-in security features, such as run pods with root privilege.
- It comes with a lot of stuff baked in, such as monitoring/logging, and various additional methods of authentication. In regular kube, you have to install and manage that stuff yourself as extras. These add-ons are also all mostly open source.
- It is easier to set up. Its installer does all the work of creating the cluster and configuring the add-ons.
- Requires less devops knowledge to operate.
- There is a version of Openshift that can be deployed directly to AWS, it again it costs much more to run than regular EKS.
Swarm is not kubernetes. It is a very simple way of orchestrating multiple containers and that’s all it does. Runs the containers and ensure they can talk to each other. No real security, access control, traffic routing policies or any of that.
Thank you for that very good explanation. It strikes me then that the main use case of OpenShift would be to set up a K8S cluster in your on-prem network, although there may be some cases where you use it instead of your cloud provider’s K8S offering.
Mare particularly, that you are prepared to spend a lot of extra money for a lower administrative burden of running a cluster.