Openshift vs Rancher
As a professional, who had experience on both platforms, this Blogpost aims to reveal the realities.
The statements in this comparison are based on real Openshift/Rancher customer feedbacks, technical evidences and wider experience...
As an administrator or developer on an Enterprise Kubernetes platform, one would expect the following features on the top of a vanilla Kubernetes:
- Not to be in a vendor-lock
- Painless version upgrades
- Native Kubernetes Experience
- Support to the Kubernetes/CNCF ecosystem
- In Sync with the latest Stable Kubernetes versions
- To be able to switch to the vanilla Kubernetes without reinstalling
- Community focus and Open Source approach
OpenShift
IBM Closed Landmark Acquisition of Red Hat during 2019, which means that Openshift is now an IBM product.
Openshift cluster deploys and manages itself. Installation and deployment consists of a bundle of Ansible Playbooks built into a container image with an Ansible runtime. The infrastructure is immutable. Openshift initial installation is reported that it usually takes days, in some cases weeks, especially in complex environments.
Openshift uses operators for everything, which brings operational knowledge into how Kubernetes applications are deployed.
Openshift Includes tools like S2I and Builds that enable developers to build container images without Dockerfiles.(Although, this approach may implicitly move you away from the general community practices)
Openshift has security controls for cluster administrators when designating what applications users are permitted to deploy.
The cluster can scale itself and Kubernetes resources as needed.
Openshift has relatively longer installation and upgrade times. This is not suitable for CI/CD solutions that deploy the entire clusters during the platform deployment pipelines. Version upgrades on Openshift are reported to be risky and painful. Some major upgrades have been reported to take all night till the morning. And a specific version upgrade has been reported that it created disruption on the overall system.
The heavy use of operators gives less flexibility overall. Operation Teams may find it harder to administrate because it relies on Operators for everything.
Openshift positions itself more packaging platform which takes open source solutions and modifies them and ships them as a proprietary solution.
Openshift in general, has an alternative toolset and approach rather than using CNCF practices. It has its own practices like its non-kubernetes-native cli called oc-client.
Once you installed Openshift, there is no way to switch back to vanilla Kubernetes.
Rancher
Rancher Labs delivers the industry’s most widely adopted open-source Kubernetes management platform, which is founded in 2014, Cupertino, CA. Rancher currently have 150+ employees, operations in 12 countries and got Series ’C’, total funding of $55M.
Rancher simplifies complex Kubernetes operations while maintaining the flexibility.
Rancher follows industry best practices and uses tooling from CNCF.
Rancher extends these best practices through automation and by making complex configurations easier to build.
The principal architect of Rancher Darren Shepherd is a Docker contributor and also on Docker Advisory Team, who later also heavily invested in Kubernetes. Rancher is a company whose Longhorn project accepted to CNCF(cncf.io --> Projects --> Sandbox Projects)
Rancher delivers a single suite of tools for managing multiple clusters.
Rancher Apps install and upgrade apps on multiple clusters at the same time. Global DNS makes them all available.
Rancher launches clusters in minutes and capable to run Kubernetes anywhere.
Rancher follows API-first design, which means that all functions have an API equivalence.
Comparison Chart
I tried to group the expectations and scored them together with several professionals:
Installation easiness |
* Installation is reported that it takes from 3 days to weeks, mostly reported takes more than a week |
*** Installation between 20 minutes to 2 days
|
CNCF/Industry Standards |
* Although RedHat contributes to the Kubernetes project, Openshift is reported to favor its own tools&practices which creates vendor-lock and moving away from common community practices |
***
Follows and extends industry standards |
Open Source |
* Community version is Open Source but Enterprise version is not |
*** %100 Open Source |
Licensing |
* vCPU based licensing makes it expensive for nodes having high number of vCPUs |
** Number of nodes are counted, number of vCPUs are not counted |
Multi-Cluster |
** Recently included to the platform |
*** Stable for many years |
Upgrades |
* Failed version upgrades and rollbacks are reported. Several disruptions after upgrades are reported. |
***
Version upgrades are smooth Zero-downtime version upgrades |
Kubernetes Version |
* Usually follows 3 version back. (OCP 4.2 is based on Kubernetes 1.14) |
** Follows max. one version back. (Rancher 2.3 support Kubernetes 1.16, 1.17Beta) |
Vendor-lock |
---
Once installed, switching to vanilla Kubernetes is not possible |
*** Rancher Management and Kubernetes are decoupled, switching to vanilla Kubernetes is possible |
Windows Container Support |
* Windows nodes are supported with limited CNI alternatives |
** Windows nodes are supported with more CNI options including Calico |
Support |
** Support for the platform is provided
|
*** Supports not only the platform but also the tools in the Kubernetes ecosystem |
Sales |
***** Noticeable sales power and marketing influence |
** Limited sales and marketing |
Partner ecosystem |
***** IBM partners worldwide are OCP partners |
** Limited partner channel
|
Bundle options |
***** Bundled into IBM deals which makes it attractive from a cost point of view |
*** Vendor agnostic approach
|
TOTAL |
26 |
34 |
That does not mean one is better than the other.
According to the Forrester report 2020, it shows strategically OpenShift and Rancher as leaders:
I hope this Blogpost was useful for the teams during their Kubernetes journey.
The statements in this Blogpost are gathered from OpenShift and Rancher customers.
Please do not hesitate to comment below in case you also have relevant experience and also you can contact me if you have any questions in mind.
Derya (Dorian) Sezen
Derya, a.k.a. Dorian, ex-CTO of an amazon.com subsidiary, is currently working as Cloud and DevOps Consultant at kloia.