Fully management container service for OCI
November 12, 2019
Fully management container service for OCI
Score 6 out of 10
Vetted Review
Verified User
Overall Satisfaction with Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE)
Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE) is used for some of the R&D related, non-customer facing workloads including applications for AI / Deep learning training in the cloud (OCI). It is currently not used for consumer facing application workloads as our use is recent.
- Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE) was straightforward and easy to use to deploy containerized applications in Oracle Cloud (OCI) that makes it one of the top choices if you are already using OCI.
- OKE can be accessed using the Kubernetes dashboard which is more or less the gold standard for Kubernetes based container deployments makes it easy to administer.
- OKE is fully managed by Oracle and resides in the cloud that keeps the cost to manage the containers low without need for a lot of customization for performance.
- Oracle Cloud Infrastructure Container Engine for Kubernetes (OKE) is a fully managed service that makes it a little hard sometimes to fine tune HPC workloads, for example AI/Deep learning training where a lot of data transfer and compute is needed.
- The initial setup can certainly be improved with less steps to make it faster and easier to deploy compared to the competition, for example Google Kubernetes Engine (GKE).
- Documentation and best practice documentation for OKE can be improved especially since there are not a lot of customer references available for large scale deployments that need fine tuned performance.
- A positive impact on ROI for us from OKE as it reduced the time to deploy and manage Kubernetes containers with little use of in-house cloud administration staff.
- Since OKE is fully managed -- after the initial setup and deployment -- there is little need for dedicated staff to manage the clusters based on OKE reducing the CapEx incurred.
- As long as fine tuning of the containers with underlying cloud infrastructure is NOT needed, OKE is able to deploy cloud native and other existing workloads in the cloud (OCI) easily reducing the time to go-live.
Since we already use OCI, OKE was one of the top choices. Other top choices were Google Kubernetes Engine, Redhat OpenShift Container Service and Microsoft Azur Kubernetes Service. The Kubernetes based container is better than the Docker platform so it was not on the top. Microsoft Azure Kubernetes Service and OKE seem to have similar performance so OKE came up on the ranking since we already use OCI. Google Kubernetes Engine is still in the mix and we may use it at a later stage and replace OKE.