Architecture
Suggest changes
PDFs
This solution is not dependent on specific hardware. The solution is compatible with any NetApp physical storage appliance, software-defined instance, or cloud service, that is supported by NetApp Trident. Examples include a NetApp AFF storage system, Amazon FSx ONTAP, Azure NetApp Files, Google Cloud NetApp Volumes, or a NetApp Cloud Volumes ONTAP instance. Additionally, the solution can be implemented on any Kubernetes cluster as long as the Kubernetes version used is supported by NetApp Trident and the other solution components that are being implemented. For a list of Kubernetes versions that are supported by Trident, see the Trident documentation. See the following tables for details on the environments that were used to validate the various components of this solution.
Apache Airflow Validation Environment
| Software Component | Version |
|---|---|
Apache Airflow |
2.0.1, deployed via Apache Airflow Helm chart 8.0.8 |
Kubernetes |
1.18 |
NetApp Trident |
21.01 |
JupyterHub Validation Environment
| Software Component | Version |
|---|---|
JupyterHub |
4.1.5, deployed via JupyterHub Helm chart 3.3.7 |
Kubernetes |
1.29 |
NetApp Trident |
24.02 |
MLflow Validation Environment
| Software Component | Version |
|---|---|
MLflow |
2.14.1, deployed via MLflow Helm chart 1.4.12 |
Kubernetes |
1.29 |
NetApp Trident |
24.02 |
Kubeflow Validation Environment
| Software Component | Version |
|---|---|
Kubeflow |
1.7, deployed via deployKF 0.1.1 |
Kubernetes |
1.26 |
NetApp Trident |
23.07 |
Support
NetApp does not offer enterprise support for Apache Airflow, JupyterHub, MLflow, Kubeflow, or Kubernetes. If you are interested in a fully supported MLOps platform, contact NetApp about fully supported MLOps solutions that NetApp offers jointly with partners.