Roadmap

Model Serving: Updates to the etcd and minio setups
Support for ODH Notebook Controller
ODH Dashboard Admin UI and Notebook Controller backend: Add UI support for admin configurations
Deployment of Data Science Pipelines based on Kubeflow Pipelines (Tekton)
ODH Re-architecture - Phase 1: Categorize existing components (as of ODH v1.3.0) into support Tiers for updates/deprecation and designate maintainers Tiered approach
- Tier 0 - ODH Core
- Tier 1 - Long Term community support with assigned MAINTAINERS and full test suite
- Tier 2 - Any components in the opendatahub-io-contrib org that are not included in the ODH Core deployment but maintain ODH integrations of interest to the community

Visit the release notes section for detailed information about the releases.

Tutorials and website cleanup
Dashboard updates: Admin - Bring your own notebooks+ Cluster Settings
Bring your own notebooks: Allow admins the flexibility to provide to users their own notebook images
Model Serving: Implementation of Model Mesh using Oauth
Community governance model
Kubeflow 1.5 support: Includes ability to install Kubeflow v1.5.0 components on OpenShift
Support for the integration of Service Mesh with Kubeflow v1.5 on OpenShift

JupyterHub Custom Notebooks User Interface improvement: Add User Interface elements for users to enter new custom notebook images
Kubeflow 1.3 on OCP 4.6/4.7/4.8: Enable Kubeflow 1.3 on OCP 4.6/4.7/4.8 by resolving all issues.
Kubeflow integration with Red Hat Service Mesh.: Replace Istio stack with Red Hat Serivce Mesh in Kubeflow
ODH and KF authentication architecture.: Architect an integrated authentication solution for ODH and KF.
KF 1.3.1 OCP Stack update.: This includes updating the Kubeflow 1.3.1 openshift stack to work on OCP 4.4+

Kubeflow 1.3: Update the Openshift Kubeflow distribution to Kubeflow version 1.3
Openshift Pipelines: Installation of Red Hat OpenShift Pipelines along with all of the required custom resources to enable a workflow supported by Tekton pipelines
Trino: Trino is a fast distributed SQL query engine that can integrate with multiple data sources such as S3, SQL databases, and NoSQL databases
JupyterHub: Updates include new Spawner user interface,ability to customize and specify JupyterHub PostgreSQL parameters and new notebook images with JupyterLab.
Open Data Hub operator: Updates include support for Operator Level 4 (Deep Insights), new Prometheus metrics collection, new Grafana Dashboard and new status field to reflect operator installation status.
Open Data Hub Dashboard: Integrated with Openshift Authentication to access the dashboard.

ODH Data Catalog: Migrate Open Data Hub Data Catalog component from Ansible to Kustomize.
Kubeflow KFServing: Enable KFServing in Open Data Hub.
Kubeflow Pipelines on Tekton: Enable Kubeflow Pipelines using Tekton in Open Data Hub.
Disconnected Deployment: Investigate and introduce ability to deploy ODH on disconnected OpenShift clusters.
JupyterHub Spawner UI Rearchitecture: Replace existing static HTML spawner UI with dynamic React base one which will allow for more customization and easier extensabiliy.

Kubeflow 1.2: Add OpenShift stack in Kubeflow 1.2 to achieve release sync between ODH and KF
Disconnected Deployment (Moved to 1.0): Investigate and introduce ability to deploy ODH on disconnected OpenShift clusters.
UBI based KF (Moved to 1.1+): Continuation of the "UBI based ODH" expanding to Kubeflow project and looking at what does it take to move Kubeflow components to UBI.
CI/CD: Continuation of the of the effort to design and create a complete CI/CD process.
Object Storage: Add an Object Storage tool based on Rook-Ceph
Enable Monitoring: Continuation of the effort to enable Prometheus enpoints in all Open Data Hub and Kubeflow components.
Open Data Hub Dashboard: Team is currently creating an Open Data Hub dashboard that will be the entry point to all installed components.
Notebooks to Pipelines with Elyra: Added a JupyterLab notebook image that includes Elyra. Elyra converts notebooks to Argo or Kubeflow pipelines.

ODH JupyterHub: Forked JupyterHub repos under Open Data Hub github repo for maintaining new changes. Added notebook images to Thoth Station for building. Added ability to launch JupyterLab images.
Notebooks to Pipelines with Elyra (Moved to 0.9): Added a JupyterLab notebook image that includes Elyra. Elyra converts notebooks to Argo or Kubeflow pipelines.
Mixing ODH & KF components- Distributed Training: Added Pytorch operator to work with Open Data Hub components.
Mixing ODH & KF components- Kubeflow Monitoring: Added monitoring to Kubeflow components by enabling monitoring to Argo and adding Prometheus and Grafana to Kubeflow installation.
CI/CD: Added more tests to Open Data Hub components including Kafka and Superset and enhanced JupyterHub testing by adding Selenium for web portal testing

KF 1.0 on OpenShift: The main goal of this initiative is to verify Kubeflow 1.0 works on OpenShift and fix the issues we find. Another goal is to document and ideally automate some of the verification process to start enabling the CI for KF on OpenSHift.
CI improvements: Extending tests for all components, enabling CI for the operator repository.
Mixing ODH & KF components (start): Proving users can mix ODH and KF components, compiling a prioritized list of components to be verified and fixed, proving on the first component (probably TF Job or Pytorch Job)
Add Object Storage Component (Moved to 0.9): Since ODH relies on S3 compatible object storage, add dependency on some minimal install of OpenShift Container Storage.
Convert Data Catalog to Kustomize (Moved to 1.0): Data Catalog is the last component missing the conversion to Kustomize.

Rebase ODH on KF operator: Use Kubeflow operator as a base for ODH.
Convert ODH components to Kustomize: Convert all ODH components to Kustomize to match Kubeflow deployment tooling.
Start CI for ODH: Investicate and kickstart ODH CI based on OpenShift CI
Move to Github: Since the goal of this release is to get closer to KF community, we need to move to Github
Add Apache Airflow: Add Apache Airflow as a component into Open Data Hub