Integrating the Tungsten Fabric SDN controller with Red Hat OpenShift

The challenge

Tungsten Fabric had to become part of OpenShift itself: installed with it, deployed as a core component and providing the networking for the rest of the installation. The integration had to meet Red Hat's requirements for an OpenShift CNI plugin for secure enterprise use, work with read-only CoreOS-based nodes, and follow the client's release cycle, custom SAFe/Agile methodology and rigorous security and corporate standards.

What was at risk:

• A networking integration held to Red Hat's CNI requirements, the client's release cycle and its security and corporate standards

What we did

• Provided a team of software and network engineers with a proven track record in similar projects, scaled to the client's needs.
• Analyzed end users' needs.
• Created an architecture, specifications and delivery plan for the Proof of Concept.
• Built the solution within the client's release cycle, developed with its custom SAFe/Agile methodology.
• Designed and implemented changes in CI pipelines.
• Validated the solution with end users in a demo deployment.
• Released the solution to the open-source community and supported it later.

How it works

Tungsten Fabric installs with OpenShift and runs as one of its core components. Four pieces make that work.

1. Operator Tungsten Fabric is integrated with the Kubernetes Operator Framework.
2. Installation Changes in the OpenShift install configuration file bring Tungsten Fabric into the OpenShift installation, where its networking deploys as a core component in parallel with other OpenShift components and provides networking for their installation.
3. CNI requirements The Tungsten Fabric Operator is aligned with all Red Hat requirements for the OpenShift CNI plugin, to ensure secure enterprise use.
4. CoreOS nodes The Tungsten Fabric architecture works with secure read-only CoreOS-based nodes, using native ignition configs attached to the OpenShift installation manifests.