ODA Reference Implementation

Overview

The reference implementation comprises four key elements:

1. Infrastructure resources
2. Kubernetes environment
3. Canvas components
4. ODA components

Infrastructure is agnostic and can use any compatible infrastructure solution providing the underlying resources are capable of running a Kubernetes cluster. Examples include bare metal, virtual machine, Cloud IaaS or Cloud PaaS. A simple Click and Deploy Canvas is available to quickly get you started with local deployment using only one Docker command.

The Kubernetes environment is completely standard; however, it requires additional configuration for an ODA compliant stack. Achieve this by extending the cluster with custom resource definitions or CRDs. The CRDs and namespaces are automatically configured for you if you utilise the Click and Deploy Canvas scripts.

The Canvas components are critical to the reference implementation. In the same way, a Kubernetes cluster has critical system components such as etcd and control plane. An ODA deployment requires key services. The ODA Component controller is central to this, while an identity management solution and service mesh are also necessary as a bare minimum. If you're using the Click and Deploy Canvas, Keycloak is deployed for identity management and ISTIO for service mesh. Observability of the mesh is provided by Kiali and supporting applications.

The components come in many forms. An ODA Component may be a single microservice application or consist of multiple microservice applications. The ODA Component specifications encourage best-practice application design to support zero-downtime upgrades. As a base, all existing APIs can be deployed as single microservices. The wider vendor community is actively uplifting their applications to support ODA plug and play through the ODA Component Accelerator program.

Quick Links

• ODA Component Accelerator
• Click and Deploy Canvas