As an OpenShift admin, I want Pods that request vendor.com/accelerator to be allocated by a DRA driver so that I can standardize on DRA (claims, policies, and lifecycle) without forcing app teams to rewrite their Pod specs.

https://github.com/kubernetes/enhancements/issues/5004

Why this matters (problem → outcome)

• Today: Extended resources (e.g., GPUs) are usually exposed via device plugins; DRA is a separate path with its own objects and workflow.

• Pain: Two parallel models to manage; hard migrations; inconsistent policy/control.

• Outcome: Unify on DRA without breaking existing Pod specs that request vendor.com/*. Pods can keep using extended-resource fields, while a DRA driver performs the actual allocation.