The use case we're interested in:

1. Users want to import VMs without having access from Openshift to the source system the VMs were running on.
2. The VMs are exported to a place that is accessible to Openshift, optionally accessible to the source system at the same time.

Assumptions:

1. The VMs are exported to OVAs, either folders or tar files, that conform the OVF specification standard
2. The OVAs are produced by vSphere or are vSphere-compatible and virt-v2v supports their conversion
3. The VMs are placed on an NFS share
4. The OVAs were exported from the same cluster, or the identifiers of the networks are unique

Out of scope:

1. Uploading OVAs from the local machine (the OVAs reside on a place that can be mounted to a pod)
2. Support sources that are different from NFS share
3. Support OVAs that were exported from oVirt/RHV