In order to add significant new features to KubeVirt you have to create VEP to get the community to agree to the design of the new feature. This design has to be approved by the appropriate sig before starting work on the implementation.

In this case we should note the following important details:

• This will require a new field where we can define the CCLM network. This is NOT the same as the cluster LM network.
• No changes will happen to the in cluster LM network (it can even be the pod network).
• The synchronization controller will now have a dual role.
  a. If a CCLM network is defined, it will act like a proxy between the incluster LM network and the CCLM network
  b. Still act as the synchronization controller between clusters, but also act as the synchronization controller between namespaces. Based on which one it might need to use a different network.
• The synchronizationAddresses logic will need to be reworked to properly pick the appropriate network based on which ones are defined.
• We need to decide if we want the proxy to terminate the SSL connections or not. This might make exchanging CAs between clusters obsolete. We would only have to exchange CAs on the CCLM network.

I envision something that looks like this