We need to create a VEP for the new API that allows us to describe the two different types of LM networks. We will have

• in-cluster LM network. This is only accessible inside the cluster, and would be the 'normal' LM network as defined in the KubeVirt CR.
• cross-cluster LM network. This network provides connectivity between clusters, this can be an L2 or L3 network, as long as the proxies can communicate it should work.

We need a new optional API field that allows us to configure the cross cluster live migration network, which in turn will cause this to be secondary network in the synchronization-controller/proxy pod. So we have a potential for 3 networks in this pod:

• CCLM network
• in-cluster LM network
• Pod network

Only of 2 of them will be actively used at once, it will be either the in-cluster LM network or the pod network together with the CCLM network. Then we need to describe how the proxies will work, and how the multiplexing of the connections will work (proxy IP/port combinations created dynamically). Note that any live migration that stays inside the same cluster, will not change behavior at all with this. So:

• compute live migration
• storage live migration
• cross namespace live migration

Will not be affected.