accommodate multinode HPC allocation rather than single-node VM instantiation

[ssenator]

To run a full model of an HPC cluster on an HPC cluster, we need to be able to create VMs on multiple nodes. However, the VM's on each node must be able to communicate with each other. Changes necessary:

1. Change private network to a bridged interface, of the public host's NIC
2. Allocate IP addresses systematically based on which node the VM will be created
3. Create the infrastructure nodes on the first node allocated
4. Create the compute nodes on other nodes in the allocation
5. Consider removing the vcgate node and/or changing it to be a vcrtr heartbeat verification node
6. Create additional verification tests, as required, mostly in prereq