When SnapVault Primary and Secondary nodes are configured with the same storage system, either within the head or across a HA pair, what path does data traverse between partners? Is it internal within the storage system or are actual front end port interfaces used?