We have a main network of 10.1.x.x for all the switches/routers (1.x), servers (10.x), desktops (30.x), and and a private network for the SANs to communication with the virtual environment (221.x and 222.x). The private network consists of 2 switches, with just the 2 SANs and 3 host VM servers plugged into them, and the switches are redundant, with redundant iSCSI paths set up in Vmware. The vMotion settings are set up on this private network, not on the main network. We have 3 host servers, each with dual 8-core procs and 128GB of ram, and all 3 run about 1% - 5% cpu utilization and 10%-15% memory utilization tops, so the environment is way overbuilt, but it's built for the future.
We put #1 into maintenance mode, which set up an automated process to migrate all VM's off of it, and after almost a day it finally succeeded in moving all VM's to the other 2 hosts. We rebooted the host, and now it's working fine, as it always has.