This image shows how each HS21 blade in a VI3 cluster would be sharing multiple fibre channel LUNs, with each LUN providing a datastore to ESX for differing purposes (more detailed information in other documents). It also shows general virtual networking configuration and how the VMs on either host would connect to the virtual networks through the phsycial nics in the ESX hosts. The method we are using to manage NIC redundancy and failover in the guest VMs is to have two NICs teamed strictly for failover. The primary LAN connection to the VMs is "Virtual Network #1" in the image, which is accessed through vNIC2 in this image. This setup is to protect against issues I have seen in the past in which the OS software or other software installed on the VM may cause a malfunction of a particular network interface (not very likely). Other than this, the redundancy is provided by two pNICs providing connectivity to each vSwitch (remember each pNIC is going through a different Cisco ESM in the Bladecenter chassis).
