When dealing with networking, we tend to arrive at the same solutions for high availability for both NFS and iSCSI. We must configure redundant links and redundant switches for both NFS and iSCSI to survive a single link or switch failure.
The previous example has a redundant pair of switches configured to protect against a switch failure. We have links going from all ESXi Hosts connecting to two separate switches. Each NFS or iSCSI storage array has a link to each switch, thus forming a redundant pair.
Switching in a VMware vSphere ESXi environment doesn't vary significantly from other highly available environments. We want to configure a redundant pair of switches to handle any switch failures that may occur and to handle any single link, NIC, or port failure that can occur. Pairs of redundant switches should be built such that they can share traffic between them. ESXi may present VMs on different NICs and (in some cases) will need to send traffic between...