There are four supported options to choose from for the deployment of virtual distributed switches to support the management and VI workload domains. The options provide a way to segment traffic based on network type based on the number of NICs configured to support the Cloud Foundation on VxRail instance.
If four Ethernet ports per node are configured to support Cloud Foundation on VxRail networking, then you can choose to either have a single virtual distributed switch be deployed for both VxRail networking and NSX-T networking, or deploy separate virtual distributed switches and move the resource-intense networks to a separate virtual distributed switch.
With a second virtual distributed switch, the vSAN network, which is the most resource-intense network required by Cloud Foundation on VxRail, can be isolated on its own virtual distributed switch with its own dedicated uplinks. Additionally, the vMotion network can either remain on the first virtual distributed switch and share bandwidth with the management and NSX-T networks, or migrate over the second virtual distributed switch.
For Cloud Foundation on VxRail workloads with very high resource requirements, you can elect to reserve eight Ethernet ports per node, and configure four uplinks on each virtual distributed switch.
If six Ethernet ports per node are configured to support Cloud Foundation on VxRail networking, then you can choose to deploy as many as three virtual distributed switches to support VxRail networking and NSX-T networking.
If you elect to deploy two virtual distributed switches, then the resource-intense networks can be moved to the second virtual distributed switch, and two uplinks on the first virtual distributed switch can be reserved for NSX-T traffic. Alternatively, you can choose to deploy a third virtual distributed switch to isolate the NSX-T traffic.