The three-pod design completely separates the functional blocks by using a Management pod, an Edge pod, and a Resource pod, each for their respective functions. The initial deployment of a three-pod design requires additional hardware when compared with a two-pod design, however each pod can scale up independently of the others. Regardless of the pod design used to create the NFVI, VNFs will perform in the same way.
The main differences between the two designs are:
In three-pod design, edge functions are no longer managed by vCloud Director and its tenants, in contrast to two-pod design. Instead, the CSP network administrator sets up all edge functions directly through the NSX Manager. This is described in detail in the Virtual Networking Design Using VMware NSX Manager section of this document.
Because in three-pod design edge functions and resource functions no longer coexist in the same pod, independent scaling plans can be created for both Edge and Resource pods. For example, adding clusters to the Resource pod no longer requires a potential VNF workload migration to provide room for growth of the edge functions.