Andre Leibovici from myvirtualcloud.net posted the new maximums and limits for Horizon View 5.2. With the new version of View, VMware has increased the hosts per cluster from 8 servers to 32 servers using a VMFS datastore. This was a limitation for both VMware View linked clones and with VMware vCloud Director fast provisioning. The limitation didn't exist for NFS storage, but there weren't many people that deployed VMware View on NFS storage.
This should affect the overall View pod design, not drastically, but we should see a reduction in the number of clusters from 8 for 8,000 desktops to 4 for 8,000 desktops. Why does this matter? In my previous role as an IT manager, it all came down to the bottom-line. How can I save capital expense? By eliminating the amount of clusters and going with 16 servers per View block, we are able to reduce the N+1 requirement by 4 physical servers. 4 servers typically cost me $25,000.00 in a scale-up model. That is $100,000.00 in infrastructure savings, cha-ching!
Lets look at it from a density stand point, if you were getting 128 virtual desktops per host (8 instances per core/16 cores per host), you gain 128 extra virtual desktops per cluster which amounts to 512 virtual desktops in the View pod scenario depicted below.
Not only does this save money, but by reducing clusters it makes the environment easier to manage. With a 10,000 desktop View pod deployment, you were typically working with 10 clusters, which should be reduced to 5 clusters with Horizon View 5.2.