Working tightly with vRealize Operations, DRS will be triggered to make sure resource is available before a VM become demanding. A scenario will be e.g. A VM typically become busy during month end closing, DRS will move certain VM out knowing it will need more resources and making sure there won't encounter a resource contention when this happens. In the past, DRS alone will only be perform when such contention happens.
With hardware vendor working to reflect meaningful sensor notification that a possible of a hardware failure e.g. DIMM Slot, vSphere HA can leverage DRS to vMotion out VM from a hardware box before a kernel panic.
vSphere HA Orchestrated Restart
Just like what you can do with Site Recovery Manager (SRM), vSphere HA Orchestrated Restart now allows you to choose the dependency of VMs especially for multi-tier applications with VM-to-VM rules during a host failure that trigger vSphere HA.
FT network now support multi-nic setup just like how multi-nic vMotion will work. This allows additional bandwidth to be possible.