We restored close to 40TB of SQL Data. It took almost 100 hours to do the restore. The backup base image is on NetApp All Flash array. All the datastores for the VMs are in this array. So I would tend to think that application restore would be a lot faster unless application restores are done through the IP network (as opposed to Direct SAN). I monitored the network interface of the destination VM and saw the traffic coming into this interface.
What is really the data flow for a SQL application restore?