SCSI pass-through disks provide a way to connect a Hyper-V virtual machine to physical storage (as opposed to relying...
on virtual hard disks). However, there are some limitations to using pass-through disks that you should be aware of.
Pass-through disks were a popular option prior to the release of Windows Server 2012. Back then, the virtual hard disk (VHD) format used by Hyper-V was limited to 2 TB, which was inadequate for some virtual machines (VMs). Pass-through disks gained popularity as a way of circumventing the 2 TB storage limit. In Windows Server 2012, however, Microsoft introduced the VHDX virtual hard disk format, which is not subject to the 2 TB limit.
In many cases, using VHDX-based virtual hard disks mitigates the need to use pass-through disks. If you plan on using pass-through disks, however, you need to understand that they are bound to the host server. As such, the use of pass-through disks can complicate live migrations (although live migrations are still possible if the pass-through disk is configured as a cluster disk with a dependency to the VM). Hyper-V is also unable to perform snapshots of pass-through disks.
Some administrators like to use SCSI pass-through disks because it makes SAN management a little bit easier. For those users who would rather connect VMs directly to SAN storage instead of creating virtual disks, I would recommend using virtual Fibre Channel (FC) or establishing an iSCSI connection from inside the VM.
The advantage to these two techniques is that they minimize the involvement of the host operating system. This helps the VM access storage more efficiently while also improving virtual machine portability. A VM with iSCSI connectivity can be moved freely as long as a network path to the storage exists. A VM that uses virtual FC can also be live migrated, as long as the target host contains the necessary host bus adapter hardware and is properly configured to support live migration.