File virtualization actually involves a variety of services, including mirroring, replication, snapshots, point-in-time...
(PIT) copies, file system aggregation, data movement, archiving, HSM stubbing (sometimes called ILM) and data abstraction across storage tiers. These features can be implemented as software running on off-the-shelf servers (a.k.a. appliances or tin-wrapped software).
Features can also be implemented on custom proprietary hardware that performs one or more of the features noted above. Some implementations are tightly integrated with a specific vendor's technology, while some may only support NFS or CIFS or Windows file sharing. Ultimately, there is no single "right" way to implement file virtualization. There are many possible means of implementing the technology and its related features. The real issue is in meeting the organization's unique needs.
Go to the beginning of the File Virtualization FAQ Guide.