Storage virtualization abstracts the space in physical hardware to software-defined storage that is accessible from any device with an end-user interface. It lets multiple storage devices appear as a single volume, and simplifies the management of data without the need to predict long-term storage requirements and pay for additional capacity upfront.
Virtualization can be accomplished in one of two ways: host-based and network based. Host-based Virtualization (typically used in HCI systems or cloud storage) relies on software to control traffic. The host, or a hyperconverged system consisting of several hosts, presents virtual drives to guest computers with any configuration, whether they be virtual machines in an enterprise or PCs that access servers for file storage or servers that access data via the cloud. The host utilizes software that maps the logical addresses of each block of physical disk data into an offset within a virtual drive.
Network-based virtualization takes an alternative approach by shifting the complexity of the storage controller to a different layer than the virtualization hardware. This is often requiring additional components, such as a network switch, in order to take on the additional I/O load. However, it is able to lower costs while increasing performance.
The layer above virtualization hardware allows for backup and recovery to be carried out without the virtualization effecting it. Additionally, it makes it easier to troubleshoot issues remotely – either by internal IT teams or managed service providers – which can improve resolution speed. It also helps with scaling by removing the dependence between the location of the files that are accessed at the level of the file, and the place the location they are kept on physical disks. This can aid in optimizing the use of storage and server consolidation as well as executing non-disruptive file migrations.