Storage encryption has historically proven unpopular because of the issues of managing the keys used in various encryption methods, and the system performance burden that encryption/decryption placed on systems hardware. These issues have been iteratively surmounted in a number of advances. Faster CPUs, disk controllers and host bus adapters (HBAs) and tough-to-break encryption now reduce the burden once placed on subsystem, disk, and device encryption.
via Enterprise Storage Encryption: An Administrator’s … – Input Output.
This is a pretty good article. Here’s one more blurb.
Subsystems are encrypted at the hardware level or at the device group level. Subsystems can also be encrypted by the use of third-party software packages designed specifically for this purpose. Subsystem encryption usually means that a single encryption key is used for the group, and that the private encryption key is stored in hardware or in system firmware. This type of protection means that a drive “stolen” or otherwise removed from the group of storage media can’t be decrypted, even by examining the disk using an identical system, as the identical system would be missing the key that unlocks the data by decrypting it.
This seems like a rather high level of security.