Data Protection and Information Lifecycle Management

The most common backup and restore media is magnetic tape. Disk-based backup systems are quickly becoming an important tool in the system administrator's toolbox, but tape is still predominant.

Like the cassette tapes used for music, magnetic tape is a thin piece of plastic with magnetic media on its surface. Data is encoded on the tape by causing magnetic particles in the tape to align along the length of the tape. Analog and digital tapes differ in how they represent data on a tape, but both use magnetic particles embedded in a plastic binder as the method of storing information. When writing, the tape is moved past the tape head, and a magnetic field is applied that polarizes the magnetic particles. To read, tape is moved past the heads again, and the state of the magnetic field from the particles is detected. The read sensor in the head transforms the magnetic field into an electrical signal, which represents a bit of data.

Tape has several important features that make it desirable for backup. It is reliable; has high capacity; and is inexpensive, reusable, and removable. It has been around since before the dawn of the computer age and is a well-understood technology. Vendors know how to make reliable tapes, tape drives, and robotics. Failures, outside of those caused by overuse and poor handling, are small.

As discussed in Chapter 2, tape data storage capacities are high compared with those of other media, with the exception of hard disk drives. CD-ROM/RW, DVD-ROM/RW, and other media are orders of magnitude smaller in capacity than common tape drives. The cost per megabyte for the media is also quite inexpensive. A 160-gigabyte (uncompressed) cartridge a typical capacity for backup tape can be found for less than $50. That's roughly $3.20 per gigabyte.

Tapes can be overwritten many more times than can other media, with the exception of hard disks. CD-R and DVD-R can be "rewritten," but that means that only the sections that are unused can be written to. Portions of the CD or DVD previously written to are no longer available for more data. After a short while, the CD or DVD capacity drops to zero, and it cannot be written to again. CD-RW and DVD-RW can be completely rewritten, but only a fixed number of times.

What makes tape technology so valuable to backup and restore is that it can be removed from the drive and placed in another location. This is a foundation component of a good backup strategy. The backup does little good if it is destroyed in the same accident that ruined the primary storage. Disks cannot be easily removed, and copying data to another location requires an off-site facility housing a backup unit.

Several types of tape technology are available. Tape systems sold today use digital tape. Several formats exist, with the most common being DLT, SuperDLT (SDLT), EXO, AIT, SuperAIT, and LTO. All come in models that can be mounted internally, externally, and in a network.

In large-scale enterprise systems, it is usual to find tape drives mounted in a library. Tape libraries are medium to large units (often as big as an industrial refrigerator) with several tape drives, a series of slots to hold tapes, and sophisticated robotics to move tapes in and out of the drives. High-end units have high-speed Ultra320 SCSI and Fibre Channel interfaces, along with embedded software that can perform automatic failover without losing I/Os already in progress. Libraries are a good option if you have a large amount of data or many nodes that need to be backed up. The robotics eliminate much of the drudgery associated with changing tapes while reducing errors caused by poor tape handling.

When media management is the issue, rather than capacity or performance, an autoloader is a good option. Autoloaders have only one drive but can hold many tapes. Tapes can then be grouped into backup sets and rotated. A typical autoloader has a carousel (similar to the old-fashioned slide projectors) that moves the tape into place and loads it into the drive.

Both libraries and autoloaders are tools to automate tape management. Libraries have the added advantage of being able to handle a large amount of data and many backup jobs. Autoloaders are much less expensive.