When it comes to data backup, many SMBs feel they have to choose between secondary disk or tape storage. While both can be viable solutions, a better approach is to combine the two in one integrated backup system.
A combined system, called disk-to-disk-to-tape (D2D2T) backs up first to a secondary set of disks, usually low-cost SATA disks and then from the secondary disks to tape.
D2D2T combines the strengths of both tape and disk. Tape backup is increasingly seen as an archival medium, used to store information for months or years, while some form of disk backup is used for immediate recover of files which are damaged or accidentally deleted. As prices drop and options multiply, the combination of these two technologies becomes increasingly practical for SMBs.
D2D2T can be packaged in several different ways. Some offerings in the D2D2T space, like EMC Corp.'s Retrospect, are software only. The user supplies the tape system and disk array. Others, such as Arkeia come as complete appliances (but you can buy software, too). Other companies, such as Hewlett-Packard (HP) Co., offer combined packages with backup disk arrays and one or more tape drives, plus HP's software to tie it all together.
Functions and features of D2D2T
The most common use for disk-to-disk-to-tape is in environments where the amount of data to be backed up threatens to overwhelm the time available for backup. The time it takes to back up to disk is much less than the time it takes to back up the equivalent data to tape. Backing up to disk first takes the pressure off the backup window and the system can later transfer the data to disk without degrading the performance of the primary storage.
Data deduplication is another important feature in D2D2T systems. By weeding out duplicate blocks of data, deduplication can reduce the size of the data to be stored by as much as 20 to 1, greatly reducing the size of the data for long-term storage.
However, putting disk in the backup chain also gives advantages in reliability. Tape is can be unreliable because of the human variable. Tape requires more handling than disk, meaning there are more chances for operator error to mess up the backup. Adding disk to the mix removes the need for as much manipulation and handling of the tape, especially since most D2D2T operations back up to the tape less frequently (such as weekly instead of daily). Even if a data transfer to tape fails, the image is still preserved on the backup disks, making it available for another try.
Backup disks also make it easier to recover individual files or folders, by far the most common use for backups. Disk access is random and tape access is linear, so it is much faster to recover a file off backup disks than it is to get it off tape, even if the particular tape happens to be sitting in the drive when you need it.
But despite all its advantages, disk can be a poor medium for long-term backup and data archiving. There are the possibilities of disk failures, and it generally isn't portable to move the data offsite for additional protection.
Off-site storage is a must for disaster recovery (DR), and physically taking a tape off the premises is the most common method of accomplishing this. Usually tapes are made using a rotation system that keeps a current tape backup on hand and one or more copies at other secure locations.
Overall, disk-to-disk-to-tape offers the best of both worlds at a somewhat higher price. The backup disks, usually cheap SATA units, can store data at as much as 10 times the space of tape, which stretches the backup window. Meanwhile, sending the backup to tape gives a removable, long-term storage image that can be kept for years if necessary.
Keep in mind that D2D2T still requires some physical manipulation of the tape, especially if cartridges are removed from the tape unit to be stored offsite. However, there is much less manipulation than there is with a pure tape solution and the copy on the backup disks provides an extra element of safety.
About the author: Rick Cook specializes in writing about issues related to storage and storage management.