Some of the arguments in favor of disk-based backup are based on erroneous assumptions.
By Dawn S. Wortman
Magnetic tape technology debuted as a data storage medium 50 years ago, and ever since then alternative technologies claiming to be a better solution than tape have appeared.
The refrain "tape is dead" recurs every few years as new technologies claim superiority. The latest challenger to tape is disk-to-disk backup using low-cost ATA arrays.
Disk-based backup has the potential to solve some unique backup issues, and when used in conjunction with tape-based backup systems, it can provide an effective data-protection strategy. While some proponents of disk-to-disk backup have presented it as a total backup solution, it is not designed to replace tape as part of an overall backup and data-protection solution.
Many of the arguments in favor of disk-to-disk backup are based on myths.
Disk-based backup is less expensive than tape
The common price comparison offered by disk-to-disk backup advocates is to examine the cost of equivalent capacity ATA disk drives to tape drives. This comparison is invalid and useless for end users. Cost comparisons should be based on the total cost of the backup solution, where tape still holds a significant advantage over disk-based backup. Additionally, disk-based backup technology does not match the scalability of tape and can become cost-prohibitive in high-capacity environments.
Disk-to-disk backup is faster than tape
Another argument by proponents of disk-to-disk backup centers on improved performance in data backup and recovery, and in restoration of applications to fully operational status.
Disk arrays can reduce backup windows, leveraging the higher data-transfer rates of hard disk drives compared to tape. However, there are potential pitfalls to reliable data recovery with disk arrays that can limit a company's ability to safely restore critical data.
In many cases, disk-to-disk backups consist simply of a block image of data, making it impossible to restore a single file. Instead, to recover one lost file requires a total restore of the entire backup volume, turning a potentially fast and easy file recovery operation into a restoration that can take hours.
Alternatively, disk arrays are used primarily for point-in-time snapshots that track only files that have been changed since the last full backup operation. To be useful in a data-recovery operation, these snapshots must be coupled with a full volume backup of the original data and re-applied in sequential order.
The very nature of snapshots, which are typically taken of a data set containing open files and active databases, also impacts their viability as a data-protection option. The result is often unsuccessful recovery and/or unreliable data.
Disk-based backup eliminates the need for tape
A central message from the disk-to-disk backup camp is that tape becomes unnecessary once disk-based backup is deployed. This is not the case because tape is a key technology in a variety of applications beyond the backup-and-restore requirements targeted by disk-to-disk backup.
Even within the backup market, disk-based backup has limitations. For example, lack of media removability may be a drawback for some organizations. With disk-based backup, there is no simple and secure method of transporting media to a secure off-site location to ensure against a catastrophic data-loss incident.
In addition, the fixed capacity of a disk array demands that at some point data must be deleted or archived to support new backup sessions, or users have to buy additional disk arrays.
These limitations may come into play with new government regulations for financial data that mandate that redundant copies of specified data must be stored at least 300 miles from the primary storage facility, or the new Health Insurance Portability and Accountability Act (HIPAA) that requires specific medical records and documentation to be retained for a minimum of six years.
Tape is a cost-effective medium to meet long-term data archiving requirements.
Enhanced backup solution
The most utility a disk array can provide in a backup environment is when it's coupled with a tape library as part of an "enhanced backup" solution.
In a disk-to-disk-to-tape configuration, the disk array functions not as a backup device, but as a front-end cache with the tape library as the backup target device.
As with many emerging technologies, hype surrounds disk-to-disk backup. It's unlikely that disk-based backup will replace tape in backup or in any other applications. In fact, the most-effective use of disk-based backup can only be achieved when it is paired with tape backup.
Disk-to-disk backup technologies may find a niche in enterprise storage environments. But they do not offer the cost advantages, scalability, and secure off-site storage capability provided by tape-based backup.
Dawn S. Wortman is senior marketing manager at Maxell Corporation of America, in Fair Lawn, NJ.