The rationale behind this is simple. Tape provides the ability to produce low-cost, off-site point-in-time copies of your vital data. If all other measures fail or a wide-scale disaster destroys them, you can recover your data from tape.
The main drawbacks to tape backup systems are that they require a lot of oversight and that your data will only be current to the last tape backup. Oversight means maintaining the tape systems (e.g., cleaning, repair, etc.) as well as making sure someone changes and stores tapes.
In addition, you should consider contracting with a storage company to house the tapes off-site. Someone will need to manage this relationship as well.
Generally speaking, someone already on staff can handle these duties, so it's rare that you'll need to hire someone to manage the tape systems. However, larger organizations might have employees dedicated to just tape and other recovery systems, so it can also incur overhead in the form of staff expenses.
Recovery point objectives (RPOs) are also a concern. Should restoration from tape be necessary, you'll only be able to recover data from the point when you last copied to tape.
Using tape systems in conjunction with other forms of DR technologies (e.g., snapshots, replication, etc.) helps minimize RPO. However, these technologies can be out of the financial reach of smaller companies.
If so, your choices are few: Either back up to tape more often (which can mean higher expenses and more management), or be ready to deal with the loss of data since the last successful backup. If your organization's RPO allows the loss of 24 hours or more of data, tape backups alone are a suitable solution. But any smaller RPO requires using other technology in conjunction with tape.
Tape also requires that you test the system—a requirement of any DR system, but one that companies often overlook with tape systems. Someone must make sure that the data you think you're committing to tape is actually there.
This includes tasks such as making sure that backups complete successfully, that you don't miss open files, and that you have sufficient tape. Most tape systems on the market send e-mail alerts if the job completes successfully or if there's an error, so this can be as easy as checking the appropriate e-mail account for the results each day.
You also need to perform test restores on a regular basis. This consists of taking a tape backup, playing back some of the data stored on that tape to a standby server, and examining the data. The process typically restores flat files and opens them with the appropriate software. This ensures that you can restore the data to the real servers if you have to.
Despite the overhead and checking, tape is still a vital part of a DR strategy. Power failures, WAN link failures, virus attacks, and other problems can create disasters that can take out both your production facility and any redundant systems.
Recent outbreaks of some rather destructive virus attacks are a prime example. Sometimes, good tape backups are your only recourse when all else fails—don't leave them out of your DR plan.
Mike Talon is an IT consultant and freelance journalist who has worked for both traditional businesses and dot-com startups.