I recently met with an organization that uses backup tapes as archive.  This is not an unusual practice, though every time I speak with someone doing it, they always express concern regarding recoverability of data down the road.  So this particular organization had to keep some data indefinitely.  In their Iron Mountain archive they have every type of tape format and probably a few backup application formats too.  Here is the crux of the problem.

* When you write data using a backup application, it puts data into into own proprietary format.  This means that you need to have the application in order to read the data on the tape.  What if you switch applications?  You might have to keep multiple backup applications or versions in order to restore the data if needed.

* Tape media continues to evolve; every couple of years a new format comes out.  Today it is LTO5 and it is different than the previous formats.  Not even 6 years ago, many organizations were beginning to move from DLT to LTO, some were even using AIT.  Now you have to keep all the different types of drives to support the media formats.  Otherwise, you have to migrate the data from older tape media to newer, and this migration could be happening as frequently as every three to five years.

* there is no guarantee that the data stored on tape will be accessible.  Tape media has a very long life span – if it is properly maintained.  It needs to be tuned periodically and it is a good idea to check that the data on tape has maintained its integrity.  A tape cartridge sitting on a shelf, even if the environment is appropriate, is not an ideal place for it over time.

* If data has to be retained for a long period of time, the more data is put off to tape for an indefinite period of time, the bigger the backup system’s catalog can get, which may cause reliability issues.  Making your backup system catalog bloated in order to retain data for a long period of time is not the best use of your investment in the data protection system.

These are the same concerns this organization expressed.  The other side of the issue though is that tape is significantly less expensive per TB than disk, especially if you have to store data for more than a few years.  As a result, organizations want to continue using tape.  So the question is, how can tape be used as an archive media without depending on the backup system and allowing seamless migration of data to newer tape format once they are implemented.  The answer is not so simple.

First, consider setting up an archive outside of the backup infrastructure.  Assign some tape resources that belong to the archive.

Second, invest in a good tape file system that has been designed to manage data on tape over a long period of time.  These systems present a CIFS or a NFS share to the application while storing data on tape in the background.

Third, identify a way by which data is moved from the primary storage environment into the tape archive.  This will depend on how much data you need to move and how frequently, whether the data is going to be accessed periodically or never, and what level of security is required.

Finally, separate the data protection tasks from the archiving tasks and understand what you are archiving, for how long, and why.  An archive is not really a backup, it is a way to preserve data for future use.  Data protection and backups in general are to protect live data from a potential corruption, deletion, or disaster.

Making the switch is a difficult process psychologically.  It might be best to take it in phases, as long as all parties understand what is at stake and where do they need to be.

Advertisements