After battling some startup and infant mortality problems with my Drobo, I’m trying to think through all of the various issues involved in long-term, i.e., life-long storage, of my photographic images and other personal data. This note is intended to share my thinking and invite comment as to how my long-term data preservation strategy could be improved.

As background information, I currently have 246 GB of digital photos in 32,964 files and 545 folders stored on my Drobo, which will be the primary storage for those files. I also have 1.58 GB of Lightroom catalog data backed up to the Drobo, but I will continue to use the hard disk in my laptop for primary storage for those files. Another 200 GB are consumed by other backups, the contents of product CDs, and private data like tax returns.

Most of the digital photos have been taken in the last two years, and as the pixel density and frame rate have increased as camera models have improved, the rate of accumulation is increasing, to 150 or even 200 GB per year (unless I become more ruthless in culling the outtakes.)

Because of the increasing volume of stored digital data, I am planning to use my first “home” Drobo as my primary storage mechanism, and then use a second “office” Drobo for offsite backup. That way I can backup my primary data from my “home” Drobo to my “office” Drobo, and at the same time back up my office files to my home Drobo. That way, I am reasonably well protected against a catastrophic failure of one Drobo, as well as fire, theft, earthquake, etc. In an ideal world, these two, mirrored, sites would in two different states, or even different countries.

In considering how best to preserve and protect all of this information for many decades, it is important to recognize that existing disk drive technology may or may not be ‘”long-life,” depending on how that may be defined, but it is certainly not “life-long.” Hard drives WILL crash or experience serious I/O errors eventually – that is an absolute fact-of-life. Perhaps even more important, drive enclosures, power supplies and USB interfaces WILL fail, either sooner, as in my case, or later. In some cases the failure could be catastrophic, to the point of melting all of the drives in the enclosure In addition, technologies WILL become obsolete and impossible to acquire or replace, and the companies that manufacture such devices WILL ”end-of-life” those devices that are no longer selling well enough to justify their continued production and maintenance. Finally, even the very best of companies often fail, or are acquired by someone else, or otherwise cease to exist.

This suggests that the strategy for long-term data protection needs to be thought out very carefully.At present, it does not appear that there are any practical “static” recording technologies that can be reasonably guaranteed to last for 50 years or more, with ink on paper being the sole exception. Although various accelerated life tests have been run on gold-plated CDs and DVD’s, the experience to date indicates that the claims for their longevity may be somewhat suspect, and most people aren’t prepared to keep their storage media in bags filled with argon or nitrogen, and stored in the freezer. In addition, because of the continuing advancement in storage technology and the risk of obsolescence, it is important to have a well-executed plan for continually migrating stored media from one storage mechanism to the next one. Human nature being what it is, the tendency will be to put off any migration until it is too late, at which point it will be a herculean task. This is bad enough for personal users, but it is much worse for governments and commercial enterprises – they can always find some way of deferring the time and expense to the next budget year, or to the next administration, or never.

That’s the bad news. The good news, in a certain sense, is that at 200 GB per year, I will soon exceed the capacity of the two 1 TB drives (1 TB total, effectively), and will have to add another drive to each unit. Now, in a year’s time, I expect the price of 1 TB drives to have dropped significantly, and 1.5 or even 2 TB drives may be available.This suggests a strategy of continuous migration and improvement, as newer drive technologies become available. But how should this be done – by filling up the unit with lower cost drives and then replacing them one at a time, or by maxing out the capacity of individual drives, and/or holding a slot in reserve? (continued on page 2)