Lynn Scott's
Lepidoptera
Index

Image Storage, Archiving and Backup

Most of us have made a considerable investment in the fruits of our photography, either as professionals or as enthusiasts. In our case, we have tens of thousands of images, sorted and cataloged, forming an irreplaceable multi-year scientific record that could not be replaced if lost.

Many photographers keep their images stored on their computer's hard disk drive — shared with all your other computer programs, e-mail, data and operating system. If the hard disk fails, everything may be lost. Alternatively, some photographers back up their photos onto CDs or DVDs; however, this may not be practical for extremely large quantities of images, and it can be very time-consuming to find a single specific image from among a large number of CDs or DVDs of archived photos.

Our strategy is simple. All images, both old and new, are always available on the computer AND all of our images are backed up, all the time, on two separate and independent hard disk drives, a technique known as mirroring. This means that when (not "if") a hard disk fails, you will not have lost any images or other critical information. A bit of paranoia is also a good thing when it comes to preserving your life's work, and being mildly paranoid, we also have a second mirror array on a second computer, onto which we periodically copy all the data from our primary mirror array. And just in case of ultimate disaster, we periodically make DVD backups of the entire mirrored disk drive. You should keep your DVD backups safely off-premises.

There are proper hardware data backup solutions readily available, using an approach known as RAID, an acronym for a Redundant Array of Independent Disks. There are several RAID algorithms. I've outlined the most popular for non-corporate environments below:

RAID 1: This is basic mirroring where the same data is written simultaneously to pairs of hard disks (one can have a RAID 1 array with many pairs of hard disks, depending upon the amount of data you have). What we like about this arrangement is that when there has been a failure, that the still good hard disk retains all your data — restoring your RAID 1 array is simple — replace the failed hard disk with a new one, and the RAID controller automatically creates a new mirror of the data from the non-failed hard disk.

RAID 0+1: This approach makes use of sets of four hard disks, really two pairs of RAID 0 striped disks, said to be somewhat faster than basic RAID 1. (Note that RAID 0 by itself, striping-only, is a misnomer — it does not provide any redundancy whatsoever, the "R" in RAID — only a speed advantage.)

RAID 5: This approach uses three or more hard disks, and shares data, parity (an error checking method) and error correction information across all the disks. When a disk fails, the other disks have sufficient information to restore the information contained on the failed disk.

Being believers in simplicity, we prefer the benefits of RAID 1 (mirroring) over all others, especially as high-capacity hard disks are now so inexpensive, about $0.50 per gigabyte of storage for the very best Seagate or Hitachi SATA II (Serial ATA) drives!

RAID controllers are included on better quality, modern computer motherboards, or are available as inexpensive PCI or PCIX add-on cards for those with older computers. Also, it is worthwhile for increased performance to obtain controllers and hard disks, which make use of SATA II, 3.0 Ghz ports — the increase in speed over the older parallel ATA connections is significant.

Finally, we suggest that it is wise to use a separate physical hard disk drive for the computer's operating system, including all your programs, games and other items — keep your RAID disks for your valued image data alone. If a virus, worm or Trojan infects your computer, the operating system and its hard disk are often prime targets — and a hard disk compromised by a malignant piece of software won't be saved by a RAID array — hence the need for DVD-backups as well as RAID!


 

Page last modified 1 March 2008
Copyright © 2001-2008 D. Lynn Scott