Backing up your data

by Glenn Koenig, 4 September 2004

These remarks are intended for those using computers at home or a small office with under 10 machines, typically.

Today, the best way to back up a hard drive is with another hard drive. The capabilities of "removable media" such as diskettes (including zip, etc. disks), CDs, DVDs, and tapes have not kept up with advances in hard drive technology. Hard drives can now hold much more data, operate much faster, and are much more secure from damage than the other four. Even the latest technology in removable media, DVDs, are especially vulnerable to damage from handling, more so than CDs. Diskettes and Zip cartridges are vulerable to heat, moisture, magnetic fields, etc. They and CDs have much smaller capacities than today's hard drives (under 1%, typically) meaning that you'd need to burn over 100 CDs to back up a 100 Gigabyte hard drive. Now, clearly, most people don't have a 100 GB drive in their computer nearly full with data that needs backing up, but the hassle of burning even a fraction of that number of CDs, or recording to diskettes, etc. still means that backing up is a distasteful task at best, and therefore often put off for so long that you are in effect not backed up.

One option is to purchase a Firewire (IEEE 1392 standard) external hard drive and simply plug it into whichever computer you wish to back up and copy your files onto it. I usually create a folder on the external hard drive named with the date and from which computer the data was backed up if you have more than one. You might put a simple calendar on the wall and mark down when you did it last on each computer at your location. How often you do it depends on how much original material you generate per unit of time. I recommmend the 'stomach test' - imagine losing all the data you've created during a period of time and think of how your stomach will ache if that were to happen. Then back up slightly more often than that.

There are many hazards to data; the most frequently cited one is computer hardware or software failure. True, this can be caused by power line 'glitches,' viruses, or other computer hardware or system software failure. Then there is fire and theft. These are both very low likelihood. Two more that are in fact more common than fire or theft are operator error and water. I know of at least two major disasters caused by water where data was lost. In both cases, there was no hurricane or major storm flooding, it was a broken pipe on an upper floor that caused water to cascade down through the ceilings, blanketing everything in sight with wet plaster, debris, etc. Stories about fires make it into the paper with headlines & photos. Water damage almost never does, probably because it almost never causes injuries or death, but is in fact somewhat more likely than a fire. Most places these days at least have smoke detectors if not an entire fire alarm system, but water alarms are still rare.

As for operator error, I have seen more cases of this than I can cite here. But one that is of particular importance is the 'reverse backup' problem. That is, when backing up, make sure you are copying the correct data to the correct place. Double check to make sure you are not obliterating the new version of something with an older copy. If you see a warning on your screen (in a small box, sometimes called an "alert"), read it carefully and make sure you know what you're copying where. External hard drives with distinctive icons and names that clearly identify them as the backup repository can help.

The best way to protect against any site-wide hazard (fire, theft, storms, water leaks, etc.) is off site backup. Yes, it cost another $150 or so for a second backup hard drive, but if your data is that important, there is no substitute for copying everything to such a drive, then taking it somewhere else, such as the house of a relative or friend, for safe keeping. If you are backing up computers in an office, you probably want to have a policy about handling of sensitive (i.e. private) data off premises. someone else, you should probably check about regulations concerning removing data that is property of the business involved, of course. How often you do this is up to you, but one good rule of thumb is to stop and think every month or so whether the stuff on that drive would be worth it if you lost your live data. If not, it's probably time to refresh your off site copy.

Here is one method to use. First, do your usual backups on your regular back up hard drive. Then take this drive to your off site location and bring the off-site drive with the older backup on it back to your regular site for use as your regular back up drive. That way, your off site location always has at least one copy (as opposed to bringing your off site drive into your primary location). Also, both your off site and on site hard drives get exercised a similar amount. This way, your data is never all in one place at the same time.