I’ve said it before and I’ll say it again (and again and again): you can never have too many backups, but you’ll always have too few. Another computer cliché I like is that there are three golden rules of computing: 1. Backup; 2. Backup; and 3. Backup. I’m not going to give step-by-step instructions on the best backup strategy (there is no single best strategy) or tell you which is the best backup program, but I am going to cover the basic concepts of backups so you can start to formulate your own strategy and pick the best program for you.
An important distinction to start with: a backup is a second copy of a file. If you copy a file to another disk or location (CD, hard disk, iDisk, whatever), then delete the original, that’s an archive copy, not a backup. The whole point of backup is to be able to replace the primary copy of the file from the backup — if there’s only one copy on a CD somewhere and that CD is destroyed, that’s it, the file is lost.
So, at all times, you want at least two copies of your files. But the files themselves aren’t the only thing you want to back up — their arrangement on the disk or in your folder structure is also worth backing up. You also want to consider whether you want a full backup of the whole of your hard disk, or only a partial backup of certain areas.
A full backup has the advantage that it’s relatively easier to recover from a catastrophic loss of your hard disk back to a fully functioning state. Because everything is backed up, you don’t have to worry that you overlooked any files when backing up.
Partial backups are useful because they are quicker than full backups — if they have just the files which differ from a basic install of your operating system (such as your user files and settings), they may be adequate for getting back to a workable state, although it may take more time and effort to get back to that workable state.
A special type of partial backup, performed after a full backup, is an incremental backup. This sort of backup only backs up the files which have changed since the last backup. This is obviously generally much faster than a full backup, while allowing the recovery to a fully working state by comparing the full and incremental backups to determine the state of the hard disk as at the last backup.
Another consideration is how your backup is stored. Technically, just copying a file on your hard disk is a form of backup, but the risk of loss is not particularly reduced when you consider the situation of the hard disk catastrophically failing — both the original and the backup would then be lost.
So, ideally, you’ll want to store your backup on a different medium. This may be another hard disk (even one on a network), CDs/DVDs, tape drives, or online storage like iDisk or Amazon’s S3. A major consideration is the amount of storage you will need. If you want to do a full backup followed by incrementals, 512MB thumb drives are not going to be very useful. You may need to use DVDs, a hard disk, or tapes. If you only really need to backup a few smallish files, thumb drives may be ideal due to their ubiquity, size and compatibility.
A related issue for the choice of storage medium is whether you’ll be utilising rotating backups, where for, say, one week you backup to one disk or set of tapes, then swap over the following week to another disk or set of tapes. This spreads the risk of loss even further, and also allows you to store a longer history of files if you do daily incremental backups. Under this regime, if you erase the storage medium on Monday before doing a full backup and do incremental backups the rest of the week, then rotate to the second storage media the next week (and back again the week following), you’ll always have at least one or two weeks’ worth of backups.
Yet another consideration is the long-term viability of that type of storage medium. Obviously, if you want access to backed-up data far into the future, you’ll want a medium with a long shelf-life which you’re confident computers will be able to read for a long time. If you just want disaster recovery for your current working state (or maybe even for a month or two), this is much less of a consideration and may allow you to utilise cheaper storage. Similarly, if you’re utilising online storage services, you’ll need to assess whether those services are likely to be around in the time frame you want continued access to your data.
Rotating media also allows you to further reduce the risk of loss by utilising offsite backups. If you are rotating two sets of media, while one set is being backed up to, you can take the set not being used to a different location in case of a catastrophic loss of the premises where the original file and the current backup set are. By their very nature, online storage facilities like iDisk and S3 are offsite, and therefore present an attractive storage medium for backups. These online services are also backed up by the service providers.
I’ll mention here only one specific backup program, simply because it’s now built into Mac OS X. Leopard’s Time Machine is like a super-duper incremental backup. It performs a full backup initially, then performs incremental backups after that, but each individual backup looks like a full backup (using some special under-the-hood Unix voodoo), so you can restore your machine to its state at the time of any incremental backup. However, Time Machine is limited in it’s choice of storage media, it’s fully onsite in its default behaviour, and, being a newish technology, some have reported significant problems on their machines. Your mileage may vary, but Time Machine backups are better than no backups, and it can perform some nice tricks for recovery of data (even within the original program such as in Mail.app and iPhoto).
One final note — if you only have one backup copy of your data, it’s worth checking that the data is recoverable every now and then. It’s better to find out there’s a problem with your backup before the primary copy is lost rather than to have that sinking feeling when you realise you have no valid copies of that file you absolutely need to work on.
What’s your backup strategy? What programs and hardware/services do you use to protect your digital assets? Tell us about it in the Australian Macworld Forums.