Recent popular discussions on the TidBITS Talk list have orbited around the issue of backing up data – what’s the best media to use, how often should one back up, what software works well, if backup devices should be built into computers, and how iMac users will back up their data. These are all important questions, and the answers affect literally every computer user today, irrespective of platform.
In the first part of this article, I’ll examine backup strategies and some topics to consider when formulating a backup plan. The next part of this article will discuss specific products you can use to back up your data.
Importance of Backing Up — Most computer users don’t back up their data. It’s easy for people – especially novices – to believe that computers are infallible. You press a key, and they just work. (Or, for those in large organizations, if something goes wrong, you call the help desk and someone else fixes the problem.)
But those of us who have been around the block a few times know that’s far from the truth. Files are deleted inadvertently, PowerBooks are dropped, hard disks fail, drinks are spilled, and of course, offices are burgled and houses burn down. Digital data stored on a disk is anything but secure, and pretending otherwise invites disaster. As an ad for Retrospect, the most popular Macintosh backup program, once said: "There are two types of people. Those who have lost data and those who will." That’s truth in advertising.
Backup Strategies — If we agree that everyone should back up their data, the next question is what they should back up. There are essentially two backup strategies, with a continuum of possibilities in between. One strategy says that when your hard disk fails, you want to be up and running as quickly as possible using a complete backup that’s as recent as possible. Call this the Complete Backup strategy. The other strategy assumes only your data files are important, since you can always reload applications from master disks or download freeware and shareware applications. Call this the Minimal Backup strategy.
The two strategies require roughly the same amount of time. If you subscribe to the Complete Backup strategy, you spend more time dealing with your backup system on a regular basis, although automating the process makes it easier. You must feed disks or tapes to your backup device and verify that everything is working. That takes a fair amount of time up front, but recovering from a dead hard disk takes only a little more time than that required to read back your files.
In contrast, adherents of the Minimal Backup strategy spend less time up front – just the occasional copying of a file to floppy, perhaps – but may require days or even weeks to restore a system to full working order. The Minimal Backup strategy puts the burden on you to backup the appropriate files. Will you remember to back up every important file you modify or create? If not, you may be forced to rehash days of work. Also, you must specifically back up preferences and other out-of-the-way files: Remember that you’ve spent time configuring your applications (think keyboard shortcuts in Microsoft Word); setting up utilities and extensions; and creating scripts for programs like QuicKeys, OneClick, and KeyQuencer. What’s more, finding and downloading new copies of freeware and shareware takes time and can prove difficult in the case of incremental updates to commercial programs or system software. Even locating serial numbers can take a surprising amount of time.
I’m strongly in favor of the Complete Backup strategy. I back up our internal machines to DAT tape using Retrospect every night (or in the case of PowerBooks and my infrequently used PC, whenever they turn on). I also use a pair of 2 GB hard disks in a RAID setup on my main machine, such that if one dies, the other will contain an exact duplicate of the data and instantly take over (in theory – it’s hard to test). I use the RAID setup because I hate losing the important email that arrives between the time my Mac backs up and the time some sort of data loss occurs.
Why do I do this? Call me paranoid, but I can’t guarantee I’ll have time to spare when something goes wrong: Murphy’s Law being what it is, it seems more likely that I’ll need to start working as quickly as possible. You must decide how important your work is to you; that decision affects the type of hardware and software you choose, plus your overall backup strategy.
My primary weakness is that I don’t have a solid offsite backup strategy. Geoff Duncan and I periodically trade DAT tapes, but if my house burned down, I’d be weeks or months out of date.
Backup Considerations — You must keep a number of issues in mind when forming a coherent backup strategy.
Historical vs. working backups: Some people rely on working backups – recent exact duplicates of their hard disks (on another disk or Jaz cartridge, say). If this is what you do or are considering, think carefully. What happens if an important file is irretrievably corrupted and you don’t notice immediately? With a working backup, the backup probably contains the corrupted file. If you use a historical backup – one that doesn’t erase previous versions of files – you can go back to the most recent version of the file that’s not corrupted. Of course, historical backups require more backup media, which increases costs.
Double-duty storage devices: Many people like backing up their hard disks to Jaz cartridges, for example, because they can use the Jaz drive for other things as well. I did this years ago with a 44 MB SyQuest drive. Although this strategy works, I don’t recommend it for two reasons. First, there’s always the temptation to use the backup cartridges for normal storage if you need some space quickly. At that point, the backup cartridge is no longer just a backup, but also contains unique data. Second, using the Jaz drive in other ways probably indicates you have other cartridges containing unique data. How do you intend to back up that data, or is it essentially worthless to you? I went through all this with my SyQuest, and I found that a dedicated DAT drive for backups doesn’t raise these thorny problems.
Media capacity: When thinking about backup devices, think about the amount of data the device or its storage media can hold. I use a DAT drive that can hold about 2.6 GB on a single 90 meter tape. However, I have about 10 GB of online storage between all of my machines. The data is not all in use, and many files are redundant (Retrospect doesn’t back up multiple copies), but a single tape won’t quite hold everything, which forces me to use multiple tape sets. The smaller the media capacity, the more media you’ll need, which drives up costs. Of course, the smaller the media capacity, the harder it will be to set up an unattended backup system. In an ideal world, you could do a full backup to a single tape or cartridge, then do incremental backups to another tape or disk for several months before needing to add additional members to that backup set.
Backup device cost: Most people worry about the cost of a backup device, whether it’s a SyJet, a DAT drive, or whatever. In my mind, and in part because I use my Macintoshes for business purposes, that cost isn’t particularly relevant, since it’s a one-time cost and the longer you put it off, the more likely you are to lose data worth far more than the backup device. Costs can range from about $150 for a Zip drive to $750 for a fast and capacious DAT drive. Don’t skimp on the device or buy something weird because it’s cheap – you don’t want the device to be a weak link.
Backup device and media format longevity: When thinking about backup devices, think conformity. You don’t want to have years of backups and archives in a format that can’t be easily accessed if your drive dies. For instance, it’s reportedly becoming difficult to obtain replacement or repaired mechanisms for SyQuest 270 MB drives from manufacturers, and there are numerous older media formats that have already gone the way of the dodo.
Backup media cost: A 90 meter DAT tape holds about 2.6 GB and costs between $5 and $10. In contrast, a 2 GB Jaz disk costs about $150. If Johnny backs up 10 GB to DAT and Sally backs up 10 GB to Jaz, who will pay more money, even taking the cost of the backup device into account? (The answer is left as an exercise for the reader.) Seriously: security costs money. Make sure to take both backup device and media cost into account.
Media reliability: Not all backup media is created equal, and corrupted backups are worse than no backups at all. From what I can tell, most backup media used today should last at least 4 years, with expectations of 10 to 30 years being fairly reasonable. CD-ROMs may last longer, though estimates of CD-R life spans are similar to the 10-year life span of magnetic tapes. Of more concern is how you treat your media – in short, it should be stored in a cool, dry, clean place, used in clean drives, and handled with care (don’t toss that cartridge into your bag!). The better you treat your media, the less likely you are to have trouble.
Backup verification: How do you know if one of your backup tapes has gone bad, or even if your backup contains the correct data? Verification. You must check the contents of your backup every now and then to make sure it’s working properly. I occasionally pull a few files from my backup sets just to check their integrity. Retrospect users should enable its verification pass, especially when backing up to tape. Backups take much longer but the safety is well worth it.
Backup redundancy: One of the best ways of reducing your exposure to media failure is to have multiple sets of backup media. That way, even if one set fails, you can still fall back on another set. Even if the alternate set is out of date, having older files is better than having no files. For instance, I currently rotate my weekly backups through three sets of backup tapes, two of which I reuse. When the third one reaches a certain number of tapes, I archive it and start a new set.
Automation: For most people, the main obstacle to backing up is that it’s a boring task they must do regularly. That’s why most backup programs offer a variety of scripting and automation features that take most of the drudgery out of backing up. I strongly recommend using these automation features – you’ll feel guilty if you lose a lot of work because you procrastinated about doing backups.
Storage location: What happens if your office is vandalized or a disaster befalls your home? Earthquakes, fires, tornadoes, and break-ins all happen. How seriously you consider offsite backup locations depends on the importance of your data. If you work in an office, taking a backup home each week is easy, and you can just as easily leave a home backup in your desk at work. If you work at home, consider giving backups to a friend you see regularly. For local backup media, consider a small, fireproof safe. However, make sure it’s designed to protect magnetic media in case of a fire – temperatures hot enough to destroy a backup tape might not be sufficient to ignite paper, which is what "fireproof" generally means. TidBITS sponsor APS Technologies has several media vaults that protect magnetic media from temperatures up to 1,550 degrees Fahrenheit for 30 to 60 minutes at prices ranging from $150 to $620.
Archiving: Although most people believe backups protect files they’re working on, a good backup strategy can also protect old files you need to keep, but don’t need on your hard disk (like graphics, scans, or video). Ideally, a backup system should address the need to archive important, unchanging files. Some people use Retrospect to back up to tape on a daily basis but archive data on CD-R on a quarterly or yearly basis.
From Strategies to Solutions — For most people, massive data loss is something that only happens to someone else. But if you’ve ever been forced to attempt the resurrection of a lost report or Quicken data file, you probably don’t want to get burned again. In the second part of this article, I’ll explore many of the backup hardware and software options available.