We’ve all heard it: “Backup, backup, backup.” Do you know why you need to back up? Yes, being prepared for the unknown is important, but this is actually a known fact: MOST HARD DRIVES WILL BE DEAD WITHIN 3-5 YEARS, and many do fail within the first year. Read that again. Now think about your 3 year-old PC… that hard drive is a ticking time bomb of pain just waiting for the most inopportune time to die or even just corrupt the data that is on there (which is just as bad or even worse from a data standpoint).
But how many of us actually do back up our data and how often? From my personal experience, very, very few people have backups that were made within the past month or three, if anything at all. But still, everyone acknowledges that they need to backup.
So why don’t they? Well, first off, it costs money. You need something to back up onto. Secondly, it takes effort and often isn’t trivial to do. Third, it’s boring (Well, it is. Do you REALLY want to spend hours trying to figure out what needs to be backed up, doing test backups, testing those backups, and then regularly performing the backups for the rest of your life? Yeah, me neither).
So instead, we take the risk, hoping and praying that nothing happens until we buy that next computer and copy everything over, thus effectively making our old computer the “backup,” even though it isn’t really one because it’s now been handed down to some other use, and the data on there now will potentially be changed or even deleted. So when we have to go back to it a year from now because that shiny new computer’s hard drive died, we are only able to recover 1/4 of our data, and we spend days upon days recreating it, and some data will just be lost forever.
Perhaps spending the time right now really is worth it in the long run. In this guide, we’ll look at the currently available options.
Types of backups
There are three types of backups. The first is a full backup. This is a complete copy of everything that is to be backed up. It will use up the same amount of space as the files on the disk, before accounting for compression. Full backups are nice because you can just find the tape/file you need and restore it with no mess.
The second type is an incremental backup. Incremental backups only store the changes or changed files since the last backup, whether it is a full backup or another incremental backup. They typically save quite a bit of space since you aren’t storing many copies of the exact same files, and they are much quicker to process during the backup. The downside is that you’ll need the last full backup plus every incremental backup you made since then in order to restore the files. So restoration is more painful, and if there is a damaged tape or other media somewhere in the middle, you could potentially lose most of your data from that point forward. Incremental backups do require special backup software to track what files have changed and upon restore, to re-join the data from all the different backup sources.
The third type of backup is a differential backup. It is a middle ground between full and incremental backups in that it stores all files that have changed since the last full backup, ignoring any previous differential or incremental backups. That makes restoration much quicker than with multiple incremental backups, but it can require quite a bit more space and take longer to perform the backup.
How often to backup
Backup frequencies depend primarily on how fast/much your data changes and how important it is to you. If you only add and change a few files per week, a monthly full backup would be fine. Also keep in mind that something is better than nothing. It’s better to have a better to have a week-old or day-old backup than a month-old backup, month-old backup than a year-old backup, and a year-old backup than nothing at all.
This is really the first thing to decide when planning a backup strategy. Businesses typically do weekly full backups on the weekends and nightly incremental backups Monday through Friday. The hardcore approach is to do a full weekly backup, duplicate it, store the copies each in separate locations (different cities/states is even better), and then do nightly incremental backups. That way, even should something catastrophic happen (such as the destruction of the building with one set of backups), the business is still never more than one week out of date.
Of course, those same businesses have an IT staff that are responsible for tasks like backup management. If you’re reading this, though, you’re probably not nearly so well equipped. Instead, you could do monthly/bi-weekly/weekly full backups, and that would probably be good enough. Depending on the backup method you choose, you could also do some incremental backups between the full backups.
Price per Gigabyte
Since money is always in the mix, this has to be factored in. How much does it cost to back up one GB of data? This can be especially important if you want to keep multiple backups (say, weekly backups for the past month or two). The price might actually be higher than the hard drive the data is stored on right now.
You’ll also need to factor in the price of the drive for something like a tape drive that is only used for backup.
Speed of backup
The speed of the backup is relevant because you often can’t use the computer during the backup process (so data doesn’t become inconsistent between the backup and hard drive storage). So unless you’re going to start the backup at the end of the day (and that is very common), you’ll want the fastest backup method possible. With some backup methods, you’ll have to swap tapes, CDs, or DVDs in the middle of the backup so you’ll have to physically be present to do the swap. There are ways of automating that, too, but that starts adding to the price.
Ease and speed of data recovery
This is often something that is glossed over when considering backup options. If the backup can’t be restored quickly and easily, even more valuable time and energy will lost. The faster the data is back, the faster everything can return to normal. The more complex the backup (greater number of incremental backups and tapes/CDs/DVDs), the more complicated and slower the recovery will be. Speed of media transfer also plays a role here as well. If you have everything backed up over the internet, it might take hours and hours to download the backup.
Of course, for individuals, the speed isn’t as important as the ease. If it’s hard, it will take more energy and focus to figure out what to do for a recovery. That’s one of the reasons it’s important to do a practice data recovery so you’re familiar with what to do when you need to recover.
Number of previous backups held
It is a good idea to hang onto old backups. The reason is that a newer backup might become corrupted, damaged, or destroyed, or the data you backed up might actually be wrong. If you only have one backup, you’re out of luck. You need to have an older backup to turn to for those situations.
Of course, holding more backups on hand adds to the price of backup. So you want to try to find the right balance.
Off-site vs. On-site
On-site backups are ones you keep on the same premises, and off-site backups are ones that are stored somewhere else. Keeping your backups in the same building doesn’t do any good if it is destroyed by fire, and keeping them in the same city doesn’t do any good if that city is wiped out by a massive flood or other catastrophe. Of course, dealing with off-site backups are annoying because you have to get the data from here to there somehow, and then you need to be able to get it back quickly if the need arises, especially if off-site backups are the only option you are using.
Most individuals can get away with on-site only backups, but it’s still a good idea to keep all your important data at a friend or family member’s house that you really trust and update it every year or so… or whatever frequency you feel comfortable with. That way a fire to your house won’t destroy absolutely everything.
How relevant durability is to you depends on your backup approach. If you do weekly full backups and keep one old backup, then you’ll be rotating through 2 sets of backup media, and it will never be more than 2 weeks since your media was last used. Usually with this scenario, the durability of the backup media (which is always measured in years) won’t really be a factor because it’s much more likely that the media will fail while you are creating the backup than during the one week where you really need to restore data from that tape.
However, if you only backup every year or so (e.g. to do the important data off-site backup), then you need to be aware of the media’s durability. If you actually wish to keep the media as an archive, then the durability becomes vitally important. You’ll still need to move the data onto new media every so often or you risk having losing the archive to media failure.
Another factor for durability is storage conditions. While CDs and DVDs often are said to have 20-100 year lifespans, a burned CD can be killed in a matter of hours by exposing it to direct sunlight. Magnetic tape is usually expected to last for 20-50 years, but data can be corrupted by a short exposure to a strong magnet. So you’ll also need to be aware of the media’s weaknesses and account for that as well. Most media prefers a cool, dark place with mild levels of humidity for maximum media lifespan.