Why Backup?

In may ways this seems like a silly question, but let me go over the major reasons you should backup regularly and often. First we have catastrophes, like the time I walked into the office to sound of loud clicking coming from my computer, complete mechanical hard drive failure, all data lost. Or what about bigger disasters like earthquake, fire, tornado, or theft. Then we have "minor" catastrophes like viruses, file corruption, and the ever popular "accidental" deletion. My favorite is the "I didn't think we'd need that file any more" kind of file loss.

How to Backup?

This is actually a more important question, largely because there are many options and methods to consider. First we must contemplate the different kinds of backups and their purposes.

Backup Types

Snapshot or Image backups are complete backups that create, as their name implies, an exact image of your system. These backups are ideal for situations where your hard drive goes to the big bit bucket in sky. Image backups quickly get you back to where you were when the backup was made. The downside to image backups is that they are very time consuming and your computer usually is completely unavailable during the backup and they may require large amounts of backup media to complete. The positive side is that since they backup everything, you just have to restore the backup and you are done, no reloading of software, or operating system, it all comes back in one shot.

Full Backups are backups of everything on your system, software and data. Unlike image backups they are not exact mirror images, but just contain all the data. Full backups share with image backups their time consuming nature, but unlike image backups, you may be able to perform full backups while the computer is doing something else. The major drawback to full backups is the time and media requirements, but if you schedule the backup to run when you are not around, they aren't too bad.

Incremental Backups are a way to not have to make full backups as frequently. Instead of backing up everything, you only backup things which have changed since the last backup (If you backup all changes since the last full backup, this is called a differential backup). The upside to incremental backups is speed. Because you are not backing up everything, the backup can proceed much faster and occupy much less space. The downside to incremental backs comes when you need to restore after a disaster. You have to restore the last full backup followed by however many incremental backups you made since then, in reverse order (unless you have been making differential backups, in which case you only need to restore the latest backup after restoring the full backup).

Data Only Backups are just what they say, you are not backing up the whole system, only the data. Data backups can also follow the Full/Incremental methods as well. The advantages of data only backups are ease and speed. Data only backups are fast enough that you can backup several times during the day, and comfortably maintain several copies for redundancy. The downside of data only backups shows up in a catastrophic situation. You will have your data, but that is all. You will have to reload the operating system, and any software.

Where and How to Backup

Traditionally, backups were always made to tape. This was because only tape could hold the vast quantities of data backups require without being prohibitively expensive. In large installations, tape is still the primary backup medium, but in smaller business there are more options. Hard drive prices have continued to fall and hard drive capacities have continued to expand. This provides both the problems and solutions. With larger hard drives, it is harder than ever to keep up with the volume of data you need to backup. On the other hand, the lower price of hard drives allows for using hard drives themselves as backup media.

External hard drives in the 100-750 Gigabyte range are now readily available and reasonably priced. External drives have the advantage of being actual hard drives so they don't suffer the from the slow speed or serial access disadvantages of tape. Using either USB 2.0 or Firewire (1394) interfaces, external hard drives can run at nearly the same speed as internal drives. They also have the advantage of being portable. A single large external drive can backup more than one computer.

A second way to backup using the hard drive as media is using the network to backup one computer to another or to a NAS (Network Attached Storage Drive). This type of backup has the advantage that no additional hardware needs to be purchased. Frequently, we use only a small fraction of the large hard drives that computers come with, leaving ample room to store backups for other machines on the network. The drawback to this approach is that this sort of back is not portable and provides no protection against hazards such as fire or theft.

Another popular media for backup are CDs and DVDs. CDs can hold up to 700MB uncompressed and DVDs can hold up to 4.7GB. Several years ago a single DVD would have been sufficient to backup most computers. Alas, hard drives have gotten so big that full backups can no longer fit on a single DVD. Even so, DVDs are large enough that most systems can be fully backed up onto two or three disks. My personal favorite in this are DVD-RAM disks. This was the first read/writeable DVD disk format and has somewhat fallen out of favor compared to DVD+/-R/RW formats, but they have one very distinct advantage. They are designed for multiple read/write cycles. Once a disk is formatted it looks like another hard disk to your computer. You do not have to "burn" a DVD-RAM, you just read and write to it like really big floppy. The only downside is that DVD-RAM seems to have nearly lost the format war. DVD+/-R/RW disks are getting much cheaper, but the jury is still out.

A new entry in the backup media area comes from Iomega, the same folks who gave us Zip disks. They call it the Rev drive. It is essentially a cartridge hard drive. It currently gets 35-70GB per cartridge uncompressed. They claim that this allows up to 90-150GB compressed using their included software. The cost of the cartridges is not cheap, but neither are tapes in the same category, and the Rev drive claims to be 8X faster than tape and random access to boot.

Another new area made feasible by faster internet connections is internet backups. Here you can sign up for an account and then send your data over the internet to the providers servers. This provides easy offsite backup, but has monthly fees and will take considerable time depending on the speed of your internet connection.

Backup Strategy

So now you know the different backup methods and media to use, which should you use? The short answer is all of them. What I mean by that is that you should use a combination of the methods and media that suit your needs and risk factors. The key thing is to have a strategy and to stick to it.

Q:How often should you backup? A:How much data can you afford to lose?

Ideally we don't want to lose any data, but we also don't want to inconvenienced either, so we need to balance the need to keep our data safe, with the need to get our daily work done.
If you are dealing with a single computer, the simplest method is to have three different types of backup working for you.
First, you should have an image backup of your system. As mentioned above this covers everything and is very time-consuming to create, but if you create one once and create a new one every six months or year that should get you covered.
Second, backup all your important data daily. The easiest way is to a DVD or external hard drive. Both of these are portable, and you should ideally have a copy offsite at all times. You should also rotate your backup media or backup sets so that if a file gets accidentally deleted or corrupted, you can go back and pull it off of an earlier backup. For this reason it is a good idea to rotate through at least three, and preferably, six sets.
Last, keep a full backup that you run at least weekly or monthly. This will allow you to get back up and running quickly if you have a hard drive crash and it has been a long time since your last image.

If you have a network, the backup situation is similar, but different. Your workstations are easy, image them after they have been set up and loaded with all their standard programs and then make sure that all important data is stored on your server.
Backing up your server is similar to backing up a single machine except that you want to be more diligent. You can save yourself from hard drive crashes by using mirrored hard drives where if one drive fails the other still has all the data on it and you don't have to be down in order to replace the bad drive, even better use a RAID array of four or more drives. This improves both safety and file speed. Even so you still need external backups. In a server this is even more critical. At this point you can backup your data using tape or Rev drives. You can also do large backups to DVD, but know there will have to be some available to swap disks.
Lastly, you can still use a designated server or workstation on your network with a large hard drive as a backup server. This can provide the speed and immediacy of backup, but doesn't satisfy offsite backup needs. For that you can continue to use traditional media or access a backup server over the internet, or a large external hard drive.

Testing Your Backups

No matter what method or strategy you use to backup your data, no backup is really complete until it has been tested. This means that your should try and restore some or all of the data you have backed up, before there is a problem. There is nothing quite like the sinking feeling you get when you go to restore a backup after a file failure, only to discover that the backup is no good. The only sure test of a backup is a test restore, until you do that, you can't really say your are backed up.

Conclusion

Backups are like insurance for your car. You hope you never have to use them, but you will feel much better if you are protected. I have watched first hand how adversely businesses are affected by losing key data. Backups are the insurance premium we must all pay if we don't want to experience that horrible sinking feeling of knowing that irreplaceable data is gone and there is now way to get it back.