Gav's Blog

After hearing about one of our GPS Society members losing their data in a computer malfunction tonight, I’ve decided to sit down and flesh out some thoughts on developing a good backup strategy for your computer(s). This is one of those get a round ‘tuit posts that I’ve been meaning to do after seeing people caught by HD failures on the Digital Photography School forums.

The topic of developing a good backup strategy for your computer surely makes most peoples eyes glaze over. It is decidedly unsexy until such time as you need it. Of course, by then it is too late. I’m hoping to combine some of my IT, risk and emergency knowledge to provide some insight into develop a suitably robust backup strategy.

If the consequence is lost data, what are the risks?

When developing a backup strategy, it is important to have a good understanding of how data can be lost – the risks – so that we can create a simple yet comprehensive plan to backup our data that accommodates the many different ways data can disappear.

So, lets pick a few. I’ve named them L1-L5 where ‘L’ is for loss.

• L1 Loss of computer (e.g. theft, smoke or water damage; electrical surge from computer power supply)
• L2 Filesystem accidents – formatting of filesystem, deletion of files, data corruption
• L3 Malicious software – formatting, deletion, or encryption of files with an unknown key (e.g. encrypt and extort)
• L4 Mechanical failure of the hard drive (the dreaded clunking sounds)
• L5 Loss of home containing computer (e.g. fire, earthquake, flood, landslide)

Whilst not comprehensive, this includes a good range of different issues we may face where a backup would be very handy and save us a lot of time, and potentially money. If we can come up with something that protects us from these losses, we should be doing pretty well.

What we need to do now, is look at various means available to backup data, and then create a quick matrix comparing each type of backup, and which losses it may/may not protect us from.

Firstly, let’s identify a number of backup solutions.

• S1 Backup to CD/DVD/HD and store on site
• S2 Backup to external HD on site
• S3 Backup to internal HD
• S4 Backup to other computer at home
• S5 Backup to Internet (service or web host)
• S6 Backup to CD/DVD and store off site
• S7 Backup to HD and store off site

Now, all these solutions are not equal. What we need to investigate now is which type of loss a given backup solution can protect against. I’ve created a sample table below to give you some idea of how it all comes together.

We start with a grid comparing types of loss and solutions. A green box means the solution generally prevents that type of loss, and red box that it generally doesn’t protect, and an orange one means that it may provide some protection.

Next, we compare solutions with various costs and constraints – in this grid green means it isn’t really a cost/constraint, red means it is a cost/constraint, and again, orange means it might be a cost/constraint.

In this example matrix, two points stand out.

• backing up to an internal hard drive does not provide much protection against data loss
• quite a few backup solutions do not protect against major losses such as the loss of a home from fire.

Additionally, every backup solution has a number of costs and/or constraints on its operation. The next step has been to add some cells that identify some of the more common costs and constraints associated with each solution.

What we can see is that there is no single perfect solution. We could extend this further and add a grid outlining some of the benefits of each backup solution – they all have some – and this would also further educate us in the development of our backup strategy.

Now, we’ll use this grid to look at selecting a couple of complementary backup solutions that avoid each others weakness.

Personally, I’m a fan of backing up my home computer using Time Machine on a Mac to an external USB hard drive (effectively S2). As you can see from the matrix, this protects me against most of the common losses, except the rather catastrophic loss-of-home. Clearly then, I can select an Internet or off-site solution as well that will provide me with more complete data protection than just backing up to an external hard drive.

Quite a few people will look at the Internet backup option (S5) and think that it looks pretty good, but be warned, there are some issues that you may face including – the speed of your internet connection when backing up files to remote servers, ongoing service fees, and potential privacy risks by storing you files on a remote server of a business.

I’d recommend selecting solutions so that you can meet the follow three four requirements.

1. You should have at least three copies of your data (source + two backups).
2. At least one backup must be reasonably current and disconnected from the computer most of the time (except when a backup is being made).
3. At least one backup file must be offsite.
4. At least one backup should be incremental.

Ian (in the GPS forums) made a good point about incorporating incremental backups into the process. Broadly speaking, there are two types of backups, full (where everything is copied at once) and incremental (where only the files that have changed since the last backup are copied). When doing incremental backups, the first backup is a full backup, and then incremental backups take place from there on. Time Machine is a good example of incremental backup software – every hour it backs-up any changed files.

As I’m not that keen personally on online backups, I’d recommend one of the following as the minimum. There is nothing wrong with making more copies on CD/DVD media to supplement the main backup solutions.

• external hard drive onsite + DVD media offsite (affordable setup)
• external hard drive onsite + external hard drive offsite (same sizes, switch them once a week or month, expensive setup)
• synchronise files between two home computers on network + external hard drive offsite (utilise existing hardware and provide backups of both computers)

There are three other tips to provide as well:

• If you use backup software, keep a copy of the install media (and licence key if appropriate) with the backups
• If you need quick access to data upon failure, make sure that at least one of your backups uses a very accessible filesystem on external hard drives (CD/DVDs are good as they generally use filesystems that are accessible in any computer). This means you can literally plug them in and access key files without having to perform a software installation and full restore
• AND TEST THAT YOU CAN ACCESS BACKED-UP DATA and/or RESTORE FROM BACKUPS

Finally, as you should have a GetAway Kit for natural disasters and the like, in addition to your other important paper information such as identification, policies and photos – you should also include a backup of your data in the kit. If you haven’t got a GetAway Kit, then now is a good time to learn about how to get ready!