Marramgrass

Back it up — online?

Time for an article-type thing that I’ve had stewing in my mind for a couple weeks. It’s very computer-y, but if you use a computer for anything at all, pretty much, then it may be worth reading. It largely deals with Mac-based solutions and software, but those that aren’t specifically available for Windows will have something very similar out there. I’m not joking, this could save you a whole pile of heartache and frustration.

Everyone who uses a computer knows you should regularly and reliably back up your hard disk, especially with all the music and the sentimentally valuable photos we all have lying around these days. Between the desktop computer and this laptop, I’ve got about 60 GB of photos (some of which could be replaced with some hard work, many of which couldn’t), maybe 30 or 40 GB of music files (all legitimate, I might add), and several GB of other stuff, include some critical work files that I really can’t lose.

What if my hard disk decides to throw a wobbly? Or the PC explodes? Or my laptop gets half-inched? That’d be bad.

Which is why backup is a good idea — taking copies of your data in case you lose the working copies/originals. For advice all should read, check out The Tao of Backup.

I’m already fairly good about taking clones of hard disks regularly. I’m seeking an additional solution, possibly an online one. Being in an unusually methodical mood, and having other more responsible things to be distracted from, I decided to have a good look at the available options.

Why online?

Here I’m not talking about full-system backups. There is no sensible substitute for using reliable software to take a file-by-file clone of your hard disk (the Leopard-ready update to SuperDuper! has been released today as I work on this). Twice, and then store them seperately. Backing up gigs and gigs of data online is never going to work while those of us lucky enough to have broadband are sitting at the base of a dinky little 256 kbps upstream (the A in ADSL means ‘asynchronous’, which is to say you receive data much, much more quickly than you can send it).

What I’m after is a redundant way to conveniently backup maybe a few hundred megabytes of data (a couple of GB, tops), largely work-related stuff I really can’t lose, to somewhere it will be safe if I suffer a local catastrophe. For this, I have a suspicion online storage may be the way to go.

Mozy and the like.

The obvious contenders are all the various services along the lines of Mozy, specific online backup services, often with their own client software that automates the backups on your computer. To take the example of Mozy, it has clients for Windows and for Mac OS X, you get the first 2 GB of storage free and can then upgrade it for US$5 per month, but they do specify home use only unless you sign up for their MozyPro service.

I can see this kind of service being exactly right for many, but there are a couple of (admittedly irrational) things that put me off. The big one is that I don’t like the idea of backup software that exists just to interface with that particular service. No real reason, it just lacks elegance. The cost is nice — I’m unlikely to reach that 2 GB limit, but the non-business use stipulation is a problem. Also, for this kind of application I need to see the longevity of the service. However, Mozy has been around since 2005.

Amazon S3 + JungleDisk.

Very different is JungleDisk. It isn’t a service, as such, although the “Plus” variant adds a few extra services for a mere US$1 per month. Rather, JungleDisk is a little bit of software that presents Amazon’s Simple Storage Service (S3) as a disk mounted on your computer.

In S3, Amazon provides internet-attached data storage at prices that are very low for smaller amounts of data, but scale up to rather a lot when you’re into the realms of tens of gigabytes and more. The price list is at that last link, and James Duncan Davidson has a useful outline of the relative costs involved (via Gruber). No free basic level, but for my level of usage it’s literally pennies a month and you only get charged for exactly what you use.

The disadvantage of this apporach is that JungleDisk requires you to supply your own Amazon S3 account, and S3 is not a service intended for direct exposure to consumers. You need an Amazon Web Services account and to then start taking note of access keys and so on. It’s not complicated, but adds an extra layer of doubt if you’re uncomfortable with it all.

The JungleDisk client costs US$20 and includes encryption and some backup functionality as well, although I’ve found the interface to be fairly unintuitive. That said, for general access to your files you don’t get much more convenient than a volume mounted to your computer. Other pluses are Amazon’s well-established position and infrastructure and that you are billed directly by them.

To actually run and maintain your backups, since JungleDisk mounts as a volume like any other, I presume you could use software like ChronoSync. I use it to keep my laptop synced with the file server at work, and it’s a gem. It costs money too, though.

Amazon S3 + Transmit (+ Automator).

Rather than JungleDisk, I’ve lately been using Transmit (an excellent FTP client) to interface with S3. Unfortunately no encryption, but solid synchronisation functions and easy to use Automator actions. So far this is working well for me. (‘Synchronisation’ in this context meaning to keep the backup copy in step with changes to the local data.)

Or...

There is another option, and it’s the least sexy but perhaps the most practical for my needs. USB memory sticks are dead cheap these days (a couple of GB for a few pounds), and very convenient for keeping handy backups of a suitable volume of data. It’s important to keep them well seperated from the computer, or theft or damage will take them too, but one in your pocket or on your keyring can accomplish that while giving you access to your backup without an internet connection. The downside, probably fairly easy to lose.

Which, then?

Financially speaking, online backup costs massively more than using local storage, especially when you get into the realms of big accounts vs. hard disks. HDD space is cheap cheap cheap. But for the volumes I’m looking at here, this isn’t an issue, and at this level the convenience of a system that makes it easiest to develop the habit of actually keeping the backups is worth the small ongoing cost.

For now I’m storing the data on Amazon’s S3 service, and getting it there using Transmit via an Automator workflow that keeps the folders I’ve selected in sync. The initial setup is a little involved, and it doesn’t support archiving different versions the way I could have using ChronoSync over JungleDisk, but once it’s going it’s a nice — if basic — system. The key to this is that without a regular habit of keeping the full clones of my hard disk up to date, I may as well not be doing it, but with that in place it gives me that added comfort of a little bit of extra redundancy.