Backup: Immediate, Full and Long-term

Preserving the availability of digital artifacts is a goal worthy of pursuit. First, I’ve got thousands of digital family photos, and I don’t want to lose them to hard drive failure or lock them up so that they’re hard to get to. Second, I’ve got my email stored on my computer for the past several years. The recent email is what’s most valuable to me, but every once in a while, I need to search through email archives to find things, like a license key for [Date Book 5](http://www.pimlicosoftware.com/datebk5.htm). Third, it took weeks to install software and configure our laptop. I don’t want to have to repeat that work if the hard drive happens to stop working — especially if a project I’m working on needs to be done soon.

There are three main types of backup that are important to me: Immediate backup, full backup and long-term archival.

__Immediate Backup__
—————–

What I’m currently working on with the computer is usually more important than what I was doing on the computer a few weeks ago. The auto-save and even the “undo” feature of most word processing programs can help me when one of my children touches the keyboard and accidentally deletes most of the text. Auto-save and undo won’t help if my laptop is stolen or the hard drive fails. That’s why I use [mozy.com](http://mozy.com) for automated, off-site backups of my Windows laptop. It’s well worth $5.00 per month for this service, and it’s easy to pay for: skip eating out for lunch once per month.

For backup to happen regularly, automation is key — especially for immediate backup. I would make full backups more frequently if it were an automated process. I use a monthly repeating reminder so I remember to backup the things that aren’t automated.

__Full Backup__
———–

Admittedly, hard drives don’t fail often, and laptops that usually stay at home aren’t often stolen (at least, not in my neighborhood). But when it does happen, it’s a pain to reinstall the myriad of applications we use on a semi-regular basis. This is why a periodic, full backup is valuable. Doing a full backup with optical media takes too much time. External hard drives are much faster, have more capacity, and are inexpensive. They plug in using standard connectors such as USB, FireWire or eSATA. I store my external USB hard drive in a fireproof box.

__Long Term Archival__
——————

I want the best of my digital memories (e.g. photos) to be preserved for decades or centuries. A CDROM may be readable in ten or twenty years, but not fifty or a hundred. There probably won’t be hardware to read it in fifty years. Will computers in fifty years recognize JPEG format? No idea!

To preserve digital artifacts for that long requires refreshing it periodically into newer formats and storage media. It’s a good idea to use open, standardized formats rather than proprietary formats. For photos, this means to use JPEG and PNG in preference to Photoshop format.

Rather than refresh constantly, there’s the option of _printing_ photos and documents. It’s going to be easier to view a physical photo or a printed document in a hundred years than to unlock the secrets of an old hard drive.

__Trust, but verify__
—————-

I tend to trust my backup solutions, but it’s necessary to verify that they’re working. My brother’s computer periodically downloads my digital photos. I trusted that this was, at least in part, a good off-site backup. I learned recently, however, that his computer deletes old photos when space gets low, which is often.

__Resources__
———-
[Preserving Your Digital Memories: What you can do](http://www.digitalpreservation.gov/you/digitalmemories.html)

A few backup solutions: Mozy, Carbonite, SyncBackSE, and JungleDisk.

Interesting projects to backup using P2P protocols (including featurs such as encryption and
fault tolerance): [Tahoe](http://allmydata.org/trac/tahoe)
with a [writeup from LWN](http://lwn.net/Articles/280483/) and [Flud](http://www.flud.org).