A blog entry by Ryan Lowe reminded me of an idea I’ve always had for backup software that seems so simple that I can’t believe I haven’t found it yet.

A lot of people–Ryan included–use CD burners to copy critical data like My Documents and store it away somewhere in the event of a crash. This solution gets the job done, but requires manual copying, which means time away from getting “real work” done. It also requires an accessible CD burner. I have a burner, but not on my laptop, which adds steps to this process to copy the data over to another machine. And, while CDs are cheap, they’re somewhat burdensome to juggle and keep track of.

I always wanted a solution to back up my data over our home network. Last year, I bought a 40 GB drive at Best Buy because I got a really great deal. It sits in our home PC, a modest P233, and is largely underused. My main PC, a laptop from work, has a 20 GB drive, about 10 of which is the data that I care about. To my knowledge, though, there really isn’t an easy way to do this kind of backup short of manually copying the files over, which just puts me back to where I started.

I want a TSR (do they call them that anymore?) program that I can set up once. I want to be able to point out the folders that contain my critical data and the backup “server” for which it should be looking (by IP address, DNS name, IPX/SPX, whatever). Then, every time I start my PC, I want the program to wait quietly until it can find the backup server on the current network (no ugly errors if I’m at work or school). Once it finds the machine, it uses just a preset percentage of my bandwidth to copy files to that machine–better yet, it would monitor my bandwidth usage and throttle up or down depending on my decrease or increase in demand.

Being a smart program, the backup software wouldn’t simply copy my entire MP3 collection across the network everytime I was in range of the WLAN. Instead, it would keep a database of which files have been changed when and copy only the files that have been recently updated. Not only would this be an intelligent use of resources, but it brings to light an interesting property of disk usage: I only use some of my files (perhaps even only 10%) most of the time and most of my files some of the time. Most of my files–like music or video–go unchanged indefinitely, while there are a handful of files–like email or databases–that are updated every day.

With the proliferation of home networks in the last few years, this would be a very practical and even profitable idea. Of course, the ideal package would be Open Source. Rather than be relegated to backing up only to a Microsoft machine running the latest and greatest operating system, I want to be able to put the hard drive in a Linux machine and copy my PC and future iMac data to it. So, the client might ideally be written in Java or be a Mozilla app.

And, just for kicks, I’m using the LazyWeb trackback to broaden exposure to this idea and see if there are any takers.