A lot will depend on the size and spread of data you wish to backup on the machines, and how they will generally be connected (do you want to update the backup of a laptop when you have it off-site, for instance).
I use a technique similar to those described in this very useful article to maintain offsite backups of my docs and settings from my main box, netbook, and mail server - it should be quite easy to implement on a number of machine if the files you backup are in consistent places (just one script to copy to all the machines, and setup a SSH key pair for each for authentication).
The rsync algotithm is a very efficient way to maintain backups of potentially large amounts of data even over slow links, and being a command line utility it is easily used in scripts that you can run manually or automate via cron or windows' scheduled tasks. rdiff is another similar option, which uses the rsync algorithm to create "delta" files for each run (rdiff-backup wraps this utility to produce an arrangement that works much the same as rsync+hardlink snapshots). There are a number of extra front-ends available for them too, which either just add a GUI front-end, wrap them to provide a specific arrangement in a more friendly way (though losing some flexibility), or both.
There are a number of ports of rsync+ssh for Windows (the one that is part of cygwin being the one most people use, cwRsync is another apparently popular option). Though you might need to be running Linux/BSD server-side to use the hardlink-based snapshot technique (I'm not sure if the Windows ports support hard links though they may as the NTFS filesytem can support/emulate them), if you are sticking using Windows server-side rsync will otherwise work well and rdiff's technique doesn't need hardlinks anyway.
