Cool, but useless.

S3 is useful as a backup system if:

• backups are easy and possibly automatic
• backups are available on all platforms (win, mac, linux)
  
• the files can be restored one at a time
• files can be accessed on any platform using multiple tools
• no proprietary junk. no vendor lockin.

Amazon S3 backup solutions. Directory schemes seem to be unique to each application. This chart shows the results of trying to read back the directory structure created by one S3 solution within another S3 backup solution.  Jungledisk (in compatibility mode) creates directories that s3fox and cyberduck can read, but it’s useless at reading the directories of others.  Cyberduck and s3fox are both reasonably good at reading the directories of others.

Proposed solution:

• Use jungledisk to do backups (due to its ease of use and convenient auto backup features) and to restore files
• Use s3fox and cyberduck to restore files if jungledisk is unavailable