Hi, I am curious about the incremental backup method suggested by
Andrew Tridgell on page 92 of his thesis describing rsync. The basic
idea is that signature information would be saved along with the data
and incremental changes could be produced using this data. Some
advantages for incremental backups would be:
1. If a large (possibly binary) file changed slightly, only a small
diff would have to be added to the archive, instead of an entire
duplicate copy of the file.
2. To produce such a diff, the large file would not have to be read
from the archive (which could be slow) in its entirety.
So has anyone implemented this yet, or begun work on it or something
similar? Thanks for any help. Please mail me directly as I am not
subscribed to this list.
--
Ben Escoto
