[Discuss] Deduplication

[richard.pieri at gmail.com (Rich Pieri)]

I think deduplication is kind of overrated and impractical. As was
pointed out several times in the EFI thread: big, fast drives are
cheap. So what if there are two or three copies of a file on a backup
set? The dedup overhead is more costly than the storage.

Where deduplication starts becoming practical is at the large scale
such as enterprise backups for an entire fleet of workstations.
Deduplicating a few dozen files is impractical. Deduplcating the
backups of thousands of workstations all running the same base OS is a
different story.

Aside: The ext# family don't have CoW capability so dupremove can't
work on them.

-- 
\m/ (--) \m/