On Thu, Apr 2, 2009 at 8:30 PM, <morty AT frakir DOT org> wrote:
> On Thu, Apr 02, 2009 at 06:11:16PM -0400, Victor Hugo dos Santos wrote:
>
>> is true that bacula don't compress encrypted files or data ??
>> and in positive case, why ???
>
> Compression works by reducing redundancy in low-entropy data
> (e.g. eliminating patterns in data with patterns.) Good encryption
> schemes produce output that is nearly indistinguishable from random,
> high entropy data (e.g. removes any obvious patterns.) So the output
> of good encryption is mathematically uncompressable. This is related
> to the reason why you can't compress data that has already been
> compressed.
I didn't know this.
but, now (with your explication) is very comprehensible !! :D
and a bit test:
==========
$ ls -l 6zW0Zf4XPFhvoPTzgb1pWdjxHE0EnPldK65*
430806 2009-04-03 09:21 6zW0Zf4XPFhvoPTzgb1pWdjxHE0EnPldK65
$ gzip -9 6zW0Zf4XPFhvoPTzgb1pWdjxHE0EnPldK65
$ ls -l 6zW0Zf4XPFhvoPTzgb1pWdjxHE0EnPldK65*
430930 2009-04-03 09:21 6zW0Zf4XPFhvoPTzgb1pWdjxHE0EnPldK65.gz
==========
LOL... the final file is a little bigger that original !! :-(
> Note that you *can* go the other way, i.e. you can encrypt compressed
> data. Encryption doesn't care whether the input is low entropy or
> high entropy.
mmm..mmm.. and mmm..
would be necessary have unencrypted files in my machine and after of
backup, compress the files in the server.
or, compress the files in my machine before backup in my encrypted
folder, but in this case I have the double information (data and
encrypted data) in my hard drive.
mmm.. I don't think it good idea !!
others alternatives ???
thanks
--
--
Victor Hugo dos Santos
Linux Counter #224399
------------------------------------------------------------------------------
_______________________________________________
Bacula-users mailing list
Bacula-users AT lists.sourceforge DOT net
https://lists.sourceforge.net/lists/listinfo/bacula-users
|
