Re: FileCopy overwrites the existing file

Grzegorz,

I would advise you to look through "Applied Cryptography", 2nd Edition
by Schneier - as far as I remember, he provides some examples of
seemingly perfect symmetric alrorithms being broken......

No matter how you look at it, large target data set will always remain
a potential risk to
the symmetric algorithms - even if you minimize "plain text risk
factor"

Anton Bassov


Grzegorz Wróbel wrote:
anton bassov wrote:
> Here we speak about the data samples of the size of *hunderds of GB*
with some "plain-text" (i.e. OS-related stuff) known in advance, so
that symmetric algorithms that are perfectly safe for encrypting some
files or messages may be not-so-reliable here - probably, you would
have to go for asymmetric ones, and, hence, pay performance penalties

The problem of "plain-text" messages and risk of "finding patterns" is
usually solved (at least I'll do it this way) by mixing the plain text
with values returned by good pseudo-random number generator (say 256bit
seed + at least 2^1000 period). Distinguish an output of pseudo-random
generator from random output is not doable in polynomial time so you are
safe with this technique, and mixed plain-text with such pseudo-random
sequence is as chaotic as such sequence itself.

I consider symmetric algorithms much safer than asymmetric ones (there
is no proof there aren't breakable in polynomial time, no one just did
it) and there are quite advanced factoring algorithms (like GNFS) that
pushes the limit of breakable keys further and further (1024bit RSA
encryption in few years won't be considered that much secure).

The asymmetric algorithms have wider range of usage, but for disk
compression nothing more than symmetric encryption is needed.

--
Grzegorz Wróbel
http://www.4neurons.com/
677265676F727940346E6575726F6E732E636F6D

.

Relevant Pages

  • Re: FileCopy overwrites the existing file
    ... that symmetric algorithms that are perfectly safe for encrypting some ... Distinguish an output of pseudo-random generator from random output is not doable in polynomial time so you are safe with this technique, and mixed plain-text with such pseudo-random sequence is as chaotic as such sequence itself. ... I consider symmetric algorithms much safer than asymmetric ones and there are quite advanced factoring algorithms that pushes the limit of breakable keys further and further (1024bit RSA encryption in few years won't be considered that much secure). ...
    (microsoft.public.win32.programmer.kernel)
  • Re: FileCopy overwrites the existing file
    ... I would advise you to look through "Applied Cryptography", ... I'm not familiar with the book, but examples of some symmetric algorithms being broken doesn't mean every symmetric algorithm is breakable. ... with values returned by good pseudo-random number generator (say 256bit ... encryption in few years won't be considered that much secure). ...
    (microsoft.public.win32.programmer.kernel)
  • Re: FileCopy overwrites the existing file
    ... already not so sure about reliability of encryption that you use... ... That how algorithms were broken in early ages of cryptography. ... No matter how you look at it, large target data set will always remain ... that symmetric algorithms that are perfectly safe for encrypting some ...
    (microsoft.public.win32.programmer.kernel)
Loading