Re: FileCopy overwrites the existing file

anton bassov wrote:
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......

I'm not familiar with the book, but examples of some symmetric algorithms being broken doesn't mean every symmetric algorithm is breakable. That how algorithms were broken in early ages of cryptography.
As for smart method of compromising cryptosystems, there are pretty fine examples of successful attacking asymmetric ones as well.


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"

There is no such thing like "large data sets" with well designed algorithm. You do realize that hundreds of gigabytes, terabytes or petabytes are nothing compared to the number like 2^2000, don't you?

I can implement symmetric algorithm which not only will be unbreakable but the encrypted code will be undistinguishable from random data (with limited time and resources of course, but I can easily push the time limit required for that far beyond, say the estimated time of existence of our universe, even if one had unattainable computing power).

I cannot say the same about any asymmetric algorithm and no one can.


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



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

Relevant Pages

  • Re: RSA vs AES
    ... Public Key Cryptography. ... 10 times weaker per bit than symmetric algorithms. ... you need a secure channel to ... once to send a randomly generated session key to the other side. ...
    (sci.crypt)
  • Re: FileCopy overwrites the existing file
    ... that symmetric algorithms that are perfectly safe for encrypting some ... with values returned by good pseudo-random number generator (say 256bit ... 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
    ... 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
    ... 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)