Re: Test for randomness



hayes.gr@xxxxxxxxx wrote:
Well I was testing my encryption routine for strengths and weaknesses.
My first test was to verify the avalanche ability of it. Using notepad+
+ I viewed 2 files at a time under the hexediter.
Both files were identical prior to encryption then. Each time I only
change 1 bit at various points along the password.
Worked good probably would have been much faster if I just wrote a
routine to do it. But at least that was done.

So next I wanted to see how good of a randomness was created out of
the encryption. I tested the method I came up with on two types of
files. The first being a jpeg I figured it should show very little
change after encryption and then being zipped. Yep I used a zip
program to test for randomness.

I then tested a text file. The US constitution 45K in size. I zipped
it with out encryption to see what type of reduction came of it. It
was reduced to 15K.

My assumption is that the more random the data the worse the
encryption will be on that data.
So I then encrypted the same file and after words tried to zip it. It
only reduced to 44K not sure the exact number of bytes that is. but
44K to 45K isn't that bad. It means at least a 97% randomness to the
pattern.

Anyone know of any software for measuring the randomness of a file? I
am just trying to prevent reinventing the wheel. Granted the zip test
is actually pretty good. I may try some other compression methods for
testing as well.

google for "diehard"
.



Relevant Pages

  • Re: Test for randomness
    ... My first test was to verify the avalanche ability of it. ... Both files were identical prior to encryption then. ... It means at least a 97% randomness to the ... DIEHARD is 1990's technology. ...
    (sci.crypt)
  • Re: Missiles in Space Combat?
    ... the ability to make it tougher. ... Quantum computers show promise for easily overcoming certain types of encryption algorithm, for example, and one could always postulate some new mathematical breakthrough to deal with some of the others. ... One time pads rely on randomness to achieve, ... Matter is fundamentally stupid:- It tries every other path first. ...
    (rec.arts.sf.science)
  • Re: Test for randomness
    ... My first test was to verify the avalanche ability of it. ... Both files were identical prior to encryption then. ... It means at least a 97% randomness to the ... You want DIEHARD on steroids: ...
    (sci.crypt)
  • Re: Test for randomness
    ... My first test was to verify the avalanche ability of it. ... Both files were identical prior to encryption then. ... It means at least a 97% randomness to the ... I may try some other compression methods for ...
    (sci.crypt)
  • Re: Disk encryption tools...?
    ... I chose AES-256 bit encryption algorithm. ... Even full version of BladeBox has less features than TrueCrypt. ... > no randomness, whereas the TrueCrypt volumes show sufficient randomness. ... > Many of you may have analyzed its source code. ...
    (sci.crypt)