0
$\begingroup$

Using CrypTool I encrypted a plaintext with "DES CBC" and I genereated the histogram of plaintext and ciphertext. The histogram shows the relative frequency of each of the characters in the document.

When I did the same with "Caesar" cipher based on online tutorials, it was possible to find the correlation between some characters comparing their frequency that I could see in their histograms.

But that it is not possible for DES. I try to find out why this is not possible. What properties of the DES are responsible for this?

$\endgroup$
3
$\begingroup$

Lookup avalanche effect either in wiki or our tags here. Essentially, the slightest change (any single bit) of the plain text leads to 50% of all the other bits (on average) changing across the full width of DES' 64 bit block size. This totally obfuscates any correlation between the plain text characters.

"Caesar" cipher doesn't exhibit avalanche being only a substitution. It simply keeps the correlations and encodes them to other letters.

| improve this answer | |
$\endgroup$
0
$\begingroup$

Studying the frequency of encrypting characters in order to link them to the corresponding clear-character is called frequency analysis. It's a very old and pretty basic cryptanalysis technique, so it's not a surprise if it doesn't work on modern crypto-systems.

It will only work on very basic substitution cipher, ie when your cipher consists on replacing each character by an other. If you take a look at how DES process a message, you can see that it's a bit more elaborated.

| improve this answer | |
$\endgroup$
0
$\begingroup$

In the histogram of the plaintext, the frequency of the characters are non zero only for specific characters. But in the histogram of the ciphertext, the majority of the characters has frequency non zero and we can see a continiously increase and decrease of the frequencies starting from the first character to the last.

So in a few words, this is because of the diffusion property which refers to the redistribution of the plaintext bits so that any surplus in it is scattered in the ciphertext. Is that corret?

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.