One Time Pad Algorithm is a simple algorithm used to cipher some message. Given a key k and a message m, the algorithm will calculate the cipher text by c = m xor k.
Both sender and receiver will share the same key k. Therefore it is easy for the receiver to decipher the cipher text by m = c xor k = m xor k xor k = m.
For more information of One Time Pad algorithm, you can refer to this link.
One Time Pad Algorithm is extremely safe if the key k is used only once and will be replaced by another key in next communication. If the convention is kept, it is generally impossible to decipher the plain text by Ciphertext-only attack
.
However, if the key k is used more than once, the attackers can then observe some patterns from only the cipher texts.
The way the attacker can do is to xor the cipher texts together and observe the result.
For example, consider three cipher texts c1, c2 and c3 which are generated with the same key k. That means c1 = k xor m1, c2 = k xor m2 and c3 = k xor m3.
If we xor c1, c2 and c3 in pairs, we can get some information about the plaintext since c1 xor c2 = k xor m1 xor k xor m2 = m1 xor m2.
Similarly, c2 xor c3 = m2 xor m3 and c1 xor c3 = m1 xor m3.
If we have some previous knowledge about the message, for example all the messages contain only [a-zA-Z] and space, we can use the ascii table to filter the result and limit the number of possible solutions.
This project will also adopt this method to decipher the cipher text.
The cipher.py accepts a file that contains the message needed to be ciphered. The first line is the number of messages. I have provided an example message file plain_text. To use the cipher, you can simply run cipher.py < plain_text > cipher_text in the command line. The key is generated randomly with python inherit random.randint() function.
The decipher.py will accept a file that contains the cipher texts. The first line of the file is the number of cipher texts (in my example, 7).
Simply run decipher.py < cipher_text > decipher_result in the command line and the decipher result will be stored in the file decipher_result.
There is an argument full to decipher.py. If full is set to 0, all the characters that cannot be determined (more than one possible characters or the character list not covering all character in the message) will be represent as *. If full is set to 1, then the characters that cannot be determined exactly will be listed in curly braces. full is set to 0 by default. To set it to 1, use --full=1 in the command.
To modify the set of possible characters in messages, simply modify the variable chars on line 1 in the file util_function.py. Each character in the string represents a possible character.
If you have any questions or better ways to solve this problem, you can raise an issue or send an email to Yiwen(yiw.che@gmail.com)