lecture 2 Flashcards
What is the Kasiski examination and what is it good for?
Kasiski examination is useful for breaking vigenere ciphers. It looks for repeated trigrams, gaps between them are likely to be multiples of the key length.
Is the Friedman test useful for other languages? How come?
The Friedman test(breaking text into columns and checking letter distributions) works on anything with an underlying letter distribution that is not uniform, as such it can work on any language.
What is a probability distribution?
A finite set is a probability distribution if the sum of probabilities of all possible evens is 1, and the probability of each outcome is greater than or equal to 0.
What is a uniform probability distribution?
One in which the probability of every outcome is 1`/ the total number of possibilities.
What is a union bound?
The probability of at least one of two or more outcomes happening together is the probability of one happening plus the probability of the others.
What is independence in statistics?
When two outcomes have no effect on the probability of the others.
What is typically used for binary sytems for encoding and decoding? Why is it useful?
XOR, this is commutative, xoring something with itself gives all 0s, the order doesn’t matter, and if a xor b is c then c xor b is a.
The key is independent from the message, and could be anything, as such the probability of any cipher message is uniform.
Why can massive keys be difficult to use?
transferring them can be difficult due to the large size.
What is Shannon’s observation? What is perfect secrecy?
If for any two messages and one ciphertext, the number of possible keys for each message to become the cipher text is the same, then the attacker can learn nothing with only ciphertext, this means the cipher has perfect secrecy.
What is a stream cipher?
A stream cipher uses a pseudo-random key, this means it uses a pseudo random generator, and a seed to produce a much longer sequence. If we agree about the seed then we will have the same long sequence of bits. This is no longer perfectly secret because there are far less possible seeds.
How can we tell if a pseudo random generator is good?
If after some time an attacker can not predict the next bit with probability significantly different from 0.5 then the generator is quite strong.
Why can using the same XOR key twice be bad?
Two messages can be XORed to allow the original message to be worked out with frequency analysis.
What is one way an attacker could modify a message based on XOR?
Modify the ciphertext by XOR with some string p, this may be gibberish, but still compromises the message.