Cryptography Reference
In-Depth Information
quadrillions of possibilities. I spare myself the calculation how long the fastest
computers in the world would take to test all of these keys.
Nevertheless, this method isn't worth much either. It can be broken effortlessly
by statistical analysis. Instructions can be found in Edgar Allan Poe's famous
novel, The Gold Bug , perhaps the first popular work on cryptanalysis. Poe
explains very vividly how cryptanalysts work: revealing information step by
step as they exploit every particularity.
I don't want to repeat the passage concerned from the topic here in detail. You
can either read it there or in [BauerMM, 15.10]. But let's have a quick look at
the keys used; it's worthwhile:
Assume we want to decode a ciphertext 203 characters long, consisting of
numbers and various typographic special characters. (It doesn't actually matter
whether letters are substituted by letters or other characters. The main thing is
that the substitution is reversible.)
Knowing the code writer, the cryptanalyst concludes that he has surely
used just a simple substitution. Bear in mind: you always have to assume
that your adversary knows the method you used.
First, the analyst will search for the most frequent character — that's
'8' — and assume that it corresponds to 'e', which is the most frequent
letter in the English, German, and other alphabets (see Table 2.1).
Table 2.1 Frequency analysis for the first chapter of the German edition of this topic
The 10 most frequent letters and
characters
The 10 most frequent pairs of letters and
characters
13.78 % ' '
13.17 % 'e'
8.09 % 'n'
6.65 % 'i'
5.67 % 'r'
5.17 % 't'
4.39 % 's'
4.03 % 'a'
3.77 % 'h'
2.99 % 'l'
total: 66.7 %
average frequency of a letter: 3.85 %
3.11 % 'e'-'n'
2.65 % 'e'-'r'
2.57 % 'n'-' '
2.35 % 'c'-'h'
2.18 % 'e'-' '
1.56 % 'e'-'i'
1.54 % 'r'-' '
1.49 % 't'-'e'
1.47 % 'i'-'e'
1.35 % ' '-'d'
total: 20.3 %
average frequency of a pair: 0.0015 %
Search WWH ::




Custom Search