Cryptography Reference
In-Depth Information
•
Second, the algorithm is extremely vulnerable to the most important
cryptanalytic approach, the plaintext attack (see above). Even if an
attacker knows nothing but the original (i.e., the plaintext) of an encrypted
message, he can recover at least large parts of the key — the remainder
results from 'idiomatic experiment', as described above in connection
with the Gold Bug example. Of course, this holds true for all substitution
methods. We will see later in this chapter that a piece of plaintext is
normally known in practice. A good encryption algorithm
must
resist
such attacks under reasonable assumptions.
Homophone substitution makes poor use of a language's inner rules. With
today's computer technologies, it would surely allow an adversary to mount
attacks even without knowing anything about the plaintext and about the key
creation. But I think this kind of stuff doesn't interest anybody anymore. The
weaknesses of this method are too serious, or are they
...
?
2.2.3 What If I First Compressed the Text?
...
you might ask. Software or hardware compression is matter-of-factly in the
computer age, because it cuts down file sizes considerably (e.g., by a factor of 3
or even 10) without losing any of their contents. The files can be decompressed
and their contents restored at any time.
Compressed files are distributed pretty much equally, i.e., all characters are
more or less equally frequent (that much for a jump into the computer age,
talking of the 256 values a byte can have rather than of 26 letters). Frequency
analysis doesn't do the trick any more. Does that mean more security? No, for
two reasons:
•
First, reconstructing the key during a plaintext attack remains as simple
as if the beginning of the message were known. You just have to first
compress the known beginning of the message and then compare the
compressed product with the ciphertext.
•
Second, though the compressed product appears to be equally distributed,
it is by no means random. To exploit the careless use of combined sub-
stitution and compression, I recommend proceeding as follows:
Every compression method places special information — so-called
magic num-
bers
— at the beginning to identify the method. A poor method would already
disclose the first elements of the key.
