06.01.2015 Views

Cryptology - Unofficial St. Mary's College of California Web Site

Cryptology - Unofficial St. Mary's College of California Web Site

Cryptology - Unofficial St. Mary's College of California Web Site

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

152 CHAPTER 8. POLYALPHABETIC CIPHERS<br />

Using mathematics similar to that used in Friedman’s Index <strong>of</strong> Coincidence,<br />

given enough text any Auto Key cipher can be broken. To perhaps explain<br />

how, notice that the plaintext-key pairs (t,O), (n,I), (s, N) and (i, D) all give<br />

rise to the ciphertext letter V. But this is the only way <strong>of</strong> creating V from two<br />

high-frequency letters. So if we suspect that an Auto Key Cipher with priming<br />

keyword <strong>of</strong> length 1 was used, then we can try each <strong>of</strong> these as the possible<br />

plaintext–key pair on each V’s in the ciphertext. By probability, we very quickly<br />

will find one that gives the correct decipherment. Similarly, the pairs re and<br />

er will occur <strong>of</strong>ten in the plaintext, producing a large number <strong>of</strong> W’s in the<br />

ciphertext. Thus an Auto Key cipher with a priming key <strong>of</strong> length 1 is not<br />

secure.<br />

Likewise, every the in a Auto Key cipher <strong>of</strong> keylength two will give many<br />

e’s enciphered by t’s into Y’s. In general, the etaoinshr letters so frequently<br />

encipher one another they give each other away, leading to a decryption <strong>of</strong> the<br />

cipher.<br />

To summarize, we broke the Vigenère Cipher by exploiting the pattern <strong>of</strong><br />

its repeating keyword. The Auto Key Cipher can be broken by exploiting the<br />

usual frequency patterns <strong>of</strong> English. Removing all such patterns must be our<br />

next goal.<br />

8.6 Perfect Secrecy<br />

All the ciphers we’ve studied so far depend, eventually, on some sort <strong>of</strong> pattern,<br />

and this pattern eventually gives them away. What is needed is a cipher system<br />

whose keyword is both endless and and senseless. The need for endless should<br />

be clear after our work with Vigenère ciphers. Once a keyword starts to be<br />

repeated, the cipher is in danger <strong>of</strong> being broken. Hence, for perfect secrecy we<br />

can never allow the it to be repeated, i.e., it must be endless. The reason for<br />

senseless is nearly the same: an extremely long keyword that is not senseless<br />

has some pattern to it. And a pattern is not much different from a repeating<br />

keyword. As Kahn puts it, the perfect cipher must “avoid the Scylla <strong>of</strong> repetition<br />

and the Charybdis <strong>of</strong> intelligibility.”<br />

Joseph O. Mauborgne (1881–1971) had the idea <strong>of</strong> “endless”. Mauborgne<br />

had a long and very distinguished career in cryptography. In 1914 he gave the<br />

first recorded solution <strong>of</strong> a Playfair cipher. (We will study these in Chapter 9.)<br />

He eventually rose to the post <strong>of</strong> Chief Signal Officer in October 1937 and as a<br />

Major General built the cryptanalytic abilities <strong>of</strong> the Signal Corps to the extent<br />

that it was reading a flood <strong>of</strong> Japanese ciphers by his retirement in 1941.<br />

Gilbert S. Vernam (1890-1960) had the idea <strong>of</strong> “senseless”. Vernam was an<br />

employee <strong>of</strong> AT&T when in 1917 he proposed the use <strong>of</strong> “stream cipher devices”<br />

for automatic encryption and decryption <strong>of</strong> telegram messages. Vernam received<br />

65 patents in the areas <strong>of</strong> cryptography and telephone switching systems and<br />

was well known for his cleverness – supposedly he asked himself “What can I

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!