Information Theory Rusty Nyffler Introduction n n Entropy

  • Slides: 10
Download presentation
Information Theory Rusty Nyffler

Information Theory Rusty Nyffler

Introduction n n Entropy Example of Use Perfect Cryptosystems Common Entropy

Introduction n n Entropy Example of Use Perfect Cryptosystems Common Entropy

Entropy n Measure of uncertainty n Entropy of a fair coin toss = 1

Entropy n Measure of uncertainty n Entropy of a fair coin toss = 1

Why is entropy important? n n n Measures the likelihood of the eavesdropper knowing

Why is entropy important? n n n Measures the likelihood of the eavesdropper knowing something about the key, given the ciphertext Greater entropy usually means greater security Entropy of a cryptosystem (K = number of possible keys (if all keys are equally likely)

Example n n 3 possible plaintexts: a(. 5), b(. 3), c(. 2) Two keys

Example n n 3 possible plaintexts: a(. 5), b(. 3), c(. 2) Two keys k 1 and k 2, equally likely (. 5) Possible ciphertexts U, V, W If someone obtains sent ciphertext, then information is known about the plaintext. ¡ ¡ If U was received, then a was the text If V was received, then b or c was the text

Example Continued n Moreover, ¡ p(b|V) =. 6 ¡ p(c|V) =. 4 ¡ p(b|W)

Example Continued n Moreover, ¡ p(b|V) =. 6 ¡ p(c|V) =. 4 ¡ p(b|W) =. 6 ¡ p(c|W) =. 4 ¡ n Which means it’s more likely that b is the plaintext when V or W is caught Entropy shows that the eavesdropper knows more about the plaintext when the ciphertext is intercepted

Perfect Cryptosystems n n n In a perfect cryptosystem (i. e. unbreakable), the ciphertext

Perfect Cryptosystems n n n In a perfect cryptosystem (i. e. unbreakable), the ciphertext should not give any more information about the plaintext or the key H(P) = H(P|C) H(K) = H(K|C)

Common Entropy n The English language ¡ ¡ n Somewhere between 1. 42 and.

Common Entropy n The English language ¡ ¡ n Somewhere between 1. 42 and. 72 English is around 75% redundant In many algorithms, the longer the key, the more entropy the system has ¡ Vigenere

Common Entropy n RSA ¡ ¡ Entropy = 0 All the info you need

Common Entropy n RSA ¡ ¡ Entropy = 0 All the info you need is in n, e, and c … except it’ll take a while to factor n

References 1. 2. 3. 4. Mac. Kay, David J. C. Information Theory, Inference, and

References 1. 2. 3. 4. Mac. Kay, David J. C. Information Theory, Inference, and Learning Algorithms. 18 April 2003. http: //www. inference. phy. cam. ac. uk/itprnn/book. l. pdf North Carolina State University. Cryptography FAQ. 21 March 2003. http: //isc. faqs. org/faqs/cryptography-faq/part 01/ Raynal, Frederick. Weak Algorithms. 2002. http: //www. owasp. org/asac/cryptographic/algorithms. shtml Trappe, Wade and Lawrence C. Washington. Introduction to Cryptography with Coding Theory. Upper Saddle River, New Jersey: Prentice-Hall, Inc. , 2002.