This preview shows pages 1–2. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.View Full Document
Unformatted text preview: CS 170 Algorithms Fall 2009 Christos Papadimitriou HW 2 Solutions 1. (10 pts.) RSA Problem 1.42 Solution We know that p is a prime and gcd ( e , p- 1 ) = 1. Using the extended Euclidean algorithm find integers x , y such that xe + y ( p- 1 ) = 1. We then have that m xe + y ( p- 1 ) = m 1 ( mod p ) and using Fermat’s little theorem we get m xe = m ( mod p ) , i.e. ( m e ) x = m ( mod p ) . So the algorithm is as follows: Input: p , e , m e mod p Compute x , y such that xe + y ( p- 1 ) = 1 using extended euclidean algorithm (in Θ ( log ( p- 1 )) steps). Compute and output ( m e ) x mod p (in Θ ( log 3 p ) steps). The correctness of the algorithm follows from the previous discussion and its running time is Θ ( log 3 p ) , clearly polynomial. 2. (15 pts.) RSA Problem 1.43 Solution We first note that we must have 3 d = 1 ( mod ( p- 1 )( q- 1 )) or equivalently 3 d- 1 = k ( p- 1 )( q- 1 ) for some k ∈ Z ; we will try to find this k . First note that d < ( p- 1 )( q- 1 ) (since it is the inverse of 3 modulo ( p- 1 )( q- 1 ) ); hence kd < k ( p- 1 )( q- 1 ) = 3 d- 1 < 3 d ⇒ k < 3. Moreover we have that k > 0 because 3 d- 1 > 0. So the only possible values for k are 1 and 2. Notice now that p 6 = ( mod 3 ) (since p is prime) and p 6 = 1 ( mod 3 ) since 3 has an inverse modulo ( p- 1 )( q- 1 ) and therefore it has to be relatively prime to p- 1. So the only option left is p = 2 ( mod 3 ) . Likewise q = 2 ( mod 3 ) , so we have ( p- 1 )( q- 1 ) = 1 ( mod 3 ) . Suppose now that k = 1; then ( p- 1 )( q- 1 ) = 3 d- 1 = 2 ( mod 3 ) , a contradiction. Hence k = 2. So now we have two equations involving the two unknowns p , q , namely: 3 d- 1 = 2 ( p- 1 )( q- 1 ) and pq = N . We can now solve the first one to get p + q = N + 1- 3 d- 1 2 , S and end up with two formulas of the form p + q = S and pq = N . We can now find p , q as the roots to the equation x 2- Sx + N = 0....
View Full Document
- Spring '09