Scribd Logo
Upload

Welcome to Scribd!

  • 298 views

    Mathematical Attack On RSA

    Uploaded by

    kgskgm
    The document discusses several algorithms for factoring large integers, which is important for breaking RSA encryption. It explains Fermat's factorization algorithm, which works if the two prime factors are close in size. It also covers the Pollard p-1 algorithm, which works if p-1 or q-1 has small prime factors. Finally, it mentions more advanced techniques like the quadratic sieve algorithm, number field sieve, and elliptic curve method that can factor even larger numbers.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd

    Mathematical Attack On RSA

    Uploaded by

    kgskgm
    298 views5 pages
    The document discusses several algorithms for factoring large integers, which is important for breaking RSA encryption. It explains Fermat's factorization algorithm, which works if the two prime factors are close in size. It also covers the Pollard p-1 algorithm, which works if p-1 or q-1 has small prime factors. Finally, it mentions more advanced techniques like the quadratic sieve algorithm, number field sieve, and elliptic curve method that can factor even larger numbers.

    Original Title

    Mathematical Attack on RSA

    Copyright

    © Attribution Non-Commercial (BY-NC)

    Available Formats

    PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    The document discusses several algorithms for factoring large integers, which is important for breaking RSA encryption. It explains Fermat's factorization algorithm, which works if the two prime factors are close in size. It also covers the Pollard p-1 algorithm, which works if p-1 or q-1 has small prime factors. Finally, it mentions more advanced techniques like the quadratic sieve algorithm, number field sieve, and elliptic curve method that can factor even larger numbers.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    298 views5 pages

    Mathematical Attack On RSA

    Uploaded by

    kgskgm
    The document discusses several algorithms for factoring large integers, which is important for breaking RSA encryption. It explains Fermat's factorization algorithm, which works if the two prime factors are close in size. It also covers the Pollard p-1 algorithm, which works if p-1 or q-1 has small prime factors. Finally, it mentions more advanced techniques like the quadratic sieve algorithm, number field sieve, and elliptic curve method that can factor even larger numbers.

    Copyright:

    Attribution Non-Commercial (BY-NC)
    Download as PDF, TXT or read online from Scribd
    Download as pdf or txt
    of 5

    Fall 2006 Chris Christensen MAT/CSC 483

    Mathematical attack on RSA If we know ( n ) and the public key (the modulus n and the encryption exponent e), then we can determine d because d is the inverse of e mod n. We can use the Extended Euclidean algorithm (in Mathematica, ExtendedGCD[integer, integer]) to determine e. Then we can read the message. Now knowing ( n ) is mathematically equivalent to knowing p and q the two prime factors of n. Why? Well, certainly is we know p and q we know ( n ) because

    ( n ) = (p 1)(q 1). Conversely, if we know n and ( n ) , then

    n - ( n ) + 1 = pq - (p 1)(q 1) + 1 = pq pq + p + q 1 + 1 = p + q. So, we know n = pq and n - ( n ) + 1 = p + q. This suggests a quadratic equation which has p and q as its roots. X ( n ( n ) + 1) X + n = ( X p )( X q )
    2

    For example, if we know that n = 27153383 and ( n ) = 27142080, then solving (by using the quadratic formula, TI-92, or Mathematica) X 11304 X + 27153383 = ( X 7841)( X 3463 )
    2

    n = 7841*3463. So, knowing ( n ) and the public key (which allows us to break any message encrypted with that key) is equivalent to factoring n. There is no known efficient algorithm to factor large integers. So, how might we factor n? We know that n is composite (in fact we know that it is the product of 2 large primes). One of the factors of n must be less than or equal to division by all positive integers less than or equal to

    n . The brute force attack is to try n . This is not recommended.

    There is a factoring algorithm due to Fermat (1601 1665) that helps if the 2 primes are nearly the same size. Here is how it works.

    Fermat factoring algorithm The algorithm is based upon the being able to factor the difference of 2 squares. x y = ( x + y )( x y )
    2 2

    If n = x y , then n factors: n = ( x + y )( x y ) . But, every positive odd integer can be written as the difference of two squares. In particular for the integers that we use of RSA moduli n = pq,
    2 2

    p +q pq n = pq = 2 2
    2

    Let k be the smallest positive integer so that k > n , and consider k n . If this is a 2 2 2 square, we can factor n: if k n = h , then n = ( k + h )( k h ) . If k n is not a
    2 2

    square, increase the term on the left by one and consider ( k + 1) n . If this is a square,
    2

    n factors. If ( k + 1) n is not a square, consider ( k + 2 ) n . Etc. Eventually, we


    2 2

    n +1 n = n 1 . will find an h so that ( k + h ) n factors. That is so because 2 2


    2 2
    2

    In this case, n factors as n = n 1. k k + h

    n +1 2

    Here is an example. n = 6699557.


    2 2 2

    n 2588.35 ; so, k = 2589.

    k n = 2589 6699557 = 58 . So,


    2

    6688557 = 2589 58 = ( 2589 + 58 )( 2589 58 ) = 2647 2531


    2 2

    Fermats factorization algorithm works well if the factors are roughly the same size.

    Here is another example. n = 26504551. 5149 26504551 is not a square. 2 5150 26504551 is not a square. 2 5151 26504551 is not a square.
    2

    26504551 5148.26 ; so, k = 5149.

    . . . 5840 26504551 = 2757


    2 2 2 2

    So, 26504551 = 5840 2757 = ( 5840 + 2757 )( 5840 2757 ) = 8597 3083 . The Pollard p 1 factorization algorithm Lets just consider the case of interest factoring n = pq where p and q are large primes. This algorithm works well if either p 1 or q 1 is a product of relatively small primes. Lets assume that p 1 is the product of small primes. First, we guess an r so that p 1 divides r. Of course, in practice we will not know p, but for the moment assume that we do know p. It might be convenient to take r to be a factorial that is large enough that p 1 divides r. For example, 7001 1 = 23 53 7 ; so, 15! Would work because 23 , 53 , and 7 each divide 15!. 4536 1 = 23 34 7 ; so, 9! would work. But, 5869 1 = 2 2 32 163 ; so, we would need to use at least 163!. This is what we mean by saying that we want to take r = k! sufficiently large so that p 1 divides r. But, remember that because we wont know p we are guessing that we have chosen r large enough. If r is not large enough, the algorithm will fail to find a factor. Assuming that p 1 divides r, we can write r = (p 1) j. Also choose a positive integer a so that 1 < a < p-1. Once again, were guessing that our a satisfies this inequality. Say, we have r sufficiently large so that p 1 divides r and a so that 1 < a < p 1. Then notice that a r = a ( p 1) j = ( a p 1 ) = 1 j = 1mod p
    j

    What this tells us is that p divides a r 1 , and because n = pq, p divides gcd ( a r 1, n ) . another a.) So, thats the algorithm. After choosing r large enough so that p 1 divides r and a so that 1 < a < p, we calculate gcd ( a r 1mod n, n ) . (We have good algorithms for gcd and for modular exponentiation.) If we have chosen r and a correctly, gcd ( a r 1mod n, n ) = p . Example: Let n = 70348807, a = 2, and r = 13!. PowerMod[2, 13!, 70348807] 17662502 GCD[% - 1, 70348807] 7723 Which is one of the factors of n = 70348807. 70348807 = 7723 9109 . If we took r too small heres what happens. Say, r = 10!. (After the fact we can see that 7723 - 1 = 2 33 11 13 ; so 10! is too small.) PowerMod[2, 10!, 70348807] 60592434 GCD[% - 1, 70348807] 1

    In fact, gcd ( a r 1, n ) = p . (Unless, by chance, a r 1 = 0 mod n . In that case we choose

    Factoring RSAs public key consists of the modulus n (which we know is the product of two large primes) and the encryption exponent e. The private key is the decryption exponent d. Recall that e and d are inverses mod ( n ) . Knowing ( n ) and n is equivalent to knowing the factors of n. One attack on RSA is to try to factor the modulus n. If we could factor n, we could calculate ( n ) and (by using the extended Euclidean algorithm) determine d. Here are some factoring techniques: Trial division: Try all the primes that are n . Its not very elegant, but in theory it would work. The problem is that, like other brute force techniques, its not practical. Fermat (1601 1665) factorization: Not a bad technique if p and q are relatively equidistant from

    n.

    (1974) Pollard p-1 algorithm: Not bad if p 1 or q 1 is the product of small primes. (1975) Pollard -algorithm: The book discusses this algorithm. (1981) Pomerance quadratic sieve algorithm QSA: Still fast for up to around 110 decimal digits. (c 1993) Number field sieve NFS: The most efficient; based on work of Pollard (1988). (c 1987) Elliptic curve method ECM: H. Lenstra. (1994) Schors algorithm: Needs a quantum computer. The existing 7 cubit quantum computer has factored 15. The RSA factoring challenge numbers http://www.rsasecurity.com/rsalabs/node.asp?id=2092

    You might also like