We argue that, contrary to common belief, it is unnecessary to use strong primes in the RSA cryptosystem. That is, by using strong primes one gains a negligible increase in security over what is obtained merely by using ``random'' primes of the same size. There are two parts to this argument. First, the use of strong primes provides no additional protection against factoring attacks, because Lenstra's method of factoring based on elliptic curves (ECM) circumvents any protection that might have been offered by using strong primes. The methods that 'strong' primes are intended to guard against, as well as ECM, are probabalistic in nature, but ECM succeeds with higher probability. For RSA key sizes being proposed now, the probability of success of these methods is very low. Additionally, the newer Number Field Sieve algorithm can factor RSA keys with certainty in less time than these methods.
Second, a simple group-theoretic argument shows that cycling attacks are extremely unlikely to be effective, as long as the primes used are large. Indeed, even probabalistic factoring attacks will succeed much more quickly and with higher probability than cycling attacks.
Category / Keywords: public-key cryptography / Date: received 30 Jan 2001 Contact author: rsilverman at rsasecurity com Available format(s): Postscript (PS) | Compressed Postscript (PS.GZ) | BibTeX Citation Version: 20010130:184629 (All versions of this report) Discussion forum: Show discussion | Start new discussion