[RC5] RC5 question
Décio Luiz Gazzoni Filho
decio at revistapcs.com.br
Fri Oct 3 20:03:10 EDT 2003
-----BEGIN PGP SIGNED MESSAGE-----
On Friday 03 October 2003 18:48, Daniel Quintiliani wrote:
> On Fri, 3 Oct 2003 18:10:33 -0300, Décio Luiz Gazzoni Filho
> <decio at revistapcs.com.br> wrote:
> > The problem is not export restrictions or anything (they've been lifted,
> > as far as I know); it's just the sheer amount of work that has to be
> > done. A book of mine states that since the inception of the computer
> > through the year 2000, there have been approximately a mole (6e23) of
> > computer operations performed on all computers on Earth. In comparison,
> > there are 3.4e38 RC5-128 keys to check. Even taking into account that
> > only 50% of keyspace is expected to be checked on average, that's 1.7e38
> > keys, or 1.7e40 cycles. Dividing out 1.7e38 by 6e23, you get a number
> > which is on the order of magnitude of a million times the age of the
> > universe.
> Still, it's more fun than the RC5-72 since it's more relevant. According to
> the stats today the keyspace of RC5-72 will reach 100% in 462,417 days.
> Divided by 365.25 is about 1266 years. (Half that is about 633 years.) So
> both of these are not very good as contests but are fun "to work towards"
> as a game.
If you factor in new participants and Moore's Law (which dictates an
exponential growth of processing power with time), RC5-72 is doable, whereas
RC5-128 is utterly hopeless, short of a very clever attack on the RC5 cipher.
Frankly I don't see the point of running a contest that has absolutely no
chance whatsoever of finishing. Then again, The Neo Project was doing just
that with the RSA factoring contests, albeit in a much larger scale (meaning
it was even more hopeless), and people still helped them.
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.3 (GNU/Linux)
-----END PGP SIGNATURE-----
More information about the rc5