[rc5] We may only need 5 years...

The Man theman at fallout.dyndns.com
Thu Oct 23 14:25:22 EDT 1997

Eric Gindrup wrote:

>      By eyeball at http://home.pacbell.net/cwax/bovineg.html, the Bovine
>      effort seems to double its number of completed keys count every 2.5
>      weeks.  Current keyrate is 5Gkps.

If the number of completed keys doubles every 2.5 weeks, then 2.5 weeks
before the keyspace is exhausted, only half the keys will be checked.
This implies that 2^63 keys must be checked in the last 2.5 weeks, or an
average rate of 6.1e+12 keys per second near the end of the contest.
This is about equivalent to 9.2 million PII/266's, and would represent a
thousandfold increase over our current rate. To assume that kind of
computing power even exists, much less can be recruited, is excessively
optimistic. You can certainly use the function of your choice to model
our growth rate, but somewhere the following must be considered:

Finite number of computers in existence
Greatest cracking speed one can expect from ANY computer
Finite number of computers connected to the net
Finite number of people willing to participate

Considering the current rate and allowing for some reasonable level of
improvement due to more participants and Moore's Law might yield a
result somewhere between 5 and 20 years. If you like unlimited optimism,
it is also possible that my client found the winning key as I wrote
this. We can choose any function we want to model our growth, but
everyone is going to have to realize that the 64-bit keyspace is
absolutely monstrous and this is not likely to be a months-long project
by any means. 

Keith Wesolowski		theman at fallout.dyndns.com
University of Nevada - CS/EE	Visit The Fallout Shelter at
The Fallout Shelter - North	http://fallout.dyndns.com/
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.

More information about the rc5 mailing list