[RC5] stop running the client?
rmalayter at bai.org
Mon Dec 13 00:12:39 EST 1999
I just read an interesting paper at
The premise of this paper is that, for long-running calculations (like the Rc5 crack) it is often better to WAIT to start it on newer hadrware, beacuse you'll actually finish the task sooner.
This is because of Moore's law. The author's calculations indicate that if any task will take longer than 26 months to complete on current hardware, you're better off waiting a wile and running it on new hardware.
Pretend for a moment you are the NSA, and you have $10,000,000 to spend on cracking a particularly sensitive communication from Beijing. You might actually finish that crack SOONER if you put the purchase of the cracking hardware off for about two years, and started then.
Obviously, this doesn't apply to d.net effort directly, because the d.net folks don't purchase their compute power in a traditional manner. New compute power is being continually added to the d.net pool, and there is no discrete, point-in-time hardware purchase.
However, this paper could possibly be used as an argument for fitting in more short-term projects into d.net and letting Moore's law take care of RC5-64... because it could be shown that the target date for RC5-64 keyspace exhaustion would only slip by a week or two if we ran another CSC or an ECC crack.
Anyone want to do the math and see how much a project like CSC is really setting us back on RC5, including d.net's participant growth rate and Moore's law ? I'm far enough out of Math 265 that I can't think of where to start without buying a Calculus textbook...
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5