[RC5] Vote to exhaust DESIII keyspace
Mike.Burgess at trw.com
Thu Dec 24 08:44:02 EST 1998
My thoughts interspersed in the original
I would not want to live in a cartoon world...
Too many people have too much access
to too much dynamite.
:) -----Original Message-----
:) Actually, we're quite aware as to what will happen to the
:) network upon full completion of a subspace.
I think the original ? was not the end of a keyspace, but the
end of all keyspaces.
:) First, the master will get into a very tight recycling
:) loop. It will have fewer and fewer available blocks and
:) will continuously recycle un-returned blocks until they have
:) all been completed.
We should hit the KEY real soon at this point.
:) At this point, the master will notify
:) us that it is out of DES-III keys. The proxies will then run
:) out of outgoing DES-III blocks and clients will switch back to
It would be nice to see what happens when THE key is found, even if
the contest was over before Distributed finds the key. Until the
the code is real-life tested, it's only a theory of what happens.
(ever hear of Bug Fixes?) Who, except microsoft, releases
known buggy code?
:) I still plan to disable DES as soon as the solution is
:) found. This will notify the proxies that the DES contest
:) is closed and they will then notify clients and pproxies.
What if there is a flaw, and we don't return a "hit" on the correct
key when we crunch it? We'd never know for sure. On a short contest,
what's a few extra days to let it run to completion. At the least,
there would be one or two D.Netters that could brag "my CPU found the
key". no payoff, just bragging rights and enthusiasm generated
:) I can see no benefit to completing the full DES keyspace;
:) all it will do is cause more load on the proxy network and
:) on the keymaster.
It's a CPU, load it down, stress it a bit, see if the pgm breaks,
if heatsunk properly, the die (CPU) won't melt. It should be
built to take a load. [ No sir, I can't pull your motorbike
out of that ditch, my truck might not like that kind of a load ]
To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest
More information about the rc5