I think it's important to understand that there's not really such a thing
as "behind" here, because we're not traversing the keys in the same order.
What those stats mean is that every key we check now has a 1 in 0.91 * 2^56
chance of being the right one, and every key they check now has a 1 in
0.933 * 2^56 chance of being the right one. It's very possible they'll find
it first even if we've eliminated a lot more keys than they have, if the
order of their key search happens to check the winning key first.
For that matter, say we've searched 99% of the keyspace. A key picked at
random from the remaining 1% of the space would be 100 times more likely to
be the winning key than one picked out of the entire keyspace. But if one
guy, all on his own, just started checking random keys right then, it's
possible he would make a guess that was much better than the random
assignment of keyblocks that deschall was doing, and he'd find the key
before we could search the last 1% of the keyspace.
Not very likely, but possible. Thinking of this in terms of a race maybe
isn't the best analogy, though. Regarding SolNET in particular, I think the
most telling stat on their page is the graph of their key rate over the
last 24 days, which has stayed pretty flat for the last 12. Meanwhile,
DESChall has roughly followed an exponential curve. It would appear they've
either stopped their evangelism efforts or just plain saturated the pool of
available foreign computing resources before we ran out of American
computers. The DoD is right about the similarities between guns and
encryption software in one very important way -- America has the most of
both of them! :)
____________________________________________________________________________
Andrew Meggs, content provider Antennahead Industries, Inc.
<mailto:insect@antennahead.com> <http://www.antennahead.com>