[rc5] keyspace flaw

Benedikt Eric Heinen beh at icemark.ch
Tue Jun 3 20:30:36 EDT 1997


> >And finally, because the bovine servers don't recycle keys, it's not
> >even like you can be sure that a majority of the keys in the first
> >1.7% of the keyspace have been searched...
> The keymaster actually does recycle keys, but it does so on over a longer
> time period--when the keyspace wraps.  When it does wrap, it will then
> begin redistributing keys that notifications have not been received back
> for. 
Hmm, since most people voting towards the sequential approach (so that
sharing the effort becomes possible), shouldn't at least one proxy be
devoted to recycle keys? I mean, if two or more competing efforts finally
closed the whole keyspace and then need to go back through their logs to
find out which blocks are still unsolved -- sharing those might then
become a problem, as at first people would be required to buid a list of
those blocks left.

What would you think about building a proxy server dedicated to recycling
key blocks?

If I look at the current executive stats, we have done 1.77% of the
keyspace, the last block that was assigned to my machine was 0x0CE3327,
which would indicate that 5.03% of all possible blocks weren given out -
1.77% reported back, 3.26% didn't. That leaves enough distance between 0
and the current block to build one proxy server that just recycles
keyblocks, without having to fear that it'll start handing out blocks that
were only given out 5 minutes ago... ;)

At the beginning that server would only be required to know those blocks
that reported back up to now and would also need to get notification for
eqach block that has been reported in.

Any comments?

  Benedikt


signoff

          Hiroshima '45           Chernobyl '86           Windows '95

----
To unsubscribe, send email to majordomo at llamas.net with 'unsubscribe rc5' in the body.



More information about the rc5 mailing list