[RC5] Active subspaces/random subspaces

Edwin ten Dam etd at st-andrews.ac.uk
Thu May 7 13:20:47 EDT 1998

Thus spoke the keyboard of Jim C. Nasby, nasbjim at charlie.cns.iit.edu on 
06/05/98 17:45:

>The basic reason for this is that the clients are coded to use the
>current subspace+1 as the random subspace. It might be more space
>efficient to issue out of random keyspaces (ie: switch from 0x64 to 0x65, 
>instead of 0x66), but it would also increase the dupe rate durring
>switchover periods. Also, redoing random subspaces right now isn't an
>option, since that would make a completed subspace the random one, meaning
>that almost all randomly generated blocks would be dupes.

Hmm. if I understand this correctly, there are a few problems with this. 
The random subspace (=current subspace+1) will never be finished at the 
same time as the current subspace; therefore you'll have to keep track of 
any subspace that once was random. So, (say curent subspace is even, 
random =current+1), for every step in current subspace we make, we'll 
gain an unfinished random subspace to keep track of. And what happens 
when we start finishing of these random subspaces? The then current 
subspaces (odd) will direct the random subspace to be the even subspaces, 
which are already done. Then there is a problem with take subspaces off 
the severs; by their nature randomly generated blocks will be from 
computers that have not been in contact for some time, they can't be 
relied on to be working in the appropriate random subspace.

Maybe we should stick to the two current random subspaces (0x67 and 0x69) 
somehow, and when we get too many dupes in for one make that a regular 
one and take a new random subspace.



Edwin ten Dam                                   etd at bigfoot.com
School of Biomedical Sciences     -    University of St Andrews
Irvine Building, North Street,  St Andrews, Fife KY16  9AL U.K.
Phone: +44 1334 463404                   - Fax: +44 1334 463400

To unsubscribe, send 'unsubscribe rc5' to majordomo at lists.distributed.net
rc5-digest subscribers replace rc5 with rc5-digest

More information about the rc5 mailing list