On Tue, Jul 23, 2002 at 06:09:30PM -0700, Ian Clarke wrote: > On Wed, Jul 24, 2002 at 01:14:58AM +0200, Oskar Sandberg wrote: > > Probabilistic selective caching should simply be based on the number of > > steps since the data was found / source was reset. This avoids all this > > silly abritrary behavior yet achieves much the same effect. > > Agreed. Didn't you do some simulations which confirmed the > effectiveness of this?
I confirmed an improvement, but I have never done any direct comparisons with reseting based on keyspace - the only simulation I know based on that is [1] which showed that it also improvemes performance if combined with some quota of "long distance" referenced and caches. [1] Hui Zhnag, Ashsish Goel, Ramesh Govindan: Using thge Small-World Mode to Improve Freenet Perfroamcen, IEE INFOCOM 2002 -- Oskar Sandberg [EMAIL PROTECTED] _______________________________________________ devl mailing list [EMAIL PROTECTED] http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
