Regarding node specialization, I think that given enough nodes, duplicate
specialization areas should appear while covering a wide spectrum of keys.
On a graph of CHKs [x-axis] (if I understand things correctly) over
concentration [y-axis] of keys for multiple nodes, that would appear as
overlapping spikes. The width of the spikes (near the base) and the height
of the spike would depend on the capacity (keeps the base wide) of the node
and the specialization accumulated by each node (makes the spike high).

I heard a while back that somehow other nodes decide for a new node where
its specialization will be. Would it be possible (and useful) that these
other nodes would vote over spots that they see as unused?

Sorry if it's a dumb question.

Doc
_______________________________________________
devl mailing list
devl at freenetproject.org
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to