I don't think the problem with searching is "exponential growth" of the
number of nodes searched (with increasing HTL).

Rather, it is "N squared growth" in the load on the network, where N is
the number of nodes.

With smart searches, the load on the network grows linearly with the
number of nodes, as long as HTL remains constant.  You put ten times as
many nodes on the system, ten times the users, and you get ten times
the traffic.  (Actually you may have to increase the HTL somewhat
for larger networks, so growth may be a little worse than linear.)
This makes traffic per node approximately constant, independent of the
size of the network.  That's scalability.

With broadcast searches, the load on the network grows as the square
of the number of nodes, again assuming HTL remains constant.  With ten
times more nodes, you get 100 times the traffic.  That means that traffic
per node increases proportionally to the number of nodes in the system.
That's not scalable, because at some size the load per node will simply
be too great.

I don't think the various optimizations suggested by Brandon will change
this fundamental scaling factor.  Even with caching and limited HTL
values, you are still searching a certain percentage of the network,
and there will still be N squared traffic load.  Unless you can show
that the percent of the network searched will DECREASE significantly as
you go to larger networks, it won't change the basic conclusion.

I think you need some quantitative analysis, even rough back of the
envelope figures, to show that a broadcast search is scalable.

Hal

_______________________________________________
Freenet-dev mailing list
Freenet-dev at lists.sourceforge.net
http://lists.sourceforge.net/mailman/listinfo/freenet-dev

Reply via email to