I have seen this thread is suggesting the use of SVM/AI techniques to 
ascertain routing capabilities of peering nodes for each request. I may have 
missed something important at the start of this thread, but is SVM or AI in 
general really called for in this case?

I have done quite a lot of SVM work in the past, and I am not entirely sure if 
such a complex predictive method would be useful for Freenet routing. It 
strikes me that a simple moving average calculation would probably do just as 
well, and would be much simpler to implement. Simply see how well (how 
quickly) the nodes in the routing table have executed requests in the past 
for particular key classes and use that as the routing metric. A moving 
average with tunable window size would probably do just as well.

While using SVMs is all well and cool, it is easy to get stuck with the idea 
of using AIs because it _sounds_ like a really clever thing to do. Often, 
however, much simpler heuristic methods can give results that are not 
measurably worse and much easier to implement.

Generally speaking, if you can define reasonably clearly what constitutes a 
"good" node to route through (e.g. one that has returned similar keys quickly 
in the past, with reasonably consistency), then using an AI is probably over 
the top.

Regards.

Gordan
_______________________________________________
devl mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org:8080/cgi-bin/mailman/listinfo/devl

Reply via email to