NGR doesn't seem to be figuring variances of the estimators. is simply making random decisions. Worse than that, it
may be making non-random choices between equally good
alternatives, when a random choice is far better.
i can see what you mean here, and i don't strongly understand the estimators (math) yet !
But if the servers accept a few requests and do a good job on those, then the estimators in different clients' routing tables will diverge depending on which node served the request, and which QR'ed it. This is exactly what we want. NGR should be arranging for 1/10 of the clients to use each of the 10 qualified servers.
your last sentence is poorly worded and confusing. Did you mean, "each client should develop a unique affinity for one (or a few) of these 10 servers" ??
Also, NGR should be doing a statistical significance calculation, and when the difference between several servers is insignificant, it should choose randomly. This will cause different clients to favor different equally qualified servers, distributing the load.
Comments?
Ed - This is absolutely inspired ! I have wondered where in the system we should be introducing randomness. I'm probably unaware of many places where it is done, but this sounds highly appropriate to me! Now for the more experienced, expert inputs ...
ken
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
