making our users continue to suffer from the lack of an effective load balancing mechanism?
Just so that we are all on the same page, we should clarify our definition of load-balancing. I believe there are two components to this definition -
1) query traffic versus data traffic needs to be managed, in so much as saying query traffic should not adversely affect data traffic (the QR overload we are addressing)
2) a node has multiple requestors. Each requestor is entitled to a 'fair share' of the the requestee. Determination of that fair percentage may need to take bandwidth, per "link," into account
These are two distinct problems, and I believe problem #1 is currently the more urgent obstacle, but both are important. If this is agreed, can we focus on #1 until it is solved ? They are clearly related, but #1 has an obvious fault to address. Regarding #2, it would behoove us to begin collecting measurements on a per node basis, even if we don't agree how to utilize them just yet... can someone suggest where in the code this would best be located ? PeerHandler ? I apologize for insufficient RTFS here.
Ken
_______________________________________________ Devl mailing list [EMAIL PROTECTED] http://dodo.freenetproject.org/cgi-bin/mailman/listinfo/devl
