On Sunday 26 November 2006 11:23, Michael Rogers wrote: > Ed Tomlinson wrote: > > Throtting alone seems to work better though. I wonder if we should disable > > the > > backoff logic and see what happens. If we trust your simulations, we > > should not > > see much of a difference at all. > > I haven't had a chance to investigate heterogeneous capacity yet - > backoff might be useful for avoiding unusually slow nodes without > throttling the whole network to the speed of the slowest nodes.
I think the above question is a very very important one. The other aspect of it is how does non symmetric bandwidth affect the network? eg. if we are basing send rates on average recieve rates from nodes with much less output bw than input we may be wasting lots of bw... What would happen if we only sent to one peer at a time (queueing other messages) or what would happen if we used peek recieve rates instead of average? Thanks Ed
