> Another idea is to keep a buffer of the times taken to pass the last 100
> messages and calculate the time from that. That would keep a measure out for
> the current "weather" on Freenet, but it might be bloat and overkill.

I think this sounds like a very nice feature. My only concern is that this
type of feedback with nodes adjusting themselves based on the settings of
other nodes might be a chaotic system rather than a stabalizing
mechanism. We have to make sure the algorithm has checks to ensure that
things don't get out of hand with nodes raising their timeouts to
compensate for a slow network, therefore they themselves responding
slower, causing other nodes to raise their timeouts, etc..



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

Reply via email to