>I have been experimenting with a node for several days now and have been
>experiencing incredibly slow transfers through freenet particularly on
>larger files.  It appears that several of the other nodes that I talk with 
>are severely bandwidth challenged.  Freenet is unlikely to be a viable 
>online resource unless a method is built in to the routing process which 
>lends prefference to connections with higher transfer rates.  For those of 
>you who are more technical than I am, would it be feasible to include a 
>transfer rate metric in the data that is stored for each connection and 
>could such a metric be used to direct traffic to faster servers.
If instead of the current best route strategy there would be a probability 
routing, coupled with a speed value of routes. Storing the round-trip value 
at each known route would add a lot more meaning the probability routing. 
The probability of a routing into certain direction should be calculated 
from closest routing found multiplied with that routings roundtrip value, 
and the lower total the better. If a routing probability fails it should 
either route into a random direction, or perhaps the second best should be 
tried. Either way this should allow finding new routes to hosts, aswell as 
routing more traffic into faster directions.

I'm kinda sleepy so I hope you excuse my flood of typo's :-D


--typo
________________________________________________________________________
Get Your Private, Free E-mail from MSN Hotmail at http://www.hotmail.com


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

Reply via email to