I hacked up a quick simulation to see how an opennet's destination  
sampling algorithm would interact with the darknet location swapping  
algorithm.

I created a network of 10,000 nodes, each with up to 15 peers, and  
ran an opennet destination sampling algorithm where the probability  
of resampling the destination was 0.1.

I then ran it until the mean path length had dropped below 11, this  
typically happened between 80,000 and 90,000 requests into the  
simulation.

I then tried varying the amount of darknet-style location swaps, from  
one per request, to 1000 per request, here are the results:

Swap Every      Sim length
1                       90000
10                      99000
100                     87000
1000            89000

As can be seen, there is no observable trend where more frequent  
swaps seem to hurt routing, which supports my hope that location  
swapping and destination sampling can co-exist.

Of course, this is just some very simple early simulations.

Ian.

Ian Clarke: Co-Founder & Chief Scientist Revver, Inc.
phone: 323.871.2828 | personal blog - http://locut.us/blog

-------------- next part --------------
An HTML attachment was scrubbed...
URL: 
<https://emu.freenetproject.org/pipermail/tech/attachments/20060820/12fa2588/attachment.html>

Reply via email to