[freenet-dev] [freenet-cvs] r16756 - trunk/freenet/src/freenet/node

[Matthew Toseland]

On Friday 21 December 2007 17:39, Robert Hailey wrote:
> 
> On Dec 21, 2007, at 8:40 AM, Matthew Toseland wrote:
> 
> > Not a good idea. HTL doesn't monotonically decrease - sometimes it  
> > doesn't
> > decrease (at 10 or 1), and it is reset whenever we get closer to the  
> > target.
> 
> So then, the standard freenet algorithm is to:
> (1) find the closest node,
> (2) search with a depth-radius of 15 hops (HTL).

10. HTL starts at 10, it's reset when the best-so-far improves, and it's 
decremented probabilistically at 10 and 1 (whether or not to decrement is 
determined at boot up time for each peer for various reasons).
> 
> What is the depth (width?) of the network currently? If (as in the  
> small-world theory) all nodes are connected at depth==6, wouldn't this  
> amount to searching every node on the network?

"small world" isn't quite that simple. :) Small world graphs should have small 
diameter, but not necessarily that small, and they do have other properties 
e.g. mostly short links, friends tend to link to friends' friends, etc.
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 189 bytes
Desc: not available
Url : http://emu.freenetproject.org/pipermail/devl/attachments/20071221/dfed7ae6/attachment.pgp