On 25 Sep 2006, at 04:47, toad wrote:
So we force the distribution? Hmmm. We don't have to force the
distribution if we path fold on successful requests, but forcing it is
better in that it will almost certainly work regardless of e.g. local
traffic levels.
It can't hurt.
I'm not sure about the "probability proportional to 1/d" bit, now that I
think about it; doesn't that mean that there will be infinite
probability that a point infinitesimally different to the current
location will be chosen?
There is some formula to randomly select numbers in the distribution we want, ask Oskar.
Also, I'm not sure about NCNUM; if we are forcing the distribution, more
would result in inaccuracy. If we avoid this by returning the NCNUM
nodes closest to the target, we'll have a lot of nodes: I don't see why
we should have hundreds of node connections on opennet, it's not
necessarily better and it's a sure-fire way to mess up a darknet opennet
hybrid.
Yeah, taking the NCNUM nodes closest to the target location would be better.
As far as I know, what oskar said would work was destination folding on
a successful request, and I can see a number of advantages to this.
Although we will need something like the described protocol for
announcement.
We could do destination folding on successful requests too (where both nodes are opennet).
Ian.
PGP.sig