On Thursday 14 October 2010 19:10:30 Robert Hailey wrote: > > Opennet may be negatively effecting network topology in this sense: > > "in a small-world network, most links are short links" (implying that > most nodes should have a few far links). > > In my observation (not having a firm connection to the larger > darknet), the open-net peer selection algorithm might be "too good", > in that *ALL* of my peers are super-short links (+-0.01). If this is > generally the case, then a request would not be able to make it around > the network in 18 hops (more like 50-100), or if it does would have to > follow "real" (darknet) links (presuming they would have a higher > likelyhood of being far-links. > > Would it greatly complicate the code to make all opennet-bound peers > have two peers "across-the-network"? maybe even ideally one at +0.33 > and one at -0.33 from the nodes location?
Yes. All the simulations show that opennet works, and adapts to the local situation, even taking performance into account to some degree. Forcing things would require extensive simulations and the support of at least one of our theoreticians. Unfortunately all 3 of them have disappeared. :|
signature.asc
Description: This is a digitally signed message part.
_______________________________________________ Devl mailing list [email protected] http://freenetproject.org/cgi-bin/mailman/listinfo/devl
