I doubt that your analysis is correct, but I would definitely be interested in a simulation. Perhaps our resident mathematician can be persuaded to run one. :)
The point about inter-country links here is simply that a large western darknet is fairly unlikely, whereas a large chinese darknet and a huge western opennet are likely. Also that there ARE things we can do about this, *if* it is a problem. These would probably have to be done before any i2p integration for premix routing, but that's probably a year in the future anyway. On Sat, Oct 15, 2005 at 05:31:00PM +0100, Ian Clarke wrote: > On 15 Oct 2005, at 17:15, Matthew Toseland wrote: > >Suppose that Freenet 0.7 is a success both in the West and in China. > >There is a large darknet in China (say 10,000 nodes). There is a large > >opennet in the West (say 1,000,000 nodes). There are a rather small > >number of nodes which connect to both systems (say 100). > > If the number of connecting nodes is that small, then yes, there will > be a problem since that will not be consistent with a small-world > link distribution. Hopefully in reality inter-country connections > will be more consistent with small-world than that. > > >This is going to be a problem. In theory, the architecture is such > >that > >the networks will talk to each other as a hybrid. In practice, with > >the > >exception of keys which happen to be close to the location of the > >gateway nodes, these are distinct networks with no common content. > > Why? What is likely is that the "west" will be in one part of > keyspace, china will be in another, and the nodes which link the two > will be in-between. Sure, if there are only 100 nodes linking the > two then they might get overloaded, but that is where caching comes in. > > I would like to see a simulation of this kind of situation before we > conclude that it will be a problem, I suspect it might not. > > Ian. -- Matthew J Toseland - toad at amphibian.dyndns.org Freenet Project Official Codemonkey - http://freenetproject.org/ ICTHUS - Nothing is impossible. Our Boss says so. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 189 bytes Desc: Digital signature URL: <https://emu.freenetproject.org/pipermail/tech/attachments/20051015/eccce0a0/attachment.pgp>
