On Thu, Aug 17, 2000 at 06:42:03PM -0400, Travis Bemann wrote: > > > NO, NO, NO! We have to keep capitalism from corrupting Freenet as > > > much as possible. > > > > If it were possible to do right it would be good. It will be hard to > > justify asking people to give up there bandwidth for free. > > So why do people run Gnutella nodes?
Because currently people are being provided with bandwidth without paying for it. This is because of the intense competition in the provider industry and because the providers calculate that people will only use a fraction of the bandwidth they are given (but they still have be able to advertise "10 MB BROADBAND!"). This of course why the Service providers then go on to have usage agreements that prohibit actually using the bandwdith for anything it would be good for (such as running Gnutella nodes). I predict this will end and people will have to go back to actually paying for the bandwidth. > > > Therefore, we must be quite paranoid when designing Freenet (which > > > very well may require out of band communication). We should use PK > > > encryption for connections, out of band initial key exchange, > > > automatically obfuscating file lengths, encrypted datastores on > > > machines with encrypted swap partitions, etc. > > > > That is still not perfect anonymity. > > It is pretty damn hard to get perfect anonymity. Even onion routing > can be subject to traffic analysis. Yeah, but onion routing with constant cover traffic is a damn site better then what we have to offer. > > > This is something that can be programmed into nodes without affecting > > > the protocol itself. Just have nodes record the average connection > > > speed with other nodes. > > > > Your still not understanding Freenet, Travis. Nodes cannot just route at > > the leisure, it will kill the network if they do. > > That is how I said you could rate connections by capacity. I never > said that it would work in Freenet. If the node uses some selection mechanism other then closest key to select which route to give a request, it effects the network. Therefore you can't say that this sort of thing is something that can be programmed into nodes without effecting the protcol, because it does. > > > -- > Travis Bemann > Sendmail is still screwed up on my box. > My email address is really bemann at execpc.com. -- \oskar _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
