Hi! On 14:37 Wed 26 Oct , Luca Dionisi wrote: ... > The problem of Denial Of Service is real. > > Also in centralized networks you suffer it. > Nobody can be sure that his ISP is acting fairly. > You cannot even be sure that your traffic will be routed. > Look at the Egyptian Internet shutdown and similar.
If you have only one connection and it is shutdown, you will obviously have problems... > Obviously in a completely decentralized network you'll have even less > reliability. I do not think so. > It goes like this: > You are node A. You have a neighbor B. > It tells you that he has a good path to a certain gnode. > Can you be 100% sure that it is not lying? No. > Could you be 100% sure that it is cheating? I think not. Agree. > Could you be sure that its path is not reliable? You should test a > connection with any possible node in that gnode. I think you cannot. The principle is simple: You just try use it. If it works, fine. If it does not, you try a different path. > I think that looking for a good solution to this sort of problem will > lead you to a very flooded network. It might have some overhead. I cannot really tell you how much. > And possibly you could ruin the principles of the decentralization and > self-configuration, self-healing of the network. I do not see why it should. > But this is just my opinion, I could be wrong. > One thing you might do is show me that a solution is feasible. Give > your solution in great details. http://michaelblizek.twilightparadox.com/projects/cor/resilience.html > Another solution is to stay with the current centralized way of doing, > choosing to trust our ISPs and their backbones to do the best for > their users. We are stuck with this now. I am not really happy with this. Even if we choose to trust our ISPs, do we trust our governments? All governments in the world? How much power do we want to give them? -Michi _______________________________________________ Netsukuku mailing list [email protected] http://lists.dyne.org/mailman/listinfo/netsukuku
