| > Single point of attack. They shut down your custom resolver, and they shut down your custom naming system. Also this proposal fails in terms of availability and resilience. Peer-to-peer networks are easy to overthrow completely even with a relatively small number of malicious nodes. > Also it's hierarchical P2P, so if you control the root servers, you control the naming system. It is decentralized only to aid availability and resilience. ICANN only controls delegation to TLDs. > I want free names for 10 websites. I want free beer. > Such a naming system would be outside the immediate control of governments, therefore democracy has nothing to do with it. In the real world people care about ownership disputes, protecting trademarks, accountability and other legal matters. > This still uses DNS, and does not solve anything. Actually, it does. DNS solves everything just fine. |
> I want free beer.
A p2p naming system would use free software and shared computing resources. There are numerous examples of both (GNU and BOINC/Gnutella/Bittorrent respectively.) So striving for a free naming system is not the same as striving for free beer.
> Peer-to-peer networks are easy to overthrow completely even with a relatively small number of malicious nodes.
Not if your p2p model uses a web of trust model like PGP. This is what the proposed model uses.
I don't agree with the proposed model, for what its worth. I think they should be looking at leveraging the work done on semantic free referencing at MIT, instead of the existing name to IP model.
(http://nms.csail.mit.edu/projects/sfr/)