Re: DHT vs Gnutella
[email protected] Tue, 5 Aug 2008 20:11:15 +0000 (UTC)
| Newsgroups | gmane.network.gnutella.devel |
|---|---|
| Organization | Home, Grenoble, France |
| Message-ID | <[email protected]> |
Quoting Nikolay Raspopov <[email protected]> from ml.gnutella.dev-forum: : RM> That's why we need to bring a different kind of structure to the : RM> network, such as the one provided by Kademlia, to be able to : RM> efficiently search the rare data with assurance to find it if it : RM> exists. : :Can not agree. Prove it. In a flood model, search horizon is limited, so you cannot reach the whole network and you cannot know which nodes hold the data you want. Therefore finding rare items is just pure luck. In a structured network such as the one provided by Kademlia, the topology guides you to the nodes where you are most likely to find the data. Raphael