MITs p2psim is a little late on the scene - check out all the other p2p sims available in the summary paper in the p2pjournal:
http://p2pjournal.com/issues/November03.pdf
Think about using the NeuroGrid simulator if you want to simulate something other than just a boring DHT:-)
Isn't the problem with trying to apply HyperCubes to a p2p network is that it takes no account for the churn rate of the peers in the network?
Hypernets are all about maintaining connections to a particular subset of nodes right?
The important thing about Gnutella and other P2P nets is that the peers are unstable, dropping in and out. How can you maintain a hypernet over the network when from one moment to the next you can't predict which peers are a part of the network?
The advantage (if I can call it that) of the Cayley tree is that you just need any set of random connections and you're away...
Slashdot Mirror
MITs p2psim is a little late on the scene - check out all the other p2p sims available in the summary paper in the p2pjournal: http://p2pjournal.com/issues/November03.pdf Think about using the NeuroGrid simulator if you want to simulate something other than just a boring DHT :-)
Isn't the problem with trying to apply HyperCubes to a p2p network is that it takes no account for the churn rate of the peers in the network?
...
Hypernets are all about maintaining connections to a particular subset of nodes right?
The important thing about Gnutella and other P2P nets is that the peers are unstable, dropping in and out. How can you maintain a hypernet over the network when from one moment to the next you can't predict which peers are a part of the network?
The advantage (if I can call it that) of the Cayley tree is that you just need any set of random connections and you're away