Peer-to-peer (P2P) file sharing applications consist of nodes that share part of their local memory. A node can download a file (also called content) from another node participating to the P2P community, and must allow others to download files stored in its own shared local memory. Since nodes join and leave the P2P community at their will in an uncoordinated way, copies of the same file must be stored in many nodes. By means of this redundancy, the probability that a given content cannot be retrieved because all nodes storing it are temporary inactive can be kept small. Content management policies are the distributed strategies implemented at the nodes to decide the set of files to be stored in the local memory. In this paper, we propose, analyze and compare different content management policies. The considered performance indices include the probability that content cannot be retrieved due to lack of available copies of the content, and the amount of traffic generated in the network.

In peer-to-peer networks, nodes can be both resource providers and resource consumers at the same time. In this sense, the services offered by a peer-to-peer network rely on resource sharing among peers. This work focuses on how peers share their access link capacity between upload and download rates. In our model peers are rational agents, and choose their strategy in order to maximize their own utility. We suppose that the bottleneck is not in the network core, but in the network edge: the access link capacity of each peer connected to the network is a scarce resource and the peers have to compete for it. Instead of trying to settle the controversies which can arise, we imagine that every peer organizes an auction to give away its bandwidth. While the service rate is always granted to the peer who makes the lowest request, the amount of allocated rate depends on the implemented auction mechanism. Numerical experiments show that in a peer-to-peer game where the access link capacities are homogeneous, a second-price auction guarantees that the equilibrium rate allocation is optimal.