Adaptive Content Management in Structured P2P Communities
Jussi Kangasharju (University of Helsinki, Finland)
A fundamental paradigm in P2P is that of a large community of
intermittently-connected nodes that cooperate to share files.
Because nodes are intermittently connected, the P2P community must
replicate and replace files as a function of their popularity to
achieve satisfactory performance. We develop a suite of distributed,
adaptive algorithms for replicating and replacing content in a P2P
community. We do this for structured P2P communities, in which a
distributed hash table (DHT) overlay is available for locating the
node responsible for a key. In particular, we develop the Top-
MFR replication and replacement algorithm, which can be layered on
top of a DHT overlay, and in addition adaptively converges to a
nearly-optimal replication profile. Furthermore, we evaluate the
file transfer load caused by the adaptive algorithms on each peer,
and present two approaches for achieving a better load balance. Our
evaluation shows that with our two algorithms, an arbitrary load
distribution is possible, hence allowing each peer to serve requests
at the rate it wishes.