An Optimal Replica Placement Algorithm for Symphony P2P Networks

Placing replicas in a P2P network can reduced the number of hops required by a node to search an object. One existing replica placement strategy for structured P2P networks allocates replicas at nodes that are continuously close to the destination. In this paper, we investigate the effectiveness of such an approach. We propose an optimal algorithm to place replicas for an object, in terms of number of hops, in the deterministic P2P network (Chord) and randomized P2P network (Symphony). Particularly, our replica placement algorithm is based on the technique of dynamic programming. We perform several experiments to find the optimal replica placements for a large number of Symphony networks and analyze the results. Our simulation results show that locations of optimal replica placements are not necessarily continuously close to the destination.