Structured peer-to-peer (p2p) overlay networks like CAN, Chord, Pastry and Tapestry [3, 6, 5, 9] provide a self-organizing substrate for large-scale peer-to-peer applications. These systems provide efficient, fault-tolerant routing, object location and load balancing within a self-organization overlay network.

In this paper, we show how redundant information that is collected as part of the normal overlay maintenance protocol can be exploited to improve the performance of routing, in terms of both the number of routing hops and routing delay penalty. We use Pastry as a concrete example to describe the set of optimizations and to evaluate their improvement in routing performance via a large scale simulation using a realistic network topology model. We then discuss how these optimizations can be applied to other structured p2p overlays.