Recent peer-to-peer (P2P) networks, represented by CAN, Chord, and Pastry, offer an administration-free and fault-tolerant application-level overlay network. For these systems to function efficiently, they must make effective use of the underlying network topology. Existing techniques for discovering network proximity information, such as landmark clustering and expanding-ring search, are either inaccurate or expensive. Moreover, the lack of global proximity information in overlay construction and maintenance results in either bad proximity approximation or excessive communication. To address these problems, we propose the following: (1) Combining landmark clustering and RTT measurements to identify the closest node, achieving both efficiency and accuracy. (2) Controlled placement of global proximity information on the system itself as soft-state, such that nodes can independently access relevant information efficiently. (3) Publish/subscribe functionality that allows nodes to subscribe to the relevant soft-state and get notified as the state changes necessitate overlay restructuring. 10 Pages