Distributed Lookup in Structured Peer-to-Peer Ad-Hoc Networks Rapha¨ el Kummer, Peter Kropf, and Pascal Felber Computer Science Department , University of Neuchˆatel, Emile-Argand 11, CP 158,CH-2009 Neuchˆatel, Switzerland {raphael.kummer,peter.kropf,pascal.felber}@unine.ch http://www.unine.ch/iiun Abstract. Various peer-to-peer (P2P) architectures for ad-hoc networks have been proposed over the last few years. Most of them are unstruc- tured and use some form of flooding to locate content, because the phys- ical constraints of the underlying network make the construction of ar- bitrary application-layer overlays impractical. In this paper, we study the problem of applying distributed hash tables (DHT) to ad-hoc networks. Our approach to efficiently lookup content in such networks exploits physical proximity of peers when establishing and maintaining the DHT based routing tables. The efficiency of our method is demonstrated by simulation of large networks. 1 Introduction Peer-to-peer (P2P) systems, in which peer nodes form a cooperative network and share their resources (storage, CPU, bandwidth), have attracted a lot of interest lately. Roughly speaking, P2P networks can be classified as either struc- tured or unstructured. Unstructured P2P networks (e.g., the Gnutella [1] and KaZaA [2] file sharing systems) have no precise control over the file placement and generally use “flooding” search protocols, which typically generate a large number of query messages. In contrast, structured P2P networks (e.g., Chord [3], CAN [4], Pastry [5], P-Grid [6]) use specialized placement algorithms to assign responsibility for each file to specific peers, as well as “directed search” protocols to efficiently locate files. Structured P2P networks using deterministic algorithms are based on the distributed hash table (DHT) paradigm. Each data item is identified by a key (data identifier). The DHT maps keys to the nodes of an overlay network and provides facilities for locating the current peer node responsible for a given key. DHT designs differ mostly by the way they maintain the network and perform lookups: there is a fundamental trade-off between the degree of the network (the number of neighbors per node, i.e., the size of the routing tables) and its diameter (the average number of hops required per lookup) [7]. Directed search protocols are particularly efficient, because they accurately route queries toward the peers responsible for a given file. They require few communication stages generating only little traffic, and do not produce false