Welcome to D
SIGMOD 2005
PODS 2005
SIGMOD-RECOR
CIDR 2005
CIKM 2005
COMAD 2005
CVDB 2005
DaMoN 2005
Data Enginee
<<< = DE-Bulletin >>>
DEBS05
DMSN 2005
DOLAP 2005
GIR 2005
GIS 2005
Hypertext 20
ICDE 2005
ICDM 2005
IHIS 2005
IQIS 2005
JCDL 2005
KRAS 2005
MDM 2005
MIR 2005
MobiDE 2005
P2PIR 2005
RIDE 2005
SBBD 2005
SIGIR 2005
SIGIR-FORUM
SIGKDD 2005
SIGKDD-EXP
SSDBM 2005
TIME 2005
TKDE 2005
TODS 2005
VLDB 2005
VLDBJ 2005
WebDB 2005
WIDM 2005

Structured Peer-to-Peer Networks: Faster, Closer, Smarter


P. Felber, K.W. Ross, E.W. Biersack, L. Garcs-Erice, and G. Urvoy-Keller

  View Paper (PDF)  

Return to March 2005, In-Network Query Processing, Volume 28, Number 1


Abstract

Peer-to-peer (P2P) distributed hash tables (DHTs) are structured networks with decentralized lookup capabilities. Each node is responsible for a given set of keys (identifiers) and lookup of a key is achieved by routing a request through the network toward the current peer responsible for the desired key. DHT designs are usually compared in terms of degree (number of neighbors) and diameter (length of lookup paths). In this paper, we focus three other desirable properties of DHT-based systems: We first present a topology-aware DHT that routes lookup requests to their destination along a path that mimics the router-level shortest-path, thereby providing a small "stretch." We then show how we can take advantage of the topological properties of the DHT to cache information in the proximity of the requesters and reduce the lookup distance. Finally, we briefly discuss techniques that allow users to look up resources stored in a DHT, even if they only have partial information for identifying these resources.


©2006 Association for Computing Machinery