Practical, distributed network coordinates

Vivaldi is a distributed algorithm that assigns synthetic coordinates to internet hosts, so that the Euclidean distance between two hosts' coordinates predicts the network latency between them. Each node in Vivaldi computes its coordinates by simulating its position in a network of physical springs. Vivaldi is both distributed and efficient: no fixed infrastructure need be deployed and a new host can compute useful coordinates after collecting latency information from only a few other hosts. Vivaldi can rely on piggy-backing latency information on application traffic instead of generating extra traffic by sending its own probe packets.This paper evaluates Vivaldi through simulations of 750 hosts, with a matrix of inter-host latencies derived from measurements between 750 real Internet hosts. Vivaldi finds synthetic coordinates that predict the measured latencies with a median relative error of 14 percent. The simulations show that a new host joining an existing Vivaldi system requires fewer than 10 probes to achieve this accuracy. Vivaldi is currently used by the Chord distributed hash table to perform proximity routing, replica selection, and retransmission timer estimation.

[1]  Tony DeRose,et al.  Surface reconstruction from unorganized points , 1992, SIGGRAPH.

[2]  Jon Crowcroft,et al.  Lighthouses for Scalable Distributed Location , 2003, IPTPS.

[3]  Marcel Waldvogel,et al.  Efficient topology-aware overlay network , 2003, CCRV.

[4]  Erik D. Demaine,et al.  Anchor-Free Distributed Localization in Sensor Networks , 2003 .

[5]  David R. Karger,et al.  Wide-area cooperative storage with CFS , 2001, SOSP.

[6]  Hui Zhang,et al.  Predicting Internet network distance with coordinates-based approaches , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[7]  Krishna P. Gummadi,et al.  King: estimating latency between arbitrary internet end hosts , 2002, IMW '02.

[8]  Paul Francis,et al.  IDMaps: a global internet host distance estimation service , 2001, TNET.

[9]  Marcel Waldvogel,et al.  Routing and Data Location in Overlay Peer-to-Peer Networks , 2002 .

[10]  Krishna P. Gummadi,et al.  King: estimating latency between arbitrary internet end hosts , 2002, IMW '02.

[11]  Robert Morris,et al.  Chord: A scalable peer-to-peer lookup service for internet applications , 2001, SIGCOMM 2001.

[12]  Mark Handley,et al.  Topologically-aware overlay construction and server selection , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

[13]  Ben Y. Zhao,et al.  OceanStore: an architecture for global-scale persistent storage , 2000, SIGP.

[14]  Larry L. Peterson,et al.  Proceedings of the 5th Symposium on Operating Systems Design and Implementation the Effectiveness of Request Redirection on Cdn Robustness , 2022 .

[15]  Peter Druschel,et al.  Exploiting network proximity in peer-to-peer overlay networks , 2002 .

[16]  Yuval Shavitt,et al.  Big-bang simulation for embedding network distances in Euclidean space , 2004, IEEE/ACM Transactions on Networking.

[17]  Frank Dabek,et al.  Learning euclidean coordinates for internet hosts , 2002 .

[18]  Mischa Schwartz,et al.  ACM SIGCOMM computer communication review , 2001, CCRV.

[19]  Yuval Shavitt,et al.  Big-Bang simulation for embedding network distances in Euclidean space , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[20]  David R. Karger,et al.  Chord: a scalable peer-to-peer lookup protocol for internet applications , 2003, TNET.

[21]  Scott Shenker,et al.  Geographic routing without location information , 2003, MobiCom '03.