Experimental study on neighbor selection policy for Phoenix Network Coordinate system

Network Coordinate (NC) which provides efficient distance prediction with scalable measurements, brings benefits to various network applications. Phoenix Network Coordinate system is a recently proposed dot product based NC system with high prediction accuracy and better robustness. In this paper, we discuss the neighbor selection policies for Phoenix and propose a modified Phoenix with Hybrid neighbor selection policy. Extensive experiments have been carried out to evaluate the performance of Phoenix NC systems with different neighbor selection policies. The results show that the system with Hybrid policy achieves lower relative errors (REs) and obtains a higher accuracy in selecting the nearest neighbor host. Moreover, the Hybrid policy based Phoenix NC system has a better performance in the application of Overlay Multicast. It reduces the tree cost of the ESM and MST multicast tree construction by at least 20%.

[1]  Jonathan M. Smith,et al.  IDES: An Internet Distance Estimation Service for Large Networks , 2006, IEEE Journal on Selected Areas in Communications.

[2]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[3]  Eng Keong Lua,et al.  Embeddable Overlay Networks , 2007, 2007 12th IEEE Symposium on Computers and Communications.

[4]  Margo I. Seltzer,et al.  Network-Aware Overlays with Network Coordinates , 2006, 26th IEEE International Conference on Distributed Computing Systems Workshops (ICDCSW'06).

[5]  Sonia Fahmy,et al.  Impact of the Inaccuracy of Distance Prediction Algorithms on Internet Applications - an Analytical and Comparative Study , 2006, Proceedings IEEE INFOCOM 2006. 25TH IEEE International Conference on Computer Communications.

[6]  David R. Karger,et al.  Finding nearest neighbors in growth-restricted metrics , 2002, STOC '02.

[7]  Bo Zhang,et al.  Towards network triangle inequality violation aware distributed systems , 2007, IMC '07.

[8]  Bobby Bhattacharjee,et al.  Scalable application layer multicast , 2002, SIGCOMM '02.

[9]  Miguel Castro,et al.  PIC: practical Internet coordinates for distance estimation , 2004, 24th International Conference on Distributed Computing Systems, 2004. Proceedings..

[10]  Eng Keong Lua,et al.  Internet Routing Policies and Round-Trip-Times , 2005, PAM.

[11]  Antony I. T. Rowstron,et al.  Pastry: Scalable, Decentralized Object Location, and Routing for Large-Scale Peer-to-Peer Systems , 2001, Middleware.

[12]  Klara Nahrstedt,et al.  Effective location-guided tree construction algorithms for small group multicast in MANET , 2002, Proceedings.Twenty-First Annual Joint Conference of the IEEE Computer and Communications Societies.

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

[14]  Peter Druschel,et al.  Pastry: Scalable, distributed object location and routing for large-scale peer-to- , 2001 .

[15]  Xiao Wang,et al.  Phoenix: Towards an Accurate, Practical and Decentralized Network Coordinate System , 2009, Networking.

[16]  Robert Tappan Morris,et al.  Vivaldi: a decentralized network coordinate system , 2004, SIGCOMM '04.

[17]  Jibin Zhan,et al.  Early Experience with an Internet Broadcast System Based on Overlay Multicast , 2004, USENIX Annual Technical Conference, General Track.