X hits on this document





8 / 8

lief, parallel downloading is far from being optimal in terms of reducing the download time. Its performance is very much dependent upon the heterogeneity in service capacities in the network and upon the degree of competi- tion in the network. In a network consisting of many het- erogeneous sources (as in current P2P networks), we are better off keeping fewer connections so that not every user gets stuck in the bad sources. Overall, from Figures 4, it is evident that the random periodic switching generally performs the best. Although it may not compete with parallel downloading in the case where all sources offer the same service capacity with very light competition, those conditions are not practical in most cases. For all other situations with realistic heterogeneity, competition among peers, as well as with temporal correlations in each path, the random periodic switching gives the most stable and optimal performance in the sense that its download time is minimal, robust with respect to network config- urations, and consistent for different users in a random environment.


[1] M. Adler, R. Kumar, K. Ross, D. Rubenstein, D. Turner, and D. D. Yao. Optimal peer selection in a free-market peer-resource economy. In Proceedings of Workshop on Economics of Peer- to-Peer Systems (P2PEcon), Cambridge, MA, Jun. 2004.

[2] M. Adler, R. Kumar, K. Ross, D. Rubenstein, D. Turner, and D. D. Yao. Optimal peer selection for p2p downloading and streaming. In Proceedings of IEEE Infocom, Miami, FL, Mar. 2005.

[3] Y. Chiu and D. Y. Eun. Minimizing File Download Time over Stochastic Channels in Peer-to-Peer Networks. Technical re- port, North Carolina State University, Raleigh, NC, Dec. 2005. available at “http://www4.ncsu.edu/dyeun/pub/techrep- ciss06-chiu.pdf”.

[4] G. Bianchi F. Lo Piccolo, G. Neglia. The effect of heterogeneous link capacities in bittorrent-like file sharing system. In IEEE International Workshop on Hot Topics in Peer-to-Peer Systems (HOT-P2P), Oct. 2004.

[5] K. P. Gummadi, R. J. Dunn, and S. Saroiu. Measurement, mod- eling, and analysis of a peer-to-peer file sharing workload. In Proceedings of ACM Symposium on Operating Systems Princi- ples (SOSP), 2003.

[6] N. Hu and P. Steenkiste. Evaluation and characterization of available bandwidth probing techniques. IEEE Journal on Se- lected Areas in Communications, 21(6), Mar. 2003.

[7] M. Jain and C. Dovrolis. End-to-end estimation of the available bandwidth variation range. In Proceedings of ACM Sigmetrics, Jun. 2005.

[8] S. Koo, C. Rosenberg, and D. Xu. Analysis of parallel down- loading for large file distribution. In Proceedings of IEEE Inter- national Workshop on Future Trends in Distributed Computing Systems (FTDCS), May 2003.

[9] A. M¨uller and D. Stoyan. Comparison Methods for Stochastic Models and Risks. John Wiley & Son, New York, NY, 2002.

[10] D. Qiu and R. Srikant. Modelling and performance analysis of bittorrent-like peer-to-peer networks. In Proceedings of ACM Sigcomm, Aug. 2004.

[11] K. K. Ramachandran and B. Sikdar. An analytic framework for modelling peer to peer networks. In Proceedings of IEEE Infocom, Mar. 2005.

[12] S. M. Ross. Stochastic Processes. John Wiley & Son, New York, second edition, 1996.


[13] S. Saroiu, K. P. Gummadi, and S. D. Gribble. A measurement study of peer-to-peer file sharing systems. In Proceegins of ACM Multimedia Computing and Networking (MMCN), 2002.

[14] X. Yang and G. de Veciana. Service capacity of peer to peer networks. In Proceedings of IEEE Infocom, Mar. 2004.

Document info
Document views26
Page views26
Page last viewedTue Jan 17 23:44:14 UTC 2017