Availability and locality measurements of peer-to-peer file systems

Although peer-to-peer networking applications continue to increase in popularity, there have been few measurement studies of their performance. We present the first study of the locality of files stored and transferred among peers in Napster and Gnutella over month-long periods. Our analysis indicates that the locality of files is skewed in all four cases and fits well to a log-quadratic distribution. This predicts that caches of the most popular songs would increase performance of the system. We also took baseline measurements of file types and sizes for comparison over time with future studies. Not surprisingly, audio files are most popular, however a significant fraction of stored data is occupied by videos. Finally, we measured the distribution of time peers in Gnutella were available for downloading. We found that node availability is strongly influenced by time-of-day effects, and that most user's tend to be available for only very short contiguous lengths of time.

[1]  Stefan Saroiu,et al.  A Measurement Study of Peer-to-Peer File Sharing Systems , 2001 .

[2]  Ian T. Foster,et al.  Mapping the Gnutella Network: Properties of Large-Scale Peer-to-Peer Systems and Implications for System Design , 2002, ArXiv.

[3]  Evangelos P. Markatos,et al.  Tracing a Large-Scale Peer to Peer System: An Hour in the Life of Gnutella , 2002, 2nd IEEE/ACM International Symposium on Cluster Computing and the Grid (CCGRID'02).

[4]  Donald F. Towsley,et al.  Modeling peer-peer file sharing systems , 2003, IEEE INFOCOM 2003. Twenty-second Annual Joint Conference of the IEEE Computer and Communications Societies (IEEE Cat. No.03CH37428).

[5]  Krishna P. Gummadi,et al.  Measurement study of peer-to-peer file system sharing , 2002 .

[6]  Brian Neil Levine,et al.  Evaluation of a novel two-step server selection metric , 2001, Proceedings Ninth International Conference on Network Protocols. ICNP 2001.

[7]  Yakov Rekhter,et al.  Address Allocation for Private Internets , 1994, RFC.

[8]  Li Fan,et al.  Web caching and Zipf-like distributions: evidence and implications , 1999, IEEE INFOCOM '99. Conference on Computer Communications. Proceedings. Eighteenth Annual Joint Conference of the IEEE Computer and Communications Societies. The Future is Now (Cat. No.99CH36320).

[9]  Eytan Adar,et al.  Free Riding on Gnutella , 2000, First Monday.