I saw this on Edupage, a news summary mailing list, and thought it would be of interest here. The full article is from the 21 Aug 2000 issue of the New York Times.
MANY TAKE, BUT FEW GIVE ON GNUTELLA The file-exchange service Gnutella is not operating at peak efficiency, largely because more users take than give, according to two researchers at Xerox. Eytan Adar and Bernardo A. Huberman studied the Gnutella network for a 24-hour period this month and recorded 3,019,405 files being exchanged by 31,395 users. However, the researchers noted most users downloaded files but did not open their own files to others on the network. This slows the system considerably because everyone must then rely on only a few users to provide files. The researchers called this behavior a "tragedy of the digital commons." Moreover, the researchers believe Gnutella may be open to legal challenge despite its decentralized structure. During the period the researchers studied, only 1 percent of users, or 314 computers, made available 40 percent of the files circulating through the network. These users could then be identified by their Internet protocol addresses. Record industry insiders say they are "analyzing" Gnutella but will not say if they will launch a legal attack similar to the one against Napster. The full paper can be found at: http://www.parc.xerox.com/istl/groups/iea/papers/gnutella/ The paper tangentially mentions Freenet, but apparently they have not looked into it very deeply. They state: FreeNet, for example, forces caching of downloaded files in various hosts. This allows for replication of data in the network forcing those who are on the network to provide shared files. Unfortunately, such a system is prone to replication of "bad" or illegal data and "tainting" hosts [5]. The second cost of the automatic replication as implemented in FreeNet is the unique identifiers for files that forces users to know exactly what they are looking for. The point I find of interest in the study is that very few people were sharing disk space for others to use. There were many "takers", but very few "givers". Obviously, Freenet is quite different from Gnutella in terms of architecture, so the findings may not be applicable. For example, more people might be willing to share disk space and bandwith if they had greater assurances about the prospects of getting sued. Even so, this is a rare, refreshing, real-world measurement of usage patterns in a large, voluntary file-sharing network. Freenet may well face a strong propensity for users to share very little disk space, start their nodes only when they want a file, and stop the nodes as soon as they get the file they want. There is no doubt that Freenet server variants and clients should be able to handle a broad range of usage patterns without compromising basic design goals. All the same, we do have to strike a balance between security, efficiency, and several other factors. In the absense of good evidence to the contrary, it seems reasonable to suppose that the 'selfish' model is the best guess available regarding Freenet usage patterns in the near future. --Will (never, ever speaking for his employers) willdye at freedom.net _______________________________________________ Freenet-dev mailing list Freenet-dev at lists.sourceforge.net http://lists.sourceforge.net/mailman/listinfo/freenet-dev
