Peer-to-peer file sharing

Napster and eDonkey2000 both used a central server-based model. These systems relied on the operation of the respective central servers, and thus were susceptible to centralized shutdown. Their demise led to the rise of networks like Limewire, Kazaa, Morpheus, Gnutella, and Gnutella2, which are able to operate without any central servers, eliminating the central vulnerability by connecting users remotely to each other. However, these networks still relied on specific, centrally distributed client programs, so they could be crippled by taking legal action against a sufficiently large number of publishers of the client programs. Sharman Networks, the publisher of Kazaa, has been inactive since 2006. StreamCast Networks, the publisher of Morpheus, shut down on April 22, 2008. Limewire LLC was shut down in late 2010 or early 2011. This cleared the way for the dominance of the Bittorrent protocol, which differs from its predecessors in two major ways. The first is that no individual, group, or company owns the protocol or the terms "Torrent" or "Bittorrent", meaning that anyone can write and distribute client software that works with the network. The second is that Bittorrent clients have no search functionality of their own. Instead, users must rely on third-party websites like Isohunt or The Pirate Bay to find "torrent" files, which function like maps that tell the client how to find and download the files that the user actually wants. These two characteristics combined offer a level of decentralization that makes Bittorrent practically impossible to shut down. File-sharing networks are sometimes organized into three "generations" based on these different levels of decentralization.[4][5] Illegal darknets, including networks like Freenet, are sometimes considered to be third-generation file-sharing networks.[6]

Peer-to-peer file sharing is also efficient in terms of cost.[7] The system administration overhead is smaller because the user is the provider and usually the provider is the administrator as well. Hence each network can be monitored by the users themselves. At the same time, large servers sometimes require more storage and this increases the cost since the storage has to be rented or bought exclusively for a server. However, usually peer-to-peer file sharing does not require a dedicated server.[8]

The economic effect of copyright infringement through peer-to-peer file sharing on music revenue has been controversial and difficult to determine. Unofficial studies found that file sharing had a negative impact on record sales.[11][12][13][14][15] It has proven difficult to untangle the cause and effect relationships among a number of different trends, including an increase in legal online purchases of music; illegal file-sharing; drop in the prices of compact disks; and the closure of many independent music stores with a concomitant shift to sales by big-box retailers.[16]

The Motion Picture Association (MPAA) reported that American studios lost $2,373 billion in 2005 (equivalent to $3,556 billion in 2022) representing approximately one third of the total cost of film piracy in the United States.[17] The MPAA's estimate was doubted by commentators since it was based on the assumption that one download was equivalent to one lost sale, and downloaders might not purchase the movie if illegal downloading was not an option.[18][19][20] Due to the private nature of the study, the figures could not be publicly checked for methodology or validity.[21][22][23] In January 2008, as the MPAA was lobbying for a bill which would compel Universities to crack down on piracy, it was admitted by MPAA that its figures on piracy in colleges had been inflated by up to 300%.[24][25]

A 2010 study, commissioned by the International Chamber of Commerce and conducted by independent Paris-based economics firm TERA, estimated that unlawful downloading of music, film and software cost Europe's creative industries several billion dollars in revenue each year.[26] A further TERA study predicted losses due to piracy reaching as much as 1.2 million jobs and €240 billion in retail revenue by 2015 if the trend continued. Researchers applied a substitution rate of ten percent to the volume of copyright infringements per year. This rate corresponded to the number of units potentially traded if unlawful file sharing were eliminated and did not occur.[27] Piracy rates for popular software and operating systems have been common, even in regions with strong intellectual property enforcement, such as the United States or the European Union.[28]

Peer communities are formed on the basis of common interests. For Khambatti, Ryu and Dasgupta common interests can be labelled as attributes "which are used to determine the peer communities in which a particular peer can participate".[36] There are two ways in which these attributes can be classified: explicit and implicit attributes.

Explicit values are information that peers provide about themselves to a specific community, such as their interest in a subject or their taste in music. With implicit values, users do not directly express information about themselves, albeit, it is still possible to find information about that specific user by uncovering his or her past queries and research carried out in a P2P network. Khambatti, Ryu and Dasgupta divide these interests further into three classes: personal, claimed and group attributes.[36]

A full set of attributes (common interests) of a specific peer is defined as personal attributes, and is a collection of information a peer has about him or herself. Peers may decide not to disclose information about themselves to maintain their privacy and online security. It is for this reason that the authors specify that "a subset of...attributes is explicitly claimed public by a peer", and they define such attributes as "claimed attributes".[36] The third category of interests is group attributes, defined as "location or affiliation oriented" and are needed to form a...basis for communities", an example being the "domain name of an internet connection" which acts as an online location and group identifier for certain users.

Cost reduction influences the sharing component of P2P communities. Users who share do so to attempt "to reduce...costs" as made clear by Cunningham, Alexander and Adilov.[37] In their work Peer-to-peer File Sharing Communities, they explain that "the act of sharing is costly since any download from a sharer implies that the sharer is sacrificing bandwidth".[37] As sharing represents the basis of P2P communities, such as Napster, and without it "the network collapses", users share despite its costs in order to attempt to lower their own costs, particularly those associated with searching, and with the congestion of internet servers.[37]

User motivation and the size of the P2P community contribute to its sustainability and activity. In her work Motivating Participation in Peer to Peer Communities, Vassileva studies these two aspects through an experiment carried out in the University of Saskatchewan (Canada), where a P2P application (COMUTELLA) was created and distributed among students. In her view, motivation is "a crucial factor" in encouraging users to participate in an online P2P community, particularly because the "lack of a critical mass of active users" in the form of a community will not allow for a P2P sharing to function properly.[38]

Usefulness is a valued aspect by users when joining a P2P community. The specific P2P system must be perceived as "useful" by the user and must be able to fulfil his or her needs and pursue his or her interests. Consequently, the "size of the community of users defines the level of usefulness" and "the value of the system determines the number of users".[38] This two way process is defined by Vassileva as a feedback loop, and has allowed for the birth of file-sharing systems like Napster and KaZaA. However, in her research Vassileva has also found that "incentives are needed for the users in the beginning", particularly for motivating and getting users into the habit of staying online.[38] This can be done, for example, by providing the system with a wide amount of resources or by having an experienced user provide assistance to a less experienced one.

Users participating in P2P systems can be classified in different ways. According to Vassileva, users can be classified depending on their participation in the P2P system. There are five types of users to be found: users who create services, users who allow services, users who facilitate search, users who allow communication, users who are uncooperative and free ride.[38]

In the first instance, the user creates new resources or services and offers them to the community. In the second, the user provides the community with disk space "to store files for downloads" or with "computing resources" to facilitate a service provided by another users.[38] In the third, the user provides a list of relationships to help other users find specific files or services. In the fourth, the user participates actively in the "protocol of the network", contributing to keeping the network together. In the last situation, the user does not contribute to the network, downloads what he or she needs but goes immediately offline once the service is not needed anymore, thus free-riding on the network and community resources.[38]

Media industries have made efforts to curtail the spread of copyrighted materials through P2P systems. Initially, the corporations were able to successfully sue the distribution platforms such as Napster and have them shut down. Additionally, they litigated users who prominently shared copyrighted materials en masse. However, as more decentralized systems such as FastTrack were developed, this proved to be unenforceable. There are also millions of users worldwide who use P2P systems illegally, which made it impractical to seek widespread legal action. One major effort involves distributing polluted files into the P2P network. For instance, one may distribute unrelated files that has the metadata of a copyrighted media. This way, users who downloads the media would receive something unrelated to what they have been expecting.[56]