Other Free Encyclopedias » Online Encyclopedia » Encyclopedia - Featured Articles » Contributed Topics from K-O

Multimedia File Sharing - Background, Early, Private file sharing, Public file sharing, The legal issues

files system peer server

Simone Santini
University of California, San Diego, USA

Definition: In multimedia file sharing we define two, quite separate, modalities: one, which we might call private file sharing, consists in being able to send multimedia files to a well identified person or to a small group of well identified persons; the other, which we might call public file sharing, consists in publishing multimedia files so that an unspecified number of possible unknown people may copy them.


File sharing, in one form or another, is as old as computer networks. Programs such as uucp (unix to unix copy) and, later, ftp (from the file transfer protocol that it used) constituted the early means through which files were shared among network users. The birth of open source software, and the creation of communal development projects was facilitated and, to a certain extent, made possible by these programs, which allowed people located even at a great distance from one another to work on the same project, and to other people to enjoy the fruits of their labor without moving any physical data support around. Many open source programs in use today— gnu emacs , just to make an example—were at first developed and distributed in such communities.

One of the problems presented by these file sharing programs is that they lacked any function to search for a specific file: in order to retrieve a file using ftp , for instance, it is necessary to know the name of the file, that address (or network name) of the computer where it is stored, the directory where the file is, and, depending on the circumstances, a password to gain access to the computer. This information was not available within the system, and had to be distributed using other means. In the 80’s, during the first massive distribution of the internet inside universities, these means were mainly user groups and e-mail messages . In the early 1990’s, the first systems for assisting users in locating the diles that they desired to download (the term download also made its appearance around the same time) made their appearance. The most diffuse were archie , a name based search engine, and gopher , a taxonomization of available files that could be visited using menus.

The Web was, in its early years (roughly 1994-1998), relatively uninterested in file sharing: the emphasis was on documents, that is, on html files created for the sole purpose of being displayed in a browser and, while more and more people started putting files on the web for downloading, the underlying protocol was ftp, and the file localization services were rather primitive, since the early search services ( lycos, altavista) were meant to search for pages, not for files.

Early multimedia file sharing

During all this time, there were some quite formidable obstacle to the wide diffusion of multimedia file sharing habits and services: for one thing, bandwidth was still a problem for many users, who connected to the internet throughout modems that, although their speed increased quite rapidly from 14Kbps to 28, 33, and 56, were still too slow for a convenient transfer of a relatively large multimedia files; for another thing, most computers didn’t have any means of reproducing the content of audio and video files, and the lack of accepted standards for these files made the problems worse.

Just as it had happened two decades before with the VCR, the first market interested in multimedia file sharing was that of pornography. In the early 1990’s, a number of bulletin boards, with direct phone line connectivity, offered modem access to collections of pornographic images and, to a lesser extent, video, using proprietary access protocols. These systems survived well into the web era, until a satisfactory and secure way of making payment on the web was devised, and until the increasing bandwidth made acceptable the worse performance of the general-purpose http protocol with respect to the specially designed proprietary protocols (see mime types ). As with the VCR, pornography was an eager early adopter since the social stigma attached to it makes the search for privacy an overwhelmingly important goal, worth the uncertainties, the costs, and the problems of a new medium. But, while these experiments provided valuable expertise in multimedia file sharing, pornography, by its very nature, was destined to remain a niche market, as large as it may be, or, in any extent, a market with which companies don’t like to be publicly associated; a truly massive development of multimedia file sharing required a different market and a great convenience of use in order to appeal to the great public.

Private file sharing

In multimedia file sharing we can talk, in general terms, of two, quite separate, modalities: one, which we might call private file sharing, consists in being able to send multimedia files to a well identified person or to a small group of well identified persons; the other, which we might call public file sharing, consists in publishing multimedia files so that an unspecified number of possible unknown people may copy them.

The main technical factor behind the success of private file sharing was the agreement on standards for “attaching” binary files to e-mail messages. These standards are in general based on the so-called mime types , which are the same types that the browsers use in order to display properly data of different types. The e-mail protocol doesn’t allow sending binary files as they are, since many of the bytes with a value less than 32 (corresponding to the ASCII control codes ) are interpreted and, possibly, transformed. To make an example, Unix systems use the character 0D (hexadecimal), the ASCII carriage return to mark the end of a line, while Windows system, inheriting from DOS, use the pair 0A 0D ( line feed and carriage return ) for the same purpose. Many Windows programs, upon observing a 0D in input convert it in a pair 0D 0A. Clearly, if the character 0D was part of a binary file, this will result in an incorrect interpretation of the file. To obviate to this problem, binary mime types are in general encoded so that only “printable” ASCII characters, that is, only bytes with a code between 32 and 128, are used.

While this technical standard for attachments is important, one should not underestimate the importance, for the diffusion of private multimedia file sharing, of e-mail programs: the ease with which a file can today be attached to a message through an opportune interface, and the simplicity of “playing back” the attachments that one received are instrumental in the success of private file sharing.

Public file sharing

Public file sharing, especially of audio files, is the form of multimedia file sharing that most readily comes to mind to most people, and has configured itself, at this historical juncture, almost as the epitome of all file sharing, a not entirely fair situation. The first widely known and used audio file sharing system was called napster and various factors of different nature contributed to its success in 2000. For one thing, the network infrastructure had reached a level where a significant number of users had the bandwidth necessary to make sending and receiving relatively large files (5-20 Mbytes) reasonable; the choice of napster to concentrated on audio rather than on video, the fact that audio had an obvious market among the many young people that in the three years before had started using the internet, as well as the diffusion of the mp3 compression standard that kept the size of audio files within manageable limits, all these factors were crucial in the success of audio file sharing. To this, napster added a well integrated application, distributed through their Web site, which made it easy to search, download, manage, and play files, and the idea of using a peer-to-peer system rather than a client-server one.

The idea of using an application that the user had to download and install was in itself a minor challenge to the corporate conventional wisdom of the time that insisted that users would not download and install applications, that requiring them to do so would result in dramatic loss of audience, and that, therefore, only interfaces composed of web pages should be used. It turned out, as it is not seldom is the case, that the corporate conventional wisdom was more conventional than wise. Napster’s choice to develop an application allowed them to overcome the drastic limitation of the page based web interfaces, to provide an easier way to use the system and, in the end, it was an essential element of its success, helped in this by the fact that the security paranoia was not raging as high in 2000 as it is in 2005.

Technically, the main contribution of napster and other similar services (gnutella, audiogalaxy, kazaa,) was the introduction in the arena of multimedia file exchange of peer ro peer architectures [gnutella]. At the time, and probably today as well, the most common form of distributed system was the client server architecture. In a client server data system, all the data are kept in a single computer or in a small group of computers, and the “clients” send their request for data to them. Since all the requests go to the same server (or small group of servers), its performance and that of its network connection are crucial for the whole system, and servers are prone to become the bottleneck of a client server system, a problem that is exacerbated by the large file sizes typical of multimedia. An economic and political consequence of this fact is that setting up a client server multimedia architecture requires a massive up front investment, a situation that makes it relatively hard for a start-up to initiate such a service. By contrast, in a peer to peer system, all the participating computers contain some data: the participant seeking a particular file will broadcast a request for it (an operation whose modalities depend on the specific type of peer to peer system: the problem of location is indeed one of the most studied into the peer to peer system literature [Chord,Lv] ), and all the participant that possess that file will answer back with some information, such as their network bandwidth or the number of requests that they are serving at the moment, from which the requestor can determine from which is access more convenient. The two computers will then use a standard protocol (e.g. Ftp) to transfer the file. The etiquette of a peer to peer community now requires that the user who just received a file should keep it available, becoming an additional source for it.

The details of the transition may vary from system to system. Napster, for example, was, so to speak, an impure peer to peer system, in that the information about the participants was kept in a central server to which all localization requests were made, a technical solution that might have contributed to make Napster vulnerable to the legal attacks of which it was object (see below). The file exchange services that came on the scene after the demise of Napster realized a more peer to peer system which proved more robust technically and legally. These systems use the redundancy of the community—that is, the fact that the same file is present on many participants– to avoid sending all the requests to all the participants, without having to keep information in a central server. In Kazaa, a typical representative of the post-Napster file exchange services, whenever a new participant enters the system, it is assigned a (rather large) number of neighbors. When a file is requested, the search request goes only to the neighbors; if the result of the search is not satisfactory, the user can request to extend it to the neighbors’ neighbors, then to the neighbors of the latter, and so on until a number of steps fixed in advance has been reached (2 or 3, typically). In these systems, the role of the central server is limited to the technically peripheral but commercially essential function of distributing advertisement to the users.

Today’s multimedia file sharing systems are in general not limited to audio files, but allow their users to exchange videos and images as well. In many cases videos are musical, due both to the relatively slow transition of the users community out of the musical arena, and to the duration of these videos, which is typically that of a song (3–6 min.), leading to more manageable file sizes. The exchange of full-length films is rapidly developing although, due to legal problems with copyright owners, it is still largely an “underground” activity.

Finally, one should give a brief mention to the exchange of open source software that although, stricti ductu, does not fit in the area of multimedia as commonly intended, shows some of the same problems and uses some of the same techniques (noticeably, mime types). In the area of software, however, the different characteristics of the user community (which is formed both of developers and of users of software) appears to have steered the interest towards centralized repositories that can serve, at one time, as locations for program downloads, and as a meeting point for the developers.

The legal issues

Beyond its technical significance, multimedia file sharing is having important social and legal repercussions, probing and, in all likelihood, changing, the power relations within the music industry between musicians, recording labels, and music listeners. Ever since the legal confrontation that led eventually to the demise of Napster, multimedia file sharing services have been the point of contention between two different and often irreconcilable views of music and the musical profession.

The recording industry association of America (RIAA) considers virtually every form of musical copy and of musical exchange outside of its control as a form of copyright infringement. During the last 300 years (roughly since the Statute of Anne, in 1712 [Author]), the notion of copyright has evolved from that of a natural right of the author, designed to protect the author’s interests, to a form of good that, like any other good, can be bought and sold. This has made it possible, together with the possibility of the mechanical reproduction of sound, the creation of copyright holding organizations, such as the recording labels, with great economic strength and (at least in the US) considerable power to steer the political process. At first, in the case of Napster, the legal strategy of the RIAA has been to go after the service provider, facilitated, in this, by Napster’s architectural decision to have a centralized indexing server: this helped the RIAA make a credible case that Napster was actively participating in copyright violation, and not merely providing the technical tools through which the violation was perpetrated (which, in itself, would not have been illegal, much like a gun manufacturer is not legally liable for a murder). With the more decentralized systems such as gnutella or kazaa, such strategy couldn’t work, and the RIAA moved directly against the users of these systems, obtaining from the courts the authorization to force the internet providers to reveal the names of the possessors of unauthorized copies of copyrighted songs, a decision that has been widely regarded as a violation of the privacy of communication and of the civil rights of the users.

On the side of file sharing, the arguments were essentially two: a softer one, so to speak, held that, much like the repeatability of video tapes has not prevented the creation of a vast, copyright compliant rental market, providing in the end a major source of revenues for the very film studios that, in the beginning, were fighting the introduction of VCR, so, in the end, the file exchange phenomenon will result beneficial to those copyright holders that are willing to embrace it. The problem, in other words, would not be in file sharing, but in the timidity of the music industry, that is not able to see beyond the current business model.

The second, more radical, argument [Santini] sees copyright as a product of the historical circumstances of the industrial revolution, and sees the recording industry as a product of mechanical reproduction on a physical medium: much like the automobile made the horse carriage construction business disappear, so the disappearance of the recording industry based on copyright and of the “star system” in the new musical phase is a natural economic fact, and that the revenues of the vast majority of the world’s musicians depend on live performances, not on record sales.

In the meantime, other providers are experimenting with copyright compliant pay-persong services. An example of this is Apple, with its services connected to its audio player “ipod.” In these cases, however, we are typically not in the presence of file sharing systems, but of client-server style downloads.

Multimedia Fingerprinting - Collusion-resistant, c-secure, Traitor Tracing Codes, Random Distortion-Constrained Codes [next] [back] Multimedia Entertainment Applications - Concepts, Enabling Technologies, Enabling Technologies: Formats and Compression, Enabling Technologies: Networks

User Comments

Your email address will be altered so spam harvesting bots can't read it easily.
Hide my email completely instead?

Cancel or

Vote down Vote up

over 2 years ago

i will try to

Vote down Vote up

over 5 years ago

All THE MEGA UPLOAD FANS should definitely check TAWKLE OUT!