The Internet, File Sharing and The Darknet

Major Integrative Project

BA in eBusiness Systems

David Ganly

01/03/04

Acknowledgements

This one is for the team. [TN] all in. Thanks to the lads, particularly to Adnans Causevic for the help with the live tests of WASTE and all the other stuff.

David Ganly, March 1^st 2004

Executive Summary

The internet has allowed people from across the world to share knowledge in unprecedented ways. The speed at which technologies have been adapted to take advantage of this is phenomenal. With the advent of technologies such as the MP3 file format for music the ability to copy and share music of all types has fallen into the hands of the average home user. On a standard broadband connection available in this country entire albums of music can be downloaded within an hour.

Such file sharing includes all types of digitised data; both legal and illegal. In 2004 the average movie is available for download illegally within days of its release to cinemas. This is an untenable situation for those involved in the content industry; these works are protected under copyright. Copyright is designed to reward the creativity of individuals or organisations. The advent of global file sharing is undermining this. Right now various laws are being put in place to protect copyright and illegalise file sharing.

Laws alone will not be enough to stem the tide of file sharing and complex Digital Rights Management tools are being created to protect digital files and the right to access them. The biggest computing companies in the world (such as Microsoft, Intel and IBM) are working on a new architecture for computers called Trusted Computing. Among other things, Trusted Computing allows for far more robust and unbreakable Digital Rights Management (DRM) to be embedded into files.

Due to the illegality of the majority of file sharing, it has been pushed onto what has become known as the Darknet. The Darknet is defined as the collection of networks and other technologies that enable people to illegally share copyrighted digital files with little or no fear of detection. Various types of global and small world darknets have grown exponentially as sharing illegal files has become more popular.

This paper charts the rise of file sharing and its move onto the world’s darknets. Trusted Computing and specifically how it applies to DRM and copyrighted content is also analysed. The paper concludes with discussion on whether or not the darknet file sharing phenomenon can be stopped or if the days of copyright in the traditional sense are ending.

1.0 Introduction

The advent of the internet has had an unquantifiable impact on communications, lifestyles and the world of business. The internet is undoubtedly one of the most important inventions of the last century. It has opened doors for many new types of business and also has the unique ability to empower the innovation of individual users. – Never before has their existed an uncensored medium like the internet and this has led to incredible benefits (and some very real dangers) in various aspects of humanity.

One huge benefit of the internet is its ability to share files with little fear of censorship across international boundaries. However, this ability has led to the wide spread sharing of copyrighted materials. As intangible items such as movies, games, software etc. have grown in value so has the will to copy them illegally.

The objective of this paper is to examine the following research question: Now that various technologies and laws are emerging that have a huge impact on file sharing, what will be the fate of copyrighted materials on the internet?

This paper traces the details of file sharing and what various technologies and laws are impacting on it. The paper concludes with predictions of what will happen in these related areas in the future.

1.1 The Nature of a File

Material on computers is stored in a series of files. A single file could be a document such as this one, a piece of music, a photograph, or a more complex dataset containing elements of all of the above. One of the obvious reasons for storing data in such a way is that so it can be shared. Up until the 1990’s the most common way of sharing such data among end users was by transferring it to a physical medium (ie. a floppy disk) and moving it to another machine where it could be copied and stored. In the workplace, local computer networks existed where each individual computer was connected to all the others through a series of cables and servers. In this case files could be shared simply by browsing electronically to the relevant machine on the network and copying the file. These networks were predominately isolated from each other. Each company operated its own network.

1.2 The Internet

The massive exponential growth of the internet in the 1990’s changed the above maxim. With nearly all computers everywhere connecting to a global network through standardised protocols, nearly all computers everywhere could share files in a manner that had never been possible before.

The internet is an information medium; and it has become prolific in the information it can provide. This is because all the time it is becoming more and more ideally suited to the four types of media that are sought – text, audio, photograph and video. The buzz word to describe these formats taken together is multimedia. By it’s nature, content like this, once digitised, takes up large amounts of memory. Very quickly, hardware and software developers sought to digitise every thing more efficiently so that it could be shared over the limited speed connections of the internet. This gave rise to the now ubiquitous JPEG file format for images of all types, MP3 for audio, and MPEG for video. Improved codecs [1] have meant file sizes have dropped while the overall quality of content has been maintained.

As the formats became accepted, so sharing increased. Home computer users connected in their millions. Pictures were the most widely trafficked in the early 1990’s as they were small enough to be uploaded and downloaded on the narrowband connections of home users (< 56k, across standard copper wire telephone lines.) As the MP3 standard evolved, home users now had the power to encode CD quality music digitally, and music tracks started to be shared. Given the narrowband access that users had, circa 1998 in Ireland at least, it took around ten to fifteen minutes to download a four to five minute music track at sufficient quality.

The motivation for the protection of copyright has its roots in the idea that society should reward those who create original material. Copyright is defined as (www.dictionary.com, 2004): The legal right granted to an author, composer, playwright, publisher, or distributor to exclusive publication, production, sale, or distribution of a literary, musical, dramatic, or artistic work.. Copyright, under law, allows intellectual intangible content (such as music or software) to be owned by an individual or company. If copyright is infringed, legal remedies are available if the infringement is proven.

Web sites operated that provided unauthorised direct links to copyrighted files. However, due to their centralised client-server nature (ie. It is easy to find out who runs a web site) these sites were easily shut down; usually with letters from music/software organisations to the owners threatening legal action. It became clear that these web sites could only operate in small social environments; once they became too well known, (in particular, became listed on search engines) they then attracted the attention of the materials owners.

1.3 Client-Server and P2P Technology

Two architectures are important to the understanding of the evolution of file sharing and the dynamics of it. When a user views a web site, they are using a client-server architecture. – The client (the user) makes requests and the server carries out the necessary computation and returns a result to the client. In Peer to peer (P2P, sometimes also referred to as person to person) architecture, each user shares equivalent capabilities and responsibilities – For example all peers are capable of sending requests and handling them as well.

1.4 Napster

On June 1^st, 1999, software called Napster became available. Categorised as a P2P file sharing utility, it allowed users who were connected to the Napster service to view selected files on each others computers and download from each other directly. The key difference between the Napster P2P sharing utility and web sites offering files was that Napster held no content itself: Napster was merely a search & discovery system for the files; the files resided on the computers of its users. Once a song was located, Napster users sent & received files from each other. Napster brought P2P file sharing into the mainstream and allowed non-technical users to quickly and easily share music. Because of this, online music sharing exploded and a global network of file sharers evolved. Napster came out at the height of the dot com hype cycle, and as such received financial backing despite being a free service with its only revenue coming from advertising.

Figure 1 Napster architecture

Napster was not immune to legal attack due in part to one dynamic of its operation: It held a central database of all the files stored on its network; which users searched to locate files they wished to download. This database became the focus of litigation. Napster was brought to court and initially certain queries had to be removed from its network; e.g. Metallica – any references to Metallica music files had to be removed because these were copyrighted. Napster was then forced to police its network for any copyrighted material. As the amount of material on Napster shrank, so did the number of users. After a year in operation Napster was shut down by a federal court judge in a case brought by the Record Industry Association of America (RIAA) for copyright infringement.

1.5 Gnutella

Napster clones appeared, and these were far more difficult to shut down due to their de-centralised nature. A decentralised network is one that allows each user of the network to distribute information and to connect to other peers directly, without relying on a central server. These new de-centralised programs were based around the Gnutella protocol, which was created by Nullsoft in 2001. Gnutella operates by finding each connected peer in the network and allowing each peer to view files on other peer’s computers. Since it doesn’t rely on a centralised server or database, all that is needed for a Gnutella network is a few connected peers, but some networks have grown to millions. For example, Kazaa, a file sharing utility runs on a derivative of Gnutella called the FastTrack network. FastTrack works by allowing certain peers with fast enough internet connections to become ‘super peers.’ These super peers temporarily become hosts that index files held by other, smaller users to enable reliable and fast searching of the network. Because Gnutella is an open protocol, anyone can create a client to use the technology.

The Gnutella network and those similar to it have many legitimate uses, which makes it difficult to shut down. Email can be used to share copyrighted material, but it is in no danger of being shut down due to its legitimate uses.

Figure 2 On the Gnutella network, there is no centralised server. Each peer distributes information and executes queries on each other.

Now, in 2004, on a standard broadband connection available in Ireland an entire album of music can be downloaded within an hour. An entire movie, ready to be burnt on to a DVD or simply watched on computer can be downloaded in one or two days. Pre ‘cracked’ (their copyright protection broken) applications can be downloaded and installed as well.

1.6 Winning the battle against P2P

This situation is untenable for those involved in copyright protection. Copyright is intended to protect the rights of those who create. With the proliferation of online sharing, profits are being eroded and sales are falling. In the United States, the RIAA is aggressively pursuing a campaign of suing users identified as file sharers. In the Asia-Pacific region, the Business Software Alliance is targeting P2P web sites and using web crawlers (automated computer programs that ‘crawl’ the internet for specific information) to find copyrighted software. Once found, the BSA notify the relevant Internet Service Provider and the sites are taken down.

Various organisations around the world are at present trying to shut down the P2P file sharing phenomenon. In Australia, Sharman Networks, owners of Kazaa, one of the world’s most popular file sharing applications was recently raided by MIPI (Music Industry Piracy Investigations) as well as the homes of its top executives. This follows on from various claims by business organisations that Kazaa was primarily used for illegal file sharing. Right now, Kazaa has 2.9 million users connected and reports that 4.2 million gigabytes of data is being shared. A search under the band ‘coldplay’ reveals over 300 hundred matches, all of which is copyrighted material. [2] This small example shows that the business organisations are correct; Kazaa is used for illegal file sharing. In the last week alone, Sharman Networks reports 2.6 million downloads of the Kazaa application. Sharman networks make money from Kazaa through advertisements, through forcing users to install spy ware (monitoring software that catalogues a users actions on the world wide web) and by having files available to buy. Along with various search results, Kazaa would show legally available files that can be purchased with a credit card.

Undoubtedly, Kazaa will loose much of its current popularity. It’s global proliferation has made it a huge target for legal attack and it could easily be enjoined to police the network it operates on (FastTrack) for copyrighted material, much like Napster was in 2000. The raid by MIPI is seen as a first step in this direction. It has been shown by various clients using Gnutella or Gnutella-like networks that they can filter the material available on them; this has been done to block child pornography. Content industry advocates rightfully ask why this cannot be done just as easily for copyrighted materials.

1.7 A file sharing culture

As the amount of copyrighted material available through Kazaa decreases, so will the number of users. Why are people so interested in file sharing and so readily able to ignore the illegality of it? Those chiefly involved in illegal file sharing, namely teenagers and those in their early twenties have grown up sharing copyrighted materials. Before the evolution of the internet in the mid 1990’s, sharing was done through small networks of friends. Music was copied via tape cassette, and computer based material was transported physically with disks. While this type of sharing was prolific, usually the amount of material available was limited and the barriers to copying high enough to keep it small. - Someone in the group of friends had to have an original copy of a tape to make a first copy, subsequent copies made from the copies tended be of lower quality etc. Also, sharing things meant actually physically getting the material to the other person which increased the time it took to share things. However, the idea was engrained into culture and the advent of the internet was a natural progression in sharing.

Of all the reasons for file sharing one stands out; people share files because it is interesting to do so. People share files of all types for a large variety of reasons – to show friends, to enhance work, etc. At present, it is equally as easy to share a legal file as an illegal file.

It is often toted that if the cost of music, software and video was decreased, the public would be more inclined to purchase the music rather than pirate it. This is undoubtedly true but it is also true that people will always seek to share files, whether it is their own intellectual property or not.

A data study (Pew Internet & American Life Project, 2004) showing a significant drop in Kazaa usage was released in January 2004. The study points to the start of the RIAA’s campaign of legal action against file sharers and their data appears to support this. However, this decline in file sharing could be misleading; the file sharers are not necessarily stopping, all that is proven is that the high profile networks are losing users. What does seem to be shown is that ‘casual’ file sharers are being influenced. The Pew study shows a drop from 29% of all internet users in the United States engaging in illegal file sharing in May 2003 to 14% in December 2003. However, new less public file sharing applications are coming on stream and the Pew study ignored these in favour of the high-profile sharing applications such as Kazaa, WinMX and BearShare. Given the fluidity and contradictory nature of many studies on the amount of file sharers the only thing that appears clearly is that in the United States at least the mass-movement style file sharing popularised by Kazaa is being curtailed. However, even if there are only 10 million or so active file sharers in the United States they constitute a huge source of lost revenue for the content industry. File sharing by its nature is viral (meaning it travels quickly in various forms) and travels through many methods of distribution; a single file sharer among a group of friends will almost certainly spread files through un-traceable channels, namely through copied cds, or sent through instant messaging programs such as MSN Messenger or Yahoo Messenger. These programs allow small groups of friends to chat and send files to each other.

2.0 Intellectual Property Law

Intellectual Property is defined as ‘The general term for intangible property rights which are a result of intellectual effort. Patents, trademarks, designs and copyright are the main intellectual property rights.’(UK Patent Office Glossary, 2004) In terms of file sharing, it is illegal because such files are copyrighted. The International Federation of the Phonographic Industry (IFPI, 2004) describes copyright as the following: ‘Copyright is the means by which a person or a business makes a living from creativity. Copyright springs from a simple notion: the people that create, produce or invest in creative work should be the ones that decide how that work should be reproduced and made available to the public.’

Entire industries are based around intellectual property; music, movies, games, TV shows and software are all IP based. The RIAA (who represent over 90% of legitimate recording businesses in the US) states on its site (www.riaa.com, 2004) that taken together copyright based industries are worth 360 billion euro and account for more than 5% of gross domestic product. As stated above, the advent of convenient file sharing has impacted on the sales of all these industries. Taking the example of the music industry, the Record Industry of America’s mid year statistics show total worth on CD shipments at 4.6 billion dollars. An annual decrease of approximately 10% year on year is cited by the record industry as being caused by the explosion of online piracy.

Given the speed of change over the last decade, litigation is playing ‘catch-up’ with a very fluid and dynamic technological landscape. The internet and file sharing has significantly altered the content industries ability to counter piracy and counterfeiting. They find their intellectual property under attack from the average internet end user, as opposed to large scale counterfeiters.

2.1 The US Digital Millennium Copyright Act

In the United States in 1998, the Digital Millennium Copyright Act (DMCA) was enacted. It was passed in response to growing concerns over copyright protection and circumvention in the evolving digital environment. The act made it illegal to circumvent anti-piracy measures in copyrighted materials (such as software authentication measures or CD copy-prevention mechanisms.) The DMCA also contained provisions for limiting the liability of internet service providers in terms of the information carried over their networks. However, under the terms of the act service providers are expected to remove copyrighted material from their servers if found. The act allowed the RIAA to use the IP addresses (unique identifiers for each computer on the Internet) of known file sharers on file sharing networks (most notably Kazaa) to force Internet Service Providers in the United States to link those IP addresses to users personal information. This allowed the RIAA to file law suits against the users. The RIAA promptly did this and forced several ISP’s to identify users alleged to be file sharing based on their IP addresses. However, Verizon Communications, one of the United States largest ISP’s appealed the decision and in December 2003 the RIAA’s strategy was undermined when the court ruled that the subpoenas used by the RIAA infringed the privacy rights of individuals. The RIAA identified the users based on their IP address before the alleged file sharing had been heard in court. This practice was deemed illegal. This has not stopped the RIAA however, they are now filing ‘Jon Doe’ legal actions - filing 531 legal summons with just the IP addresses of alleged infringers. This means that the RIAA cannot settle out of court with alleged file sharers because they cannot be identified until legal proceedings are underway. It is the media attention that these law suits have gained, as well as the launch of legal online sources of files (see below) that is believed to be the primary reason for the drop in illegal file sharing in the United States.

2.2 The EU and Copyright

Enacted November 1^st 2003, the EU Copyright directive made downloading content such as movies and music illegal. It also made circumventing copy-protection mechanisms a crime punishable by law. This has already inconvenienced users; certain new music CD’s are designed to run only on stereos and not on computer CD players. This legislation makes circumventing this or any protection designed to prevent copyright theft illegal.

Currently, a new directive (Council of the European Union, 2004) is before the European Union relating to Intellectual Property. Free speech and consumer activists have dubbed the directive ‘the DMCA on steroids.’ The EU IP Directive is quite a far reaching piece of legislation. The legislation was originally intended to clamp down on criminal counterfeiting and piracy but the directive has now been widened in scope to include a wide range of IP copyright infringement cases. Now included is the possibility of RIAA/DMCA style law suits against individual file sharers in Europe. In particular, Article 8 of the directive provides for ‘Anton Pillar’ orders (secret court authorisations of raids for evidence by the plaintiff's agents) and for the ability to demand extensive personal information in the pursuit of legal action (Foundation for a Free Information Infrastructure UK, 2004). The Foundation for a Free Information Infrastructure UK organisation points out ‘In Europe these kinds of investigatory procedures are more usually associated with criminal proceedings with a much higher standard of proof, and are quite alien to the civil (non-criminal) justice systems of most of the member states.’ The directive leads the way for a far more hostile legal-action based intellectual property environment.

The main cause for concern over the directive is caused by the fact that it appears to be being rushed through the legal system and their has been very little public consultation. The final text of the directive was released 17^th February 2004. The directive will be up for consideration on the 22^nd February 2004. At time of writing, the directives future is not certain, but it is clear that the European Union Legal Affairs committee wishes to see the directive passed on it’s first reading. Many parties, including the UK House of Lords and rights activists organisations have serious concerns over various aspects of the directive and worry that it will over-rule members states current legal standings on IP law without due consultation.

3.0 Digital Rights Management and Trusted Computing

Digital Rights Management (DRM) is the management of intellectual property rights in the digital environment (www.epic.org, 2004). There is an inherent problem DRM faces: with the advent of broadband the barriers to getting individual music files has disappeared, so now people exchange entire albums and collections of music. So called ‘ultra-broadband’ is inevitable; at present standard broadband connections for home users run at 50 to 55 kilobytes a second. Once this rises above 1 megabyte per second, any data, whatever it be, will be available within minutes. Entire movies at broadcast quality will be available. It is estimated (O’Brien, 2003) that by 2010, a single hard drive will be able to store every song ever recorded. (There is estimated to be about five million recorded songs in existence.) Even today, in 2004, it is not uncommon to see people with music collections in the tens of thousands. It raises serious questions: Why would anyone buy a DVD when they can download it for free? How can record/movie/software developers continue to make a profit when users can acquire their products for nothing?

‘The internet treats censorship as damage and routes around it’ (Gilmore, 1993) This quote has proved to be depressingly true for DRM activists. Any time they manage to get a court injunction stopping a particular system (e.g. Napster) another three pop up in its place. (e.g. Kazaa, BearShare, Bit Torrent) The inherent problem is the control of a file. Once its on a computer, the person who owns the intellectual property rights to that file has no control over what happens to it. At present, material stored on a computer is controlled by the user of that computer. In that, regardless of the origin of that material, once it resides on the users machine then he/she has sole command over its use.

3.1 Current DRM

At present, digital rights management systems are meeting with some success. One example of a DRM system in use is the activation built into Windows XP. Within 30 days of installing the operating system, it must be authenticated, which for most users means allowing it to connect to a Microsoft server to verify it is indeed a legally purchased copy. This is being circumvented through the availability of illegal XP master CDs which require no such authentication to operate.

Another form of DRM that is gaining significant media focus at present is that of Apple iTunes and the re-launched Napster 2.0. Currently both stores are only officially available to US residents, but similar European based services are becoming available. These two online music stores allow music tracks to be downloaded at $0.99 a track or $9.95 an album. The rights to the songs and what the user can do with them is controlled via DRM technology. In the case of Napster, it uses Microsoft Windows Media Player 9 to control the media files and how they are used. Under their DRM system with purchased music you can perform: Unlimited burns of individual tracks, unlimited transfers to supported Portable Players and copy your music to up to 3 computers

However, the systems are not fool proof and Jon Johansen, famous for hacking the region encoding on DVDs released a file in November 2003 that has shown weaknesses in the iTunes DRM protection. A quick Google search reveals methods for circumventing the protection in a myriad of ways.

True and complete control of a file is needed for illegal file copying to be stopped. This is the issue that is currently being worked on by some of the biggest computer companies in the world. Their question: ‘How can we control access (the rights) to our intellectual property?’ Their solution is Trusted Computing.

3.2 Trusted Computing

‘Trusted Computing (TC) refers to the totality of protection mechanisms (hardware, firmware and software) that provide a secure computing environment. TC includes everything that must be trusted -- access control, authorization and authentication procedures, cryptography, firewalls, virus protection, data backup, and even human administration -- in order for the right level of security to work.’ [3] (Webopedia.com, 2004)

Current PC’s are what are technically known as open systems. This means that they can be changed to behave maliciously towards another system. An example of this would be hacking another computer over the internet or initiating a virus attack through an email client. On the other hand there are closed systems, examples of this would be an ATM machine. The system is closed, maintained and monitored by a central system and as such has very limited vulnerability to attacks.

Trusted computing is an effort to bring elements of closed systems to the personal computer operating environment, which up to now has been largely an open system.

Currently, Trusted computing is being rapidly developed by the worlds leading technology developers (e.g. AMD, Intel, Microsoft, Philips, Sony, Sun Microsystems) The companies have formed the Trusted Computing Group and are working together towards standards to enable trusted computing.

Microsoft has coined the acronym NGSCB – Next Generation Secure Computing Base, (pronounced ‘enscub’) which contains the core elements of Trusted Computing. Trusted Computing has huge benefits in all areas of security. It enables software to run in a wholly secure environment. It allows each application running on a PC to run absolutely independently; for example a malicious attack exploiting a security flaw in a web browser would be confined to the web browser; it could not execute commands outside of that application. This would provide huge benefits in areas such as virus protection, data stability and in limiting the ability of hackers to compromise a system or network. A cynic might suggest that big business could release viruses that attack non-trusted computing systems in order to force trusted computing take-up. What is more likely is that they will just stop supporting non-TC-enabled products thus forcing upgrades.

3.3 Trusted Computing Technology

In Microsoft’s outline (Microsoft.com, 2004) of the trusted computing base they define four major technology changes needed. These are:

3.3.1 Memory curtaining

Memory curtaining means each program running on a PC cannot interfere with (as in read/write) to another programs memory. Using this technology, even the operating system is prevented from accessing memory, so the system is much more secure against malicious attack.

3.3.2 Secure input and output

Secure I/O refers to securing the path from the keyboard to the application and from the application back to the screen. This stops programs such as key loggers (programs that record what a user types) and screen grabbers (programs that record what displays on screen) from operating.

3.3.3 Sealed storage

Sealed storage secures data created in a given application and prevents it being opened by another user or computer program. At present, if you add a password to a file or your operating system, the passwords are stored alongside the files so they can in most cases be quite easily circumvented. With sealed storage, a unique system of hardware and software encrypted keys is used to prevent such access. If a user manages to get a copy of a protected file onto their machine, they will not be able to open it due to the lack of appropriate keys.

3.3.4 Remote attestation

Considered by many as the most significant and controversial aspect of the proposed trusted computing architecture remote attestation allows for the authentication of hardware and software on a given system. So if you, another user, a virus or spy ware alters one of your applications or your operating system it would be detected and other users would be alerted. For example, each time the system is started or an application executed data is sent to a Microsoft server which checks that everything is certified and then allows the application to continue working.

Taken together, the above four high-level technologies aim to create a system by which data on a PC acts like a safe – certificates are needed to gain access to the data and only those with rights are allowed to manipulate the data. Attempts to circumvent the system will leave evidence that such an attempt was made. Up until now, these technologies have been implemented with software. Software is inherently easier to circumvent, and it is hoped that by using hardware supported architecture a trusted platform can be created.

3.4 Trusted Computing and DRM

Those interested in the protection of intellectual property have realised that in order for files to be secure, radical new thinking like Trusted Computing is needed to regain control over their property. Currently, once in the users possession data can be manipulated and copied in an unlimited fashion. Trusted computing can determine the level of control that the user has, even when the data resides on their hard drive. For example, secure video and audio prevent the files from being ‘freed’ – their rights management removed allowing the file to be copied.

In the Trusted Computing Group’s (TCG) Frequently Asked Questions page (www.trustedcomputinggroup.org, 2004) question 29 states: ‘Was TCG formed to specify Digital Rights Management technologies? ‘ And the answer: ‘No. It is not a goal of the TCG to enable or embed digital rights management (DRM) technology in computing platforms….’ It goes on to say: ‘…The architecture, merits, and uses of any application that makes use of TCG building blocks and software interfaces is entirely the responsibility of the developer of the application. Hence, any issues related to applications should be directed to the application.’ So no, Trusted Computing is not specifically designed for DRM. But it provides the base for DRM applications to run securely.

With trusted computing comes personal certificates, meaning that a remote server can accurately identify exactly who it is that is connecting to their web site. This increases accountability for actions. Microsoft have stated that it will be possible to turn features like these certificates off. But to use popular applications of the future, e.g. Online shopping, the certificates would have to be switched on. It is thought that eventually users would just leave the certificates on all the time as they become needed for a large amount of applications.

In The Digital Imprimatur (Walker, 2003) the point is made that circumventing trusted computing will become increasingly difficult as it’s adoption spreads. Bugs and exploits will be taken advantage of, but these will be fixed and in time ‘the screw will tighten’ until circumventing Trusted Computing requires too much effort and skill to make it a significant threat.

Also made possible by DRM and Trusted Computing, watermarking allows a permanent and invisible mark to be embedded into content. This way, the watermark could be detected by hardware or software in a system and the right to access/play the file determined. Another possible technology is fingerprinting. This means that when a file is sold, the file is fingerprinted with the buyers details, making them far more reluctant to share them, because the original source could easily be identified by investigators. Corbis, an online provider of stock photography and news prints are using watermarking and finger printing in their photos to stop them being used in unauthorised works.

3.5 Trusted Computing and File Formats

Trusted computing contains another controversial element; file format locking. A core part of DRM, if a music file is downloaded, it is essential that it can only be opened by the DRM enabled application. – If the file can be opened by another program, then that program can easily convert the file to a non DRM enabled format such as existing MP3 files. Trusted computing extends this ability; file formats will be locked by hardware dependent keys to specific applications. This means that word documents would only be viewable in Microsoft word and it would be impossible to open the document in another program. Many see this as anti competitive. The ability to permanently lock out competitors creates monopolistic software; to an even greater degree than currently held by Microsoft’s Office suite of applications. Perhaps more importantly, it removes the potentially damaging law suites that can result from a competitor trying to read proprietary file formats; suing a company for trying this usually results in a negative backlash for the originator so removing this possibility is a big positive for dominant software vendors.

3.6 Micropayments

One major reason for pursuing trusted computing is to allow innovation in online commerce due to enhanced security. At present, an online transaction requires a user to provide authentication every time they make a purchase, even if they are returning customers. While some sites offer ways around this (like Amazon.com’s one click ordering system), by and large making an online purchase is a cumbersome experience requiring the filling of numerous forms. Trusted computing allows credit card and personal information to be tied to the users PC meaning a user could simply click once on any web site to make a transaction that would then auto deduct from their relevant account.

Micropayment is seen by some as the way forward to solving the file sharing crisis. While people still object to paying €1 to access a song, they may object far less to paying small amounts such as €0.10 for popular songs and continue to pay €1 for more obscure ones. The public may not be as adverse to paying small amounts if this became established as the norm for soliciting information; such as micropayment subscriptions every time a newspaper web site is accessed.

3.7 Its all about trust

Trusted computing has many positive uses. However, one fundamental issue becomes obvious when the issue is studied: Trusted Computing treats the user as a hostile, and gives more power to external forces: your PC manufacturer, software developers and those who develop content. Trusted computing limits what a user can do with their own system in order to decrease security risks and enhance stability. However, by limiting the system in this way, they also improve how companies can make money from users.

Whitfield Diffie, chief security officer at Sun Microsystems and the inventor of public-key cryptography (a system for ensuring information remains encrypted when it is past between different parties) said The Microsoft approach “lends itself to market domination, lock out, and not really owning your own computer. That's going to create a fight that dwarfs the debates of the 1990's.” (Merrit, 2003.)

Trusted computing is a warm term; it sounds like it will make a computer much more reliable and less open to attack. This feeling lies in the word trust. An excellent insight into this in provided in the Trusted Computing Frequently Asked Questions. (Anderson, 2003.)

The author cites the example of a doctor patient relationship. A patient trusts his doctor to keep his medical records secret. The patient doesn’t trust his postman to do so, because the postman doesn’t have the records, so he can’t affect the patient’s opinion that his medical records should be secret. However, the doctor can, so the fact that he is in a position to harm him is really what is meant when it is said that the patient trusts him. If you strip away the emotional aspects of trust that is what it is. Anderson concludes his FAQ with the following question and answer:

Q: So a ‘Trusted Computer’ is a computer that can break my security?

A: That's a polite way of putting it.

4.0 The Rise of the Darknet

As discussed above, the unauthorised sharing copyrighted material is not new. The advent of technologies such as tape recorders and photocopiers created panic in industries because they changed what end users were able to do with protected materials. The reason the internet is a cause for such concern for related industries is because it allows the sharing of such materials on a massive, global scale; it is estimated that one tenth of the world’s population – about six hundred million –(www.glreach.com, 2004) have access to the internet. Technological innovations have created convenience for file sharing; namely lower latency.

Four Microsoft Corporation researchers released a paper titled ‘The Darknet and the Future of Content Distribution.’(Biddle et al, 2002) A Darknet is defined as The collection of networks and other technologies that enable people to illegally share copyrighted digital files with little or no fear of detection.(wordspy.com, 2004.)

The first of these darknets they called the ‘sneaker net.’ As in, friends physically distributing disks and tapes to each other. They traced the evolution of the darknet onto the internet and into the forms discussed already in this document; namely the original Napster and Gnutella networks.

According to the Microsoft research paper the idea of a darknet is based on the following assumptions:

1. Any widely distributed object will be available to a fraction of users in a form that permits copying.

2. Users will copy objects if it is possible and interesting to do so.

3. Users are connected by high-bandwidth channels.

Assumption one has proven true so far; innumerable albums have been leaked before their release and a significant proportion of movies sent to awards boards (e.g. The Oscars) have been intercepted illegally and encoded for download. Assumption two has been proven simply by analysing the activities of the now wide spread file sharing phenomenon. Assumption three can now apply in Ireland given the advent of affordable broadband. It has been true for several years in the United States and in other EU countries.

Under these terms Napster and Gnutella clients such as Kazaa are all forms of darknets. These darknets have severe weaknesses, the main one being lack of anonymity. The internet service provider of a user can be identified via the IP address of a Kazaa user; the internet service provider can then identify the exact user. This technique was used successfully in the United States to sue Kazaa users sharing copyrighted materials. This lack of end-point anonymity also creates another weakness in current darknet structures; free riding. Currently, a huge proportion of illegal data is held by only a small number of ‘super-peers.’ The majority of users of the Gnutella network download from these peers but do not allow uploads from their computers. This makes the darknet susceptible to attack; if these super-peers are disconnected then a large proportion of the copyrighted content will be removed from the darknet.

4.1 Global Darknets

Given the movement towards stricter controls on intellectual property; namely technology (complex DRM and trusted computing) and law (the digital millennium copyright act in the US and the new EU directives) it would seem that the days of the large file sharing networks are numbered. The fact that no one ‘owns’ the new networks has long been established, and it appears as though the key to stopping file sharers, in the short term at least is by education and litigation. The global networks of today are far too obvious a target for content industry litigation to survive for long, or at least to survive with a significant amount of copyrighted material being shared.

4.2 Small Worlds Darknets

Given the increasing exposure of users of global darknets to legal attack, another option is sought by file sharers. Harking back to the days before the internet, small networks of friends are sharing files with each other through innovative, secure software such as WASTE.

On May 28^th 2003 WASTE appeared for download on the Nullsoft (the developers of the Gnutella protocol) web site. Within hours of its release, the file was retracted and a message was put up on the site stating: ‘If you downloaded or otherwise obtained a copy of the Software, you acquired no lawful rights to the Software and must destroy any and all copies of the Software, including by deleting it from your computer. Any license that you may believe you acquired with the Software is void, revoked and terminated.’ (www.nullsoft.com, 2003) The software was released under the General Public Licence (www.gnu.org, 2004) which means anyone can improve/edit the software as long as they include the source code and credit is given. For this reason, WASTE has been adopted by open source developers and its development continued despite the above warnings.

The parent company of Nullsoft, America Online, one of the worlds largest internet service providers refused to comment on WASTE or why it was removed from the site. Commentators surmise that WASTE was removed because AOL didn’t want the negative publicity associated with releasing software that could so easily be used for illegal file sharing and seemed designed for such a purpose. Others surmise that WASTE was removed because it created conflict with AOL’s Instant Messenger software – it allowed small networks of friends to chat and share files, but not in such a secure environment.

WASTE was described on the Nullsoft page (the page is no longer available, but mirrors of the page are) as ‘WASTE is a software product and protocol that enables secure distributed communication for small (on the order of 10-50 nodes) trusted groups of users. WASTE is designed to enable small companies and small teams within larger companies to easily communicate and collaborate in a secure and efficient fashion, independent of physical network topology.’

The encryption level used in WASTE to keep inter-party communication secure is higher than that used to encrypt credit card transactions in present day online shopping systems. It uses public key/private key encryption, which means that in order for friends to connect to each other through WASTE, they must first swap public keys. Public keys are strings of alpha-numeric characters which when used in tandem with private keys (which are protected by passwords) which are known only to the user allow for secure communications to occur.

Figure 4 Waste showing 1 live connection. (invented IP’s shown.)

Figure 5 Waste’s file browser in action. (details edited for demonstration purposes, these folders don’t actually exist anywhere.)

WASTE is so ideally suited to illegal file sharing due to the nature of its architecture. In order to gain access to a private WASTE network a user must be trusted enough to be given another users public key. In this way, large scale disruption of small world networks would be extremely costly to enforcement agencies due to the amount of undercover work needed to gain access. Even if one WASTE network is compromised, it has little or no bearing whatsoever on other ones.

While content would not move as quickly through various small world networks as it would through a unified global one it would still percolate through networks through ‘friends of friends.’ One person who is attached to two or more darknets would pass content from one to the next.

Figure 6 Interconnected small world Darknets: Users who are members of more than one darknet spread information between separate networks. Adapted from The Darknet and the Future of Content Distribution.(Biddle et al, 2002)

4.3 Darknet technologies

The diversity and different technological basis for the types of Darknet shown is indicative of the amount of innovation in the area. WASTE is just another example of a flourishing type of darknet. Some other in use darknets are outlined below:

4.3.1 IRC, xDCC and DotGet

Internet Relay Chat (IRC) has been around since 1988. It allows users to connect to servers using IRC client software and to chat and interact with millions of users all over the world. Built into IRC’s capabilities is Direct Client to Client (DCC) which allows users to directly connect to each other without a server in between. This allows users to send files to each other. Building on this capability, xDCC means users can now use search engines such as www.packetnews.com to search ‘bots’ on various IRC networks for files they wish to procure. These bots are software programs which can automatically handle DCC requests from users, automatically putting users in a queue for a particular file and then dcc sending the file to the user. The motivation behind these sharing networks appears to be enhancing the proliferation of legal and illegal file sharing and spreading all types of material to a global audience.

DotGet, another IRC based file sharing tool brings all of the above features of IRC and xDCC together in the mIRC client software and uses the IRC network to create a Kazaa like decentralised network complete with searching.

4.3.2 DC++

Another popular and open-source (meaning anyone can change and improve it) Darknet client is DC++. DC++ uses the DirectConnect network and allows individual users to become ‘hubs’ which other users can connect to and download files from. DC++ removes the free rider weakness by only allowing a user to connect to a hub if it meets that hub’s required for shared files. Each hub determines the amount of files needed to allow access.

4.4 Freenet

The only truly unstoppable global network would be an anonymous one. So far, this hasn’t been achieved. An attempt at such a network is currently underway. It is called Freenet. On the Freenet website, under the ‘what is Freenet?’ question it states: ‘Freenet is free software which lets you publish and obtain information on the Internet without fear of censorship. To achieve this freedom, the network is entirely decentralized and publishers and consumers of information are anonymous. Without anonymity there can never be true freedom of speech, and without decentralization the network will be vulnerable to attack.’ (Clarke, 2003)

Completely free of charge and funded solely by donations, Freenet builds on the decentralised nature of systems such as Gnutella while at the same time guaranteeing anonymity. A major motivation behind Freenet is the enforcement of freedom of speech, regarded by many as one of the central human rights. Censorship in any form is counter-democratic; and Freenet is designed to ensure the free flow of pure information without censorship from any body, be it organisational, corporate or governmental.

On the Freenet philosophy page (Clarke, 2003) copyright is discussed. The author (the creator of Freenet) states bluntly: ‘You cannot guarantee freedom of speech and enforce copyright law.’ His reasoning behind this statement is that to enforce copyright means communications must be monitored in some fashion, therefore complete freedom of speech and copyright law are incompatible. In essence, Freenet is designed to be an un-policable network.

4.5 Freenet Technology

Freenet is a decentralised network; much like Gnutella. This means it is based on a series of nodes connecting to each other with no central server involved. When a user connects to Freenet, a section of their hard drive is assigned for Freenets use only. This new ‘node’ in the Freenet network receives queries from other nodes and forwards them on. This storage space is essentially owned by Freenet; the user does not decide what gets stored in it. When a file is added by a user, the file is stored on some nodes (not decided by the user, assigned based on heuristics in Freenets programming) and the file is also assigned a globally unique identifier key (GUID) which is entirely location independent. A file may migrate and replicate through nodes throughout its lifetime depending on its popularity. Freenet decides what files to keep and remove based on the number of requests for a file. This way, popular data filters through the network quickly while irrelevant, damaged, or outdated data is removed. As nodes respond to more searches, they build up data on what’s available on neighbouring nodes. This way searches become faster as nodes increase in popularity and make connections with new nodes.

This type of architecture is perfect for the secure distribution of copyright works. Together with heavy encryption, it would be very difficult to remove copyrighted content once it was on Freenet. Also, due to the popularity-based indexing, new files (such as new movie releases) would be guaranteed to spread quickly while older ones would be culled from the network.

Already in use in areas such as China, where government censorship of the internet is common place, Freenet proves anonymous networks are possible. However it is not without significant problems: Freenets architecture is inherently difficult to improve and it suffers from long search times and overloading of popular nodes. These issues are being worked on but updates are a long time in coming. Because of its donation-based funding, Freenet only has one full time programmer working on it aside from its creator. Due to this, its evolution has been extremely slow. However, even if Freenet fails, the idea is sound and it is inevitable that other attempts will be made at an anonymous decentralised network.

5.0 Conclusions

Two main outcomes of the file sharing revolution seem possible. The first is that the internet ‘genie’ is put back in its bottle and falls under the control of businesses and governments across the world with technology based on the trusted computing principles discussed previously supported by appropriate legislation. The second outcome is that darknets continue to grow in power, popularity, convenience and speed until the idea of copyright on intangible objects becomes obsolete due to the availability of any material, copyrighted or not, instantly. The ramifications of both outcomes are huge. Which is more likely?

Given people’s aversion to ‘big brother’ style government and the right to privacy, it seems difficult to imagine a democratic government that could force the use of trusted computing style technologies. It would seem that trusted computing will enjoy a limited implementation; it will be adopted for secure tasks and will bring many benefits to many fields, but users will always want choice, and most importantly, the ability to switch it off. For example, laws that were passed in the United States in the aftermath of the September 11^th attacks on the World Trade Centre in New York to enhance security (e.g. the patriot act) are now the subject of heated debate due to their heavy impact on the privacy of individuals.

5.1 The Open Source Revolution and Trusted Computing

Open source software is software that is provided with the source code so that it can be edited and improved upon. The Linux operating system is the best known open source project, a collaboration of software developers all over the world it is constantly being worked on, customised and improved. Businesses and end users world wide are starting to realise that they can make significant cost reductions by adopting open source technology; it cuts out expensive licence fees paid to companies such as Microsoft and can also provide far more customised solutions.

If Microsoft pushes ahead with its Trusted Computing initiative built into its future operating systems (Longhorn, the code name for the next major Microsoft Operating System release and the first to contain definite TC elements is scheduled for 2006) they will, as already stated, at some point in the future limit what users can do with their own computers. Linux and other open source projects are beginning to provide an adequate alternative which in the future would allow users to turn away from Microsoft. Why would any user choose to use a system that limits what they can do with the technology they have purchased? For this reason it is difficult to see how Microsoft will be able to implement a total trusted computing system.

If a hardware manufacturer builds only trusted computing usable components, then they will lose out to a manufacturer that builds both. Competition could well be a major issue in keeping trusted computing from being universally adopted.

It is believed that Trusted Computing based solutions such as certificates will become necessary for common tasks such as online shopping – but if this is the case then users will want a system that can do this and do everything else. So they will want a system where they control what elements are trusted.

5.2 Removing Choice and regaining control

As shown above, it is unlikely that a user is going to accept trusted computing in its entirety. However, governments could try and force its adoption through laws. Again, common sense dictates that no citizen of a country is going to allow a law to pass that compromises their freedom in such ways. In a democratic country, privacy and freedom of speech are one of the most fiercely protected rights and anything that takes away from it is fiercely opposed. Governments will always seek to regulate a medium in order to exert control over it and businesses will seek to do so to extract profit. Trusted computing represents a big step in making control and censorship possible.

5.3 Darknets: An unstoppable force

Darknets of some size and description will always exist. Global darknets will likely rise and fall just as Napster and Kazaa have. The key to their continued success is innovation by interested individuals – programs such as WASTE and Freenet; released completely free of charge are evidence of peoples interest in the continuation of a completely censorship free internet. Small worlds Darknets will gain in popularity if global networks are shutdown entirely. The advent of a truly anonymous network such as that proposed by Freenet would be incredibly damaging for the content industries.

Companies and organisations charged with protecting copyright constantly debunk the idea that file sharers can become anonymous and escape detection on the various darknets. While at present there is an element of truth in this, what is also true is that as darknet technology improves, so will the cost of locating and identifying users who are engaged in illegal activities. There must come a stage where the number of users being sued drops to a small enough level as to not create a sufficient deterrent. Also, once a system is developed to identify a significant number of users on a given network then the next Darknet will already be rising. – The very nature of innovation means that their will be a lag between identifying a service and successfully compromising it.

5.4 Technology is moving too quickly to protect copyright

As already stated, people’s interest in copying material will never fade because it will always be interesting to do so. Even if new Digital Rights Management technologies including ultra encryption and fingerprinting technologies evolve there is a problem. The technology behind handheld digital camcorders is improving at a rapid pace, and it is very likely that at some stage in the near future it will become extremely difficult to differentiate between a digital version of a movie bought from an online store and a version created by pointing a camcorder at an appropriate display; e.g. a cinema screen. This technology completely ignores complicated anti-screen grabber or other DRM protection technologies. In the same way, people will always be able to record audio by placing a microphone next to a speaker, and technologies in this area can only improve as well.

5.5 A world of ends [4]

The internet has been described by Craig Burton, a Linux and open source advocate as ‘a hollow sphere comprised entirely of ends.’ There is no centre to the internet. The internet is comprised entirely of end points (individual computers, whether they be servers, ATM’s or single PC’s) with each end point connected to every other end point. The internet is an idea, and in that sense no one owns it. The internet protocol is an open protocol and needs no conditions to enter. No company controls the internet. No one needs to worry that tomorrow the internet will be upgraded and they’ll need a new machine to access it. – Maintaining the internet is in the hands of every single user.

Anyone can improve the internet. Email and the World Wide Web are perfect examples of improvements made to the internet. To be a ubiquitous improvement such as email or the world wide web the improvement needs to be for everyone. It needs to be open-source and easy to understand. This is why the above ideas flourished while others have failed. Instant messaging applications, which allow quick and easy chat and file sharing among friends, while a good innovation, have failed to become widespread because of the lack of open standards and the wish of companies to own and control them.

The internet is an entirely stupid network. The internet only knows how to route bits from one location to another. In this way, it is a reliable network, if a part of it fails then the data routes around it. Censorship of the internet must occur on its ends; and when it is looked at in this way then it is easy to see that total censorship is impossible. The darknet is simply another series of ends at the edge of the hollow sphere.

5.6 A levy on non-commercial file sharing

In the near term, legal commerce in all its forms must continue to compete with the Darknet. The Darknet provides a low cost and high quality service to users. In order to compete with this legal commerce must adapt and offer more incentives.

Recently, the Electronic Frontier Foundation in the US made an interesting proposal with regard to legalising file sharing. ‘Voluntary Collective Licensing’ would involve file shares paying a fee such as $5 a month in order to be allowed to share legally. This money would be collected by a central organisation and then be redistributed to artists and their labels based on popularity. A similar system already operates for radio stations. The RIAA has been quick to debunk the idea citing the fact that it values all music as the same, which would drastically impact their profits. The EFF proposal is interesting because it is a solution based on the fact that file sharing is here to stay in all its forms. This proposal is not new, it was originally proposed in a research paper in 2002 by Neil Weinstock Netanal, a professor of Law at the University of Texas. The RIAA and similar organisations still believe that they can ‘beat’ the file sharing phenomenon and drive uptake of legal systems such as Napster 2.0 and Apple iTunes.

At present, the record industries idea of competing with the darknet in the United States at least is to offer copy protected songs for $1 and to aggressively litigate against users of the darknets. These songs are protected. If the consumer goes and buys a new CD, then it is most likely going to be copy protected and wont play in their PC. It is a crime to circumvent such protection.

It seems that society as a whole has to choose whether to save copyright or not; in that suggestions such as the Voluntary Collective Licensing be taken seriously and other such forward thinking proposals be put forward. The RIAA and organisations like it that believe their businesses can continue without significant change are misguided.

The cost of renting a movie or going to the cinema is low enough so that businesses in this area remain much less affected by the file sharing phenomenon than the music industry. However, this situation will change as the speed of internet connections increases and the time taken to download a movie drops from days, to hours and finally to within minutes.

5.7 The end of copyright

Seeing as it seems impossible to stop the copying of any type of digital data it is inevitable that the copyright of intangible digital assets will become irrelevant. All of the techniques used by the content industries and governments so far have treated the user as a hostile with good reason; the user will continue to copy regardless of the actions of others.

Without copyright entire industries (which according to the Record Industry Association of America account for 5% of Gross Domestic Product) will fall apart. Artists will continue to create; creativity is part of human nature and with or without monetary reward it will continue. What’s far more confusing is, for example, the issue of big-budget Hollywood movies. How can a sixty million dollar film be given the go ahead if within days of its release (or possibility even before) it is available to millions of internet users at no charge?

Software has the advantage of having benefits outside of the actual computer program itself; for example people are willing to pay for software if it promises them technical support, updates or other indirect value. Software is also not as vulnerable as content-files to attack given its far more dynamic nature ie. you can’t point a camcorder at a screen and record a software program. It appears that mass-market intangible consumer goods like music and movies will be the largest casualties of the file sharing epidemic.

5.8 An inevitable ending

It’s inherently difficult to know if you are inside a bubble that’s in danger of bursting, as many dot-com investors found out in 2000. However, towards the end of that particular hype-cycle, things started happening that showed that the fun and games were over and a heavy dose of reality was needed. Much like then, warning signs for the end of the copyright bubble are appearing. The enormity of the issues at stake completely dwarf the dot-com blip of the late 1990’s - but the inevitability of it all stands out in stark contrast to the stop gap solutions presented by the content industries.

The reality is that the content industries must accept the file sharing phenomenon and legalise it in some way such as the Voluntary Collective Licensing system. Continuing to attack file sharers will just push them further onto more technologically advanced darknets.