PEER-TO-PEER CONTENT DISTRIBUTION
An Analysis of the Internal and External Potentials and Obstacles
Jorn De Boever
Centre for Usability Research, K.U.Leuven, Parkstraat 45(b 3605), 3000 Leuven, Belgium
Keywords: Peer-to-peer systems, content distribution, SWOT analysis, adoption.
Abstract: The popularity of (illegal) peer-to-peer file sharing has a disruptive impact on the internet traffic and the
business models of content providers. In addition, several studies have found an increasing demand for
bandwidth consuming content. It is widely recognised that illegal file sharing applications demonstrated the
potential of peer-to-peer systems to distribute popular content in a scalable and inexpensive way. However,
there has been relatively little economic analysis of the potentials and obstacles of peer-to-peer systems as a
legal and commercial content distribution model. Many content providers encounter uncertainties regarding
the adoption or rejection of peer-to-peer networks to spread content via the internet. The recent launch of
several commercial, legal peer-to-peer content distribution platforms, which are still mainly beta versions,
increases the importance of an integrated analysis of the internal strengths / weaknesses and the external
opportunities / threats. This paper explores these internal and external aspects by means of an integrated
SWOT analysis.
1 INTRODUCTION
Peer-to-peer systems have been associated with
illegal file sharing since the introduction of Napster
in 1999 (Shirky, 2001). But, this negative image has
shifted because of the arrival of innovative peer-to-
peer systems that demonstrate the potential of these
applications for legal content distribution.
Nevertheless, we still notice that the adoption of
peer-to-peer systems, to distribute content in a legal
way, is still not widespread. In this context, the
question arises which opportunities and challenges
peer-to-peer systems offer for legal and commercial
content distribution. Some technical aspects of peer-
to-peer networks – e.g. performance, information
retrieval, and self organisation – have been
extensively studied in recent years. However, there
has been relatively little analysis of the economics of
peer-to-peer systems. Consequently, we consider it
to be important to have a broader more business
oriented approach toward peer-to-peer platforms as a
content distribution model.
The introduction of innovations causes a certain
level of uncertainty for the industry, which must
decide whether or not to adopt it (Rogers, 2003).
The industry will try to compensate this uncertainty
by exploring the different aspects of an innovation
or by looking at what other companies have decided.
That is why it is important to understand why
companies would (not) adopt peer-to-peer systems
and with which uncertainties the industry still
struggles. We conducted an extensive SWOT
analysis (Strengths, Weaknesses, Opportunities and
Threats) by means of interviews and literature
review. The aim of the present paper is therefore to
briefly explore the obstacles and potentials of peer-
to-peer technology to distribute commercial content
in an efficient and legal way. This study will allow
us to evaluate the current business viability of peer-
to-peer technology. The results of this analysis
should enable us to understand how peer-to-peer
networks might be positioned strategically in order
to raise the chances for a successful commercial
application for legally exchanging media content.
This paper comprises a study that might assist
content providers in the adoption process.
This paper starts by reviewing some relevant
literature in the area of economics and peer-to-peer
systems, followed by the elaboration of the
methodology. The next section presents the results
of the SWOT analysis by enumerating these aspects.
The final section will confront and link the internal
and external elements so as to put meaning and
interpretation to these results.
367
De Boever J. (2008).
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles.
In Proceedings of the Fourth International Conference on Web Information Systems and Technologies, pages 367-377
DOI: 10.5220/0001517103670377
Copyright
c
SciTePress
2 LITERATURE
Substantial research effort has been devoted to the
technical aspects of peer-to-peer technology. This
has resulted in an extended body of literature about
architectures (e.g. Androutsellis-Theotokis and
Spinellis, 2004; De Boever, 2007), availability (e.g.
Bhagwan, Savage and Voelker, 2003; Chu, Labonte
and Levine, 2002), security (e.g. Divac-Krnic and
Ackerman, 2005) and etc.
As we stated in the introduction, this paper will
elaborate more on the economics of peer-to-peer
technology. One of the most characteristic aspects of
peer-to-peer systems is that – like in virtual
communities – to a large extent, value is being
created by the end users. End users are more than
solely consumers in that they at least also cooperate
to distribute the content (Lechner and Hummel,
2002). MacInnes and Hwang (2003) analysed four
cases – KaZaA, Kontiki, SETI and Groove – of
peer-to-peer business models. According to these
authors, the most important challenges for the
business models were the revenue model, security
and user behaviour. In the study, conducted by Rupp
and Estier (2003), different actors in the music
industry were compared on how they were being
confronted with and affected by new technologies
such as peer-to-peer networks. Other studies
explored the revenue models of several peer-to-peer
systems (Hummel, Muhle and Schoder, 2005;
Hummel, Strømme and La Salle, 2003). The revenue
models of most existing file sharing applications
failed on efficient allocation. This means that the
revenues do not always end up with the rightful
claimants. It is obvious that these studies already
identified several of the issues that lead to
uncertainty about the viability of peer-to-peer
networks as content distribution models.
We will elaborate, integrate and build our
findings on the studies that are stated below. The
papers that will be integrated in our SWOT analysis
are the following: Smith, Clippinger and Konsynski
(2003), Hughes, Lang and Vragov (2005), Kwok,
Lang and Tam (2002), Sigurdsson, Halldorsson and
Hasslinger (2005) and Rodriguez, Tan and
Gkantsidis (2006). The combinations of these
studies with interviews provide us with a detailed
analysis in which several new aspects have been
identified. Although we value the results of the five
selected studies, we argue that these papers suffer
from some flaws which add to the significance of
our contribution. For that reason, the drawbacks of
these studies require further examination of the
internal and external potentials and obstacles. We
will now contend why these studies are deficient to a
certain extent.
The study of Smith, Clippinger and Konsynski
(2003) mainly reflects on the use of peer-to-peer
technology within organizations (internally),
whereas our analysis emphasizes peer-to-peer
technology as a content distribution model in a
commercial B2C (Business to Consumer)
environment. Another limitation of their study is the
fact that they only held discussions with CIOs,
which results in only one point of view. In contrast,
we have tried to recruit respondents with different
backgrounds.
The research of Hughes, Lang and Vragov
(2005) provides more of a theoretical, analytic
framework for market design of peer-to-peer
networks than it offers empirical results. We have
integrated some elements of this framework in our
study so as to corroborate our results. Another
limitation of this study is that they primarily focus
on the constraints – legal, technical, economic,
structural, political, cognitive and socio-cultural –
while not paying attention to the strengths and
opportunities.
Although Kwok, Lang and Tam (2002) use the
words “risks and opportunities” in the title of their
article, they limit their examination to the free riding
problem, the related motivations for cooperation,
and piracy. In addition, the analysis of these authors
is merely from a theoretical perspective.
Furthermore, these academics do not pay a lot of
attention to the opportunities.
Sigurdsson, Halldorsson and Hasslinger (2005)
provide an analysis that primarily focuses on
technical and financial aspects while ignoring e.g.
user-related aspects. Furthermore, we pose that these
authors have failed to make the different concepts of
a SWOT analysis operational, which resulted in
confusion about the terms.
Finally, Rodriguez, Tan and Gkantsidis (2006)
described the results of a workshop on the issues
regarding legal peer-to-peer content distribution.
This paper offers a list of items that stresses the
weaknesses and threats, while paying little attention
to the strengths and opportunities.
In this paper, we will integrate the results of
these five studies, while addressing the flaws of
former research in our own analysis.
3 METHODOLOGY
Our methodology consisted of three levels: a
literature review, expert interviews and the data of
WEBIST 2008 - International Conference on Web Information Systems and Technologies
368
the former two levels were used in a SWOT
analysis. The first phase of our methodology
comprised a detailed literature review. We already
gave an overview of some of these articles above.
From this literature review, we selected five papers
to elaborate and integrate in our SWOT analysis. We
only selected studies that covered a broad range of
issues involved in peer-to-peer research and
excluded others that only focus on one or a limited
number of topics so as to make sure that
overrepresentation of some topics would be avoided.
In the second phase, we conducted 15 in-depth
interviews with people from the industry and with
some academics. We have interviewed people with
different backgrounds to obtain as accurate and as
objective results as possible. Therefore, participants,
with different backgrounds (social – 3, economic –
3, technical – 8, and law – 1), were selected from
different actors for these interviews. We interviewed
respondents working for broadcasters (4), other
content providers (4), ISPs (2), research community
(4) and copyright lobby organisations (1). Of the
researchers, two are working on research projects in
which a peer-to-peer application is being developed.
Based on the results of these interviews, we were
able to accomplish a significant contribution to the
existing literature by broadening the scope.
The findings of the former two phases resulted in
the data for the SWOT analysis. First, we clearly
defined what was meant by the concepts strengths,
weaknesses, opportunities and threats so as to avoid
confusion. Second, we scored the different topics
that were mentioned in the results section of the
papers and in the interviews for our analysis. Based
on the defined concepts, we decided – by means of
qualitative interpretation – for each item to which
category it belonged. Third, based on literature
review and the results of the in-depth interviews, we
examined why the mentioned items were issues.
Finally, we tried to confront or relate the items of the
four categories by analysing e.g. which strengths
might be used to address some weaknesses or to take
advantage of certain opportunities; which
weaknesses might be mended by further research;
which threats must be avoided and etc.
Although SWOT analyses are mainly used
within organisations (Vermeylen, 2004; Sabbaghi
and Vaidyanathan, 2004), we argue that it is an
effective tool as well to grasp the major positive and
negative issues of applications and technologies.
While studying technologies or information systems,
researchers run the risk of considering the
technology or the adoption of it as a positive fact.
This has also been termed as the pro-innovation bias
(Rogers, 2003). The danger of the pro-innovation
bias is that it becomes difficult to have a neutral and
objective perspective. The advantage of a SWOT
analysis is that the researcher is forced to analyse the
weaknesses and threats as well. A SWOT analysis
consists of two main parts, namely the internal
aspects and the external aspects (Vermeylen, 2004).
The internal aspects consist of the strengths and
weaknesses. The external aspects contain the
opportunities and threats. The external elements are
the occurring changes, events, and etcetera that are
taking place in the environment.
In the past and today, peer-to-peer networks have
been very successful in sharing files, mostly in an
illegal way. Peer-to-peer applications to share files
or distribute content in a legal way are not
widespread yet. The SWOT analysis will reveal
several internal strengths / weaknesses and external
opportunities/ threats of peer-to-peer as a
commercial application for the legal exchange of
media content (video, images, text and film).
4 RESULTS
In this section, we will explore the internal strengths
/ weaknesses and external opportunities/ threats. We
will define what we understand under these concepts
within each part of this section. The data in this
section result from 15 interviews and five selected
papers.
4.1 Internal Strengths
The strengths encompass the internal aspects that
offer peer-to-peer systems for instance possibilities
to take advantage of certain factors in the
environment. These are the strengths that are
characteristic of peer-to-peer systems.
Cost Savings. Peer-to-peer is a technology that is
capable of reducing the costs for content distribution
because it utilizes fewer resources such as
bandwidth, storage capacity and etc. (Norton, 2007;
Smith, et al., 2003). In addition, the avoidance of
centralised servers limits the costs in terms of
investment and upkeep (Sigurdsson, et al., 2005).
The lower costs make it possible for the industry to
serve a larger audience or to distribute more
bandwidth consuming content. This might lower the
need for compression techniques so that the content
has better quality. This makes peer-to-peer networks
an interesting technology for the distribution of
bandwidth consuming content. The implementation
of techniques, such as swarming, makes it possible
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles
369
for end users to more efficiently download larger
content. In other words, the industry does not have
to buy that much distribution capacity anymore. It is
also an inexpensive solution for the worldwide
distribution of content. There is no need for
adaptations on the network level, which reduces
costs as well. All these aspects result from the fact
that peer-to-peer systems utilise the available
resources that are available for instance on the end
users’ personal computers. Norton (2007) has
examined and compared several content distribution
models based on a cost analysis. The result was that
peer-to-peer systems were the most inexpensive
content distribution models.
Scalability. Peer-to-peer systems scale with the
amount of use, which results in limited marginal
costs (Ding, Nutanong and Buyya, 2005; Taylor,
2004; Sigurdsson, et al., 2005). The extra costs for
each additional user are restricted because each
additional consumer in theory also means additional
resources. These are positive network externalities
of peer-to-peer systems which makes them very
scalable. Some peer-to-peer networks are even able
to cope with ‘flash crowds’ (i.e. a sudden and
extremely increase of the number of users that
consume the same service or content which might
overload a system). Peer-to-peer technology might
be utilised to develop overlay networks so as to
distribute content around the globe. Furthermore,
these overlay networks are cost-efficient as the costs
for implementation and maintenance are limited
(Sigurdsson, et al., 2005).
Easy Implementation. From the position of
content providers, peer-to-peer systems are easy and
inexpensive to implement because it operates more
on the application layer. To distribute content on a
worldwide scale, there is little need for large
investments in infrastructure and control. However,
some peer-to-peer providers pose that the opposite is
true. It is more complicated to start a peer-to-peer
system in comparison with installing a central
server.
Fault Tolerance. Peer-to-peer systems are fault
tolerant because there are limited centralised
components in the network that are potential
bottlenecks or single points of failure (Sigurdsson, et
al., 2005). Because of the distributed nature of peer-
to-peer networks and the redundancy of nodes, the
disappearance of one or several nodes mostly has
only limited impact on the overall performance.
Availability. Some respondents consider peer-to-
peer systems to be more efficient because these
systems utilise the idle resources that are available in
the network. This leads to more availability of
resources such as bandwidth, storage capacity, and
content.
4.2 Internal Weaknesses
The internal weaknesses contain the distinctive
characteristics of peer-to-peer networks that are
disadvantageous. These are the direct constraints of
these systems.
Quality of Service (QoS). The reliability of peer-
to-peer systems in terms of QoS is a major issue for
the commercial viability of these systems
(Parameswaran, Susarla and Whinston, 2001;
Sigurdsson, et al., 2005; Hughes, et al., 2005;
Rodriguez, et al., 2006). QoS in this context relates
to the performance of the network for instance on
the level of delays, content availability, distribution
speed, scalability and etc. The end users’ personal
computers operate as nodes in the network. This
makes it hard to guarantee service quality because
these nodes can suddenly appear and disappear, and
there are still end users with low bandwidth
connections. Content might become unavailable if
nodes, that host certain content, suddenly leave the
system. This might result in delays and high buffer
times. This is especially problematic for large
content that takes more time to download. If the
download requires more time, than there is more
chance that the content exchange cannot be
completed because the hosts become unavailable.
The more decentralised a peer-to-peer network is,
the more that QoS is an issue. The service quality
might also influence the real time experience of
peer-to-peer live streaming applications. Peer-to-
peer content distribution is in this sense a best effort
model with low guarantees of service quality. The
uncertain QoS is probably the most important
obstacle for content providers to adopt peer-to-peer
systems for content distribution. In addition, it
remains unclear whether users will accept that
content, from professional content providers, that is
being shared in a legal way, is unavailable or
requires delays. In summary, it seems difficult to
guarantee a certain performance level and it is
therefore hard to formulate reliable service level
agreements.
Inefficient use of Resources. Most peer-to-peer
systems handle the available resources – e.g.
bandwidth – in an inefficient way. This results from
the fact that there are mostly no centralised
components in the system that manages the use of
these resources because peer-to-peer networks are,
to a large extent, self organising systems. This self
organisation requires much traffic to maintain the
WEBIST 2008 - International Conference on Web Information Systems and Technologies
370
operability of the system. This was e.g. a significant
issue in a former Gnutella version (e.g. Ripeanu,
Iamnitchi and Foster, 2002).
Security and Privacy. These issues are still
worrisome within the context of peer-to-peer
systems because of decentralization (Taylor, 2004;
Smith, et al., 2003). Peer-to-peer implies that other
peers can access your computer to retrieve certain
data. This makes it easy for malicious nodes to enter
your computer and cause damage. This danger for
malicious attacks has social implications as well,
related to trust (Rodriguez, et al., 2006). In most
peer-to-peer systems, users encounter unknown
anonymous peers they should trust if they want to
exchange data. Inflicting damage in this trust
relation might cause users to abandon peer-to-peer
systems. The ‘older’ generations of peer-to-peer
systems are mainly anonymous without actual social
interaction, which limits the trust relation among
peers. In addition, the lack of centralised monitoring
facilities causes difficulties in guaranteeing security
(Sigurdsson, et al., 2005). By using peer-to-peer
networks, users expose private data such as IP
address, the content they have consumed and etc.
This can inflict damage to the users’ privacy
(Rodriguez, et al., 2006).
Usability and Adoption. These are two issues that
received little attention in peer-to-peer research and
it proved to be a concern for some of the
respondents. Most peer-to-peer file sharing
applications have complex interfaces that are little
intuitive. Several terms (e.g. seeders, trackers and
etc.) might have no meaning to some users. Good
and Krekelberg (2005) have conducted usability
research on the KaZaA user interface. They came to
the disquieting result that many users were not able
to locate their shared folder, so they did not know
which content they were actually sharing. This is
probably the reason why these researchers found
personal data on KaZaA such as e-mails and even
credit card data. Today, peer-to-peer systems are
mainly used by persons who regularly surf the
internet and who probably know how peer-to-peer
networks function. If one wants to develop peer-to-
peer applications to be consumed by a wide
audience, the user interface needs to be more
accessible, usable and learnable for users who are
less familiar with peer-to-peer systems (e.g. Morris,
2007). For instance, the fact that users have to
download and install a separate client might be a
threshold to adopt these systems. Furthermore, file
sharing might frustrate users as they have to wait
until the full file has been downloaded before
consuming it (Rodriguez, et al., 2006).
Costs to Users. Peer-to-peer systems utilise
resources that are available on the end users’
personal computers. The users’ resources – e.g.
storage capacity, bandwidth, electricity – are being
used for the distribution of content, which entails
considerable costs for these users. It remains an open
question whether the average user will accept that a
content provider utilises the user’s resources to
distribute content.
Business Model. Peer-to-peer systems are being
confronted with indistinctness’s and strains
regarding several elements of business models.
There is still no proof of concept of a clear,
comprehensive business model that integrates peer-
to-peer systems and where there is a fit between the
several business model components. The main
bottlenecks in this area, according to most
respondents, are related to control. The lack of
control over large parts of content distribution is
problematic to some content providers. In addition,
there are often no means to measure the
consumption of the peers, which makes it difficult to
attract advertisers. There exists no proof of concept
of peer-to-peer systems that integrate an advertising
or payment model for revenues. Peer-to-peer
systems therefore need a centralised component
which monitors the consumption of the peers in
order to be able to implement a revenue model.
4.3 External Opportunities
The opportunities contain the external elements to
which peer-to-peer systems can anticipate from its
internal assets. In other words, these topics consist
of the aspects in the environment that might have a
positive effect on the adoption and use of peer-to-
peer systems. The opportunities of peer-to-peer
networks often result from the flaws of other
distribution models.
Costs centralized Models. In a client/server
network where there is a situation of augmenting
popularity of bandwidth consuming content – such
as audio, video and games – the costs for
distributing this content will increase. Furthermore,
the risk of a bottleneck and single point of failure
increases as well. Researchers have noticed an
increasing use of and demand for bandwidth
consuming content such as video, audio, HDTV and
etc. (IDC, 2007). These researchers predict that the
amount of digital data will expand even more in the
following years. Some experts predict that video
content will have a share of about 80% of the
internet traffic (Norton, 2007). For centralised
distribution models, this implies additional costs in
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles
371
the sphere of for instance distribution capacity,
storage capacity, bandwidth, servers, control and etc.
Disadvantages of other Distribution Models. We
will demonstrate some opportunities for peer-to-peer
systems by addressing some flaws of other
distribution models such as unicasting, multicasting
and CDNs (Content Delivery Network). Unicasting
is a centralised streaming method that has the same
drawbacks as client/server models. Unicasting does
not scale with the amount of consumption, it runs
the risk of a bottleneck and the costs rise as content
is consumed more often. Therefore, most content
providers that utilise a unicasting model, limit the
quality of the content to confine the costs when
popularity increases. Multicasting is a more
decentralized streaming method, but this technique
has some drawbacks as well. Multicasting requires
adaptations in the network infrastructure and many
networks are not multicast enabled. Therefore,
multicasting is not suitable for worldwide content
distribution. Furthermore, multicasting is mainly
interesting for live content distribution and less for
on demand content. A third example of a distribution
model is a CDN, which is a distributed network of
computers to distribute content. An elaborated
analysis of CDNs can be found in the text of Pallis
and Vakali (2006). When content providers decide
to utilise a CDN, they first make strategic decisions
about the necessary performance and QoS and
include these requirements in a service level
agreement. But, it is often difficult to predict the
popularity of services or content which might
influence the service quality.
Increasing Capacity of end Users’ Equipment.
The end users possess an expanding amount of
resources – such as bandwidth, storage and
processing capacity – that can be exploited by peer-
to-peer networks (Sigurdsson, et al., 2005). This also
means that there are more idle resources available on
the edge of the internet. Peer-to-peer systems have
become a success because of the availability of
unutilised resources. Users have become more than
purely consumers in the value chain and they now
add value by replicating and distributing content in
peer-to-peer systems.
Evolution toward Symmetric Networks. The
internet today is still organised in an asymmetric
way, i.e. it expects users more to download and less
to upload content. This is why users have more
download capacity in comparison with their upload
capacity. But this trend is shifting because of
developments such as peer-to-peer file sharing
networks and, more recently, user generated content.
Several respondents expected the internet to become
more symmetrically organised in the future so that
the down-/upload ratio will gradually equals one.
This might be an opportunity for systems such as
peer-to-peer networks that require users to utilise
their upload capacity to serve content.
User Orientation. Several respondents posed that
peer-to-peer networks focus more on the end users
by giving them more roles in the value chain. In
peer-to-peer file sharing systems, a social filtering
process occurs in which the community of peers
decide which content might remain available and
which content not. Content that does not fit the
users’ needs, or that becomes less relevant, will
slowly disappear from the system. In other words,
peer-to-peer systems offer the opportunity to involve
the end users more closely in the value chain.
4.4 External Threats
The external threats are factors in the external
environment that can interrupt or threaten the
success of commercial and legal peer-to-peer
systems.
Data Limits. ISPs or network providers still
maintain data limits in several countries as a part of
their business model and to be able to guarantee a
certain level of performance and control over the
traffic. If a content provider wants to distribute large
content such as video files, end users will soon attain
their data limit. This is problematic since more
content providers want to distribute video over the
internet in high definition. If users exceed the
allowed limit, they need to pay for the extra
consumed volume.
Asymmetric Networks. The commercial internet
has been build from a top down perspective. The
content owners distribute content, and the users
download this content. In other words it focuses
more on the download link, which resulted in
asymmetric organisation, e.g. ADSL. This
asymmetric architecture causes low upstream rates
(Sigurdsson, et al., 2005). For the efficient operation
of peer-to-peer systems, it would have been better to
have a more symmetrical relationship between up-
and downloading – for instance SDSL – because
peer-to-peer systems emphasise uploading as well.
In the current asymmetric organisation of the
internet, there is a risk that the upstream bandwidth
capacity will not be sufficient to compensate the
downstream bandwidth capacity in a situation of
widespread peer-to-peer traffic (Norton, 2007). If
the use of peer-to-peer systems would be ubiquitous,
the ISPs fear that the upload capacity might become
congested during peak hours.
WEBIST 2008 - International Conference on Web Information Systems and Technologies
372
Measurability and Control. In server oriented
models, it is easier to keep track of the different
flows of data in the system which is a prerequisite
for advertisers and content owners (Smith, et al.,
2003; Sigurdsson, et al., 2005). This measurability is
important to the revenue model within the business
model. This measurability of other distribution
models is a threat to peer-to-peer networks and it is
an internal weakness as well.
Competing distribution Models. By means of
some examples, namely multicasting and CDNs, we
will demonstrate that peer-to-peer systems are being
threatened by some characteristics that other models
display. In the assessment of using a certain content
distribution model, content providers will balance
the pros and cons of these systems against each other
in order to make the right decision. Multicasting, for
instance, is a streaming model that offers some
answers to the flaws of the unicasting model,
particularly in the area of live streaming. In addition,
multicasting is a system that reinforces itself so that
it is very scalable. Furthermore, multicasting can
offer better QoS compared to peer-to-peer
streaming. In this sense, multicasting can be
considered as a more reliable model that is mainly
suitable for popular live events. A second example
of content distribution models contains CDNs. In
contrast with peer-to-peer systems, CDNs offer high
guarantees on the level of QoS which makes it
possible to draw up reliable service level
agreements. In general, CDNs are considered to be
reliable content distribution models that are able to
address several security issues.
Position of ISPs. We will now explain why ISPs
might not sympathize with peer-to-peer content
distribution models (Rodriguez, et al., 2006).
Content providers, that implement peer-to-peer
technology for content distribution, avoid bandwidth
expenses by passing the costs on the users. The users
will try to prevent these costs by making sure that
they do not exceed their data limit. The result of the
adoption of peer-to-peer technology is an increasing
traffic volume on the network infrastructure of ISPs,
while these ISPs receive less revenues.
Measurements studies have demonstrated that more
than 50% of internet traffic is emanating from peer-
to-peer systems (Haßlinger, 2005; CacheLogic,
2004). The increasing traffic volume (especially the
upstream traffic), caused by peer-to-peer systems,
originates from the fact that peer-to-peer technology
allows exchanging large files and because there is
much message overhead required to self organise the
system. In addition, most peer-to-peer applications
lack locality awareness which implies that a lot of
traffic crosses several ISP networks whereas the
content might be available on a node of the local ISP
(Rodriguez, et al., 2006). In other words, the current
revenue model of ISPs does not fit a situation in
which the use of peer-to-peer networks is pervasive
because the traffic volume might not be in
proportion to the revenues. If large traffic volumes
are congesting the network, ISPs might decide to
squeeze, queue or even block peer-to-peer traffic.
We have to remark that ISPs will not be likely to go
that far because these actions can cause
unfavourable churn rates. The consequence might be
that many ISPs change their price policy from e.g.
flat rates into a system of layered pricing according
to the traffic users generate (Rodriguez, et al., 2006).
Legal Issues and Image. Peer-to-peer technology
has to cope with an image problem, in which there is
still a connotation of illegal file sharing, piracy and
sharing of pornographic content. This is why some
authors (e.g. Rodriguez, et al., 2006) argue that legal
uses of peer-to-peer content distribution should be
promoted. The MP3 encoding format, combined
with the exchange possibilities of peer-to-peer file
sharing, have lead to a situation in which illegal file
sharing has become an omnipresent and accessible
practice that is not likely going to disappear (Kwok,
et al., 2002; Smith, et al., 2003; Sigurdsson, et al.,
2005). Decentralisation makes it difficult to monitor
and manage peer-to-peer networks. Making content
downloadable is difficult from a business
perspective, because content owners fear copyright
infringements. If users are able to store content on
their hard drives, the chances increase that files are
being copied and spread in an illegal way. Control
over the rights remains an important issue. However,
several respondents posed that control of copyrights
was not their major concern. They emphasised the
importance of control of their business model. These
respondents accept for instance that content is being
copied and distributed by the end users, as long as
they are able to capture these ‘eyeballs’ for
advertising purposes. On the other hand, most of the
interviewees hold an opinion that solutions such as
DRM are not suitable for fighting piracy because (1)
every version will be compromised sooner or later
and (2) it does not fit the consumers’ needs. In
addition peer-to-peer providers have to cope with
different legislations in different countries. For
instance, the USA has developed the fair use
doctrine, whereas European jurisdiction utilises an
exception list. For a more in-depth analysis of
jurisdiction concerning illegal file sharing in
different countries, we refer to (Werkers and Gilio,
2006; Dogan, 2005).
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles
373
Critical Mass. The operation and viability of
peer-to-peer systems require a critical mass of
cooperating users that make sufficient resources
available (Rodriguez, et al., 2006; Kwok, et al.,
2002). In client/server systems, it does not matter
that content is consumed by little people as long as
the server stays available. If a new peer-to-peer
application is being introduced, it must generate
sufficient mass so that content would be sufficiently
available. But, as is the case in many virtual
communities, peer-to-peer networks are being
threatened by the free riding phenomenon (Adar and
Huberman, 2000; Hughes, Coulson and Walkerdine,
2005). Free riders are users that consume resources
from the network while not contributing anything in
return. In a situation without critical mass and a
large amount of free riders, content will often be
unavailable. This causes users to leave the system
with the result that the peer-to-peer network is no
longer viable. Consequently, peer-to-peer
technology is mainly interesting for the distribution
of popular content because it is easier to attain a
critical mass of cooperating peers.
Content Aggregation. Several peer-to-peer
platforms – such as Babelgum and Joost – operate as
content aggregators as well. Peer-to-peer platforms
try to acquire popular content to increase the
popularity of their service. A part of the revenue
model of several peer-to-peer platforms is based on
advertising and revenue sharing (e.g. Warner Music
Group, 09.04.2007). There are several competing
peer-to-peer companies trying to gather the same
content providers. As more peer-to-peer systems will
start up, it will become more difficult to aggregate
popular content providers and as a result to attain a
critical mass. This threat of content aggregation is
linked to the issues of competing distribution models
and the acquisition of a critical mass.
Decreasing bandwidth Prices, Network Address
Translation (NAT) and Abandonment of Open
Source. The first issue concerns the relevance of
peer-to-peer content distribution if the bandwidth
prices would decrease significantly. Peer-to-peer
content distribution is mainly regarded as a
technology to save on bandwidth costs. As
bandwidth capacity might become cheaper, the
question arises whether a content provider would
still adopt a distribution model that performs badly
in terms of service quality. The second issue, NATs,
involves mainly a technical burden that needs to be
solved (Rodriguez, et al., 2006). The use of a NAT
makes that several hosts within a private network
only use one IP address to have access to the
internet. The use of NATs is increasing for instance
because of the use of wireless routers. The exchange
with nodes that operate behind a NAT router
generates a technical issue, by limiting end-to-end
connectivity, which needs to be addressed. The final
issue relates to the abandonment of open source
because of commercialisation of peer-to-peer
technology. Most commercial peer-to-peer providers
use closed source. Closed source software is not
subject to peer review with the result e.g. that it
might be engineered poorly and that it is less
reliable.
5 CONFRONTATION
Some people criticise SWOT analyses for being just
an enumeration of several points without any
reflection (Vermeylen, 2004). We therefore will try
to discover links between the internal and external
aspects that have an impact on peer-to-peer systems.
Peer-to-peer systems can take advantage of some
external opportunities by deploying the internal
strengths. Research and experiments might be
necessary to address the internal weaknesses, while
peer-to-peer providers are sometimes not able to
change the external threats.
The results of the SWOT analysis indicate that
peer-to-peer networks still have not been able to
break through their infancy. The identified
limitations of peer-to-peer technology still entail
major issues on the way to a legal and commercial
content distribution model. The relation between the
strengths and opportunities is certainly existent.
Particularly, peer-to-peer networks can be perceived
as inexpensive, fault tolerant and scalable content
distribution models. These internal assets can be
used to address some external opportunities. Peer-to-
peer technology provides solutions for some of the
flaws of other content distribution models, such as
the client/server model (e.g. bottleneck, not
scalable), unicasting (e.g. bottleneck, not scalable),
multicasting (e.g. network adaptations). The costs of
several current distribution models will increase as
demand for popular bandwidth consuming content
grows. Peer-to-peer technology distributes content in
a cost efficient manner that requires less limiting the
quality of the content by compression techniques. In
addition, the possible evolution toward symmetric
networks might create opportunities for peer-to-peer
systems to utilise the users’ uplink even more.
On the one hand, there are several weaknesses –
for instance on the level of usability and business
modelling – that still can be mended. These are
elements that require further research and
WEBIST 2008 - International Conference on Web Information Systems and Technologies
374
experimentation, but these are not insurmountable
problems. On the other hand, other weaknesses (e.g.
QoS and security) and threats (e.g. legal issues,
control) require the integration of a centralised
component in the peer-to-peer architecture.
Moreover, as we indicated, a business model
requires that companies are able to monitor and
measure the consumption of users, to support for
instance the advertising model. To pursue these
goals, a centralised component is necessary in the
architecture. Finally, to improve the QoS of peer-to-
peer systems, it might be an option to close
agreements with ISPs to equip its infrastructure so
that the traffic becomes more manageable. In this
way, ISPs might be able to generate revenue from
the additional load on their infrastructure, caused by
traffic of peer-to-peer networks. For instance,
BitTorrent has closed a service level agreement with
GNi, a company that offers data networking
solutions, so as to be able to offer better user
experience (Del Conte, 28.06.2006).
Most of the external threats and opportunities are
beyond the manipulation of peer-to-peer providers.
Peer-to-peer providers can only take advantage of
the opportunities by promoting the criticalities of
these aspects (e.g. demand for bandwidth consuming
content, costs of centralised models and etc.). On the
level of threats, a strategy might include avoiding
these elements, although this seems to be
impossible. Peer-to-peer providers cannot ignore
issues such as data limits, asymmetric networks and
the control of ISPs. Every peer-to-peer provider, that
offers solutions for the distribution of popular
bandwidth consuming content, is being confronted
with these threats.
We will now describe a worst case scenario, in
which peer-to-peer content distribution is massively
used for legal content distribution and in which the
threats have not been addressed. In a situation in
which peer-to-peer systems would be used by the
average internet user, and not only by young more
computer literate users, the positive network
externalities of peer-to-peer networks might fall into
negative network externalities. The network
infrastructure of the ISPs would be overloaded
which would result in insufficient QoS, which would
reveal itself in for instance delays. To retain control
of the traffic, ISPs might decide to squeeze or to
block peer-to-peer traffic which would decrease user
experiences. Because of these inadequate user
experiences, users might decide to stop using the
peer-to-peer system with the result that content
would be less replicated throughout the network.
The unstable availability of content will cause
additional loss of users, which results again in less
available resources. In other words, this vicious
circle would prevent peer-to-peer systems to attain a
critical mass resulting in the extinction of peer-to-
peer systems. This is a worse case scenario that is
not likely going to occur because peer-to-peer
providers will try to avoid or address these
weaknesses and threats.
6 CONCLUDING REMARKS
The present study examined the internal/external
potentials and obstacles of peer-to-peer systems to
distribute commercial content in an efficient and
legal way. Data gathered from interviews and
literature was utilised to conduct an extensive
SWOT analysis from a business perspective.
Overall, this study reflected previous research
findings (Smith et al., 2003; Hughes et al., 2005;
Kwok et al., 2002; Sigurdsson et al., 2005 and
Rodriguez et al., 2006). We argue that our research
succeeded in addressing the limitations of former
research as was discussed in the literature section.
Interestingly, by conducting interviews with several
respondents with different backgrounds, this SWOT
analysis added several new elements to the existing
results in a structured way. Whereas most
researchers primarily examined the internal
weaknesses, this study highlighted the importance of
the internal strengths and external
opportunities/threats as well. In addition, this
examination paid important attention to the position
of the users, since the users are essential in the value
creation within the peer-to-peer value chain. On the
one hand, users have more resources – e.g.
bandwidth or storage capacity – at their disposal,
which creates opportunities for peer-to-peer systems.
On the other hand, the use of peer-to-peer systems
implies that users experience additional costs
because they have to make these resources available
for other peers in the system. Furthermore, usability
requirements of peer-to-peer systems have received
little attention. This lack of knowledge might
impede the widespread adoption of peer-to-peer
systems among end users. Finally, a large proportion
of users, in a great deal of peer-to-peer systems, are
free riding on the contributions of others. There is a
need for incentives to increase cooperation so as to
ensure network performance.
Beside users, we focussed on another often
forgotten actor as well, namely the ISPs. Peer-to-
peer systems cause an increased load on the network
infrastructure of ISPs, while the ISPs are less
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles
375
remunerated. This causes a negative attitude of ISPs
toward peer-to-peer systems, which might motivate
the ISPs to hinder peer-to-peer traffic.
In comparison with the five selected papers, our
research added the following topics: strengths (easy
implementation, availability), weaknesses (usability
and adoption, costs to users, business models),
opportunities (disadvantages of other distribution
models, evolution toward symmetric networks, user
orientation) and threats (data limits, competing
distribution models, content aggregation, decreasing
bandwidth prices, abandonment open source).
In contrast with former studies, this examination
goes beyond a mere enumeration of aspects by
confronting and linking the internal and external
elements. Several external developments offer
opportunities for peer-to-peer technology. In this
sense, peer-to-peer content distribution can be
envisaged as an inexpensive, fault tolerant and
scalable model for content distribution that offers
some advantages in comparison with other content
distribution models. Some of the internal
weaknesses can be addressed by further
examination, for instance on the level of usability
and business modelling. The QoS, peer-to-peer
systems produce, remains the major obstacle for a
widespread adoption of peer-to-peer technology
among content providers. Most of the external
threats and opportunities are beyond the control of
peer-to-peer providers. Peer-to-peer providers
cannot avoid issues – such as asymmetric networks,
data limits – which threatens their commercial
position.
Despite of the above mentioned positive qualities
of this analysis, we have to address the limitations of
this study as well. The reliability of this examination
might be threatened by the underrepresentation of
some actors in the interviews. We only conducted a
limited amount of interviews with copyright
lobbyists (1) and ISPs (2). However, we did not
manage in recruiting peer-to-peer providers for this
analysis which might have affected the results. In
this sense our results might primarily represent the
issues identified by content providers. In addition,
many issues have been raised involving users
whereas this research did not gather data from end
users. Indeed, a promising line of study would be to
interview peer-to-peer providers and end users so as
to improve the reliability of the results.
Overall, the results of this study imply that peer-
to-peer technology offers some opportunities for
commercial and legal content distribution. The
recent launch of several legal, commercial peer-to-
peer systems – e.g. Zattoo, Babelgum, Joost, Vuze –
have demonstrated the feasibility of these models.
However, the results of this study indicate that
several issues still need to be addressed. Centralising
some elements of the architecture might be
necessary to address the weaknesses – e.g. security,
QoS – of this content distribution model. In
conclusion, we pose that further development and
examination of peer-to-peer technology is still a
prerequisite for commercial viability.
ACKNOWLEDGEMENTS
FLEET (an interdisciplinary research project on
FLEmish E-publishing Trends) is an IWT SBO
project, with research partners IBBT SMIT, Cemeso,
LSTS, MOFI, ICRI, CUO, MICT, ECDC and TNO.
REFERENCES
Adar, E. & Huberman, B.A. (2000). Free Riding on
Gnutella. First Monday, 5(10), Retrieved March 8,
2006, from http://www.firstmonday.dk/issues/
issue5_10/adar/.
Androutsellis-Theotokis, S. & Spinellis, D. (2004). A
Survey of Peer-to-Peer Content Distribution
Technologies. ACM Computing Surveys, 36(4), 335-
371.
Bhagwan, R., Savage, S. & Voelker, G.M. (2003).
Understanding Availability. Paper presented at the 2
nd
International Workshop on Peer-to-Peer Systems,
Berkeley, USA.
CacheLogic (2004). The True Picture of Peer-to-Peer
Filesharing. CacheLogic, Retrieved November 23,
2006, from http://www.cachelogic.com/home/
pages/research/p2p2004.php.
Chu, J., Labonte, K. & Levine, B.N. (2002). Availability
and Locality Measurements of Peer-to-Peer File
Systems. Paper presented at the Conference on
Scalability and Traffic Control in IP Networks,
Boston, USA.
De Boever, J. (2007). Peer-to-Peer Networks as a
Distribution and Publishing Model. Proceedings of the
11
th
International Conference on Electronic
Publishing, Vienna, Austria, 175-188.
Del Conte, N.T. (28.06.2006). BitTorrent pays for direct
delivery. PC Magazine, Retrieved September 26,
2007, from http://www.pcmag.com/article2/
0,1895,1983053,00.asp.
Ding, C.H., Nutanong, S. & Buyya, R. (2005). Peer-to-
Peer Networks for Content Sharing. In R.
Subramanian & B.D. Goodman (Eds.). Peer-to-Peer
Computing: the Evolution of a Disruptive Technology
(pp. 28-65). London: Idea Group Publishing.
Divac-Krnic, L. & Ackerman, R. (2005). Security-Related
Issues in Peer-to-Peer Networks. In R. Steinmetz & K.
WEBIST 2008 - International Conference on Web Information Systems and Technologies
376
Wehrle (Eds.). Peer-to-Peer Systems and Applications
(pp. 529-545). Berlin Heidelberg: Springer-Verlag.
Dogan, S.L. (2005). Peer-to-Peer Technology and the
Copyright Crossroads. In R. Subramanian & B.D.
Goodman (Eds.). Peer-to-Peer Computing: the
Evolution of a Disruptive Technology (pp. 166-193).
London: Idea Group Publishing.
Good, N.S. & Krekelberg, A. (2005). Usability and
privacy: a study of KaZaA P2P file-sharing. In L.F.
Cranor & S. Garfinkel (Eds.). Security and Usability:
Designing Secure Systems That People Can Use (pp
651-668). Cambridge: O’Reilly.
Haßlinger, G. (2005). ISP Platforms Under a Heavy Peer-
to-Peer Workload. In R. Steinmetz & K. Wehrle
(Eds.). Peer-to-Peer Systems and Applications (pp.
369-381). Berlin Heidelberg: Springer-Verlag.
Hughes, D., Coulson, G. & Walkerdine, J. (2005). Free
Riding on Gnutella. IEEE Distributed Systems Online,
6(6).
Hughes, J., Lang, K.R. & Vragov, R. (2005). Electronic
Market Design Principles in the Context of Peer-to-
Peer Filesharing Systems. Paper presented at the
Pacific Asia Conference on Information Systems,
Bangkok, Thailand.
Hummel, T., Muhle, S. & Schoder, D. (2005). Business
Applications and Revenue Models. In R. Steinmetz &
K. Wehrle (Eds.). Peer-to-Peer Systems and
Applications (pp. 473-489). Berlin Heidelberg:
Springer-Verlag.
Hummel, T., Strømme, Ø. & La Salle, R.M. (2003).
Earning a Living among Peers – the Quest for Viable
P2P Revenue Models. Paper presented at the 36
th
Hawaii International Conference on System Sciences,
Island of Hawaii (Big Island), Hawaii.
IDC (2007). The Expanding Digital Universe: A Forecast
of Worldwide Information Growth Through 2010.
EMC. Retrieved April 4, 2007, from
http://www.emc.com/about/destination/digital_univers
e/pdf/Expanding_Digital_Universe_Executive_Summ
ary_022507.pdf.
Kwok, S.H., Lang, R.L. & Tam, K.Y. (2002). Peer-to-Peer
Technology Business and Service Models: Risks and
Opportunities. Electronic Markets, 12(3), 175-183.
Lechner, U. & Hummel, J. (2002). Business Models and
System Architectures of Virtual Communities: From a
Sociological Phenomenon to Peer-to-Peer
Architectures. International Journal of Electronic
Commerce, 6(3), 41-53.
MacInnes, I. & Hwang, J. (2003). Business Models for
Peer to Peer Initiatives. Paper presented at the 16
th
Bled Electronic Commerce Conference, Bled,
Slovenia.
Morris, I. (2007). Crave Talk: Five things that are wrong
with BBC iPlayer. CNET. Retrieved July 31, 2007,
from http://crave.cnet.co.uk/software/
0,39029471,49291924,00.htm.
Norton, W.B. (2007). Video Internet: The Next Wave of
Massive Disruption to the U.S. Peering Ecosystem.
Paper presented at the Stanford Networking Seminar,
Stanford, USA.
Pallis, G. & Vakali, A. (2006). Insight and Perspectives
for Content Delivery Networks. Communications of
the ACM, 49(1), 101-106.
Parameswaran, M., Susarla, A. & Whinston, A.B. (2001).
P2P Networking: An Information-Sharing Alternative.
IEEE Distributed Systems Online, 34(7), 1-8.
Ripeanu, M., Iamnitchi, A. & Foster, I. (2002). Mapping
the Gnutella Network. IEEE Internet Computing, 6(1),
50-57.
Rodriguez, P., Tan, S.-M. & Gkantsidis, C. (2006). On the
Feasability of Commercial, Legal P2P Content
Distribution. ACM SIGCOMM Computer
Communication Review, 36(1), 75-78.
Rogers, E.M. (2003). Diffusion of Innovations. New York:
Free Press.
Rupp, P. & Estier, T. (2003). A Model for a Better
Understanding of the Digital Distribution of Music in
a Peer-to-Peer Environment. Paper presented at the
36
th
Hawaii International Conference on System
Sciences, Island of Hawaii, (Big Island), Hawaii.
Sabbaghi, A. & Vaidyanathan, G. (2004). SWOT Analysis
and Theory of Constraint in Information Technology
Projects. Information Systems Education Journal,
2(23), 3-19.
Shirky, C. (2001). Listening to Napster. In A. Oram (Ed.),
Peer-to-Peer: Harnessing the Power of Disruptive
Technologies (pp. 21-37). United States of America:
O’Reilly & Associates, Inc.
Sigurdsson, H. M., Halldorsson, U.R. & Hasslinger, G.
(2005). Potentials and Challenges of Peer-to-Peer
Based Content Distribution. Paper presented at the 2
nd
International CICT Conference: Next Generation
Broadband - Content and User Perspectives, Lyngby,
Denmark.
Smith, H.A., Clippinger, J. & Konsynski, B. (2003).
Riding the Wave: Discovering the Value of P2P
Technologies. Communications of the Association for
Information Systems, 11(4), 94-107.
Taylor, I.J. (2004). From P2P to Web Services and Grids:
Peers in a Client/Server World. London: Springer.
Vermeylen, S. (2004). Werken met SWOT-analyse.
Brussel: Uitgeverij Politeia.
Warner Music Group (09.04.2007). Warner Music Group
and Joost announce agreement to offer WMG video
content on new internet television platform. WMG.
[26.09.2007, WMG: http://www.wmg.com/news/
article/?id=8a0af81211b8cea50111d683c13e0f66].
Werkers, E. & Gilio, F. (2006). De complexe verhouding
tussen peer-to-peer netwerken en de exceptie van de
privékopie: kan de driestappentest een evenwicht tot
stand brengen? Computerrecht, 166, pp. 283-292.
PEER-TO-PEER CONTENT DISTRIBUTION - An Analysis of the Internal and External Potentials and Obstacles
377