04-04-2012, 12:08 PM
A Study On Peer-to-Peer Systems
[attachment=19429]
Introduction
It was Napster in 1999 that introduced to the public a new approach of distributed
computing known as peer-to-peer networking. Immediately attracting millions of
users, such systems came in the forefront of a range of technologies that were going
to change the way we see the Internet. This new approach in networking, very
soon got the attention of the internet community. Not unexpectedly, its apparent
success fuelled academic research which gave birth to new generations of systems
but also new challenges that need to be faced. Nowadays, the term peer-to-peer
is for distributed computing what the term “Web 2.0” has become for internet
applications.
But besides being a buzz-word, peer-to-peer’s promise has provoked considerable
activity in the academic community and also the industry. Many new applications
have appeared and Internet Service Providers report large proportions of their overall
traffic being caused by peer-to-peer implementations for file sharing. The academic
world has responded to this newly found network architecture with a vast amount of
publications since the beginning of the millennium. The initial wave was “re-active”
in that it mainly examined the existing systems. Following was a flourish of new
ideas that either improved extensively current ideas or proposed completely new
approaches.
Background
It was only very recently, around 2000, that people started becoming aware of peerto-
peer technologies. It was companies -or rather phenomena- like Napster that
brought peer-to-peer in the public’s eye. Napster started providing its services in
September 1999 and by 2001 it was eventually shut down by the music industry. By
then its network consisted of 50 million users (Barkai, 2001). The characteristics
of this technology, its popularity, potential and its effect on the field of distributed
computing has lead many to call it a “disruptive technology” (Christensen, 1997).
Definition
According to the Oxford English Dictionary a peer is “a person of the same age,
status or ability as another specified person”. Based on this definition we can have
a -rather simplistic- view of what is meant by a the term “peer-to-peer network” :
A network of equals. A step forward from existing architectures where the client is
nothing more than a dump terminal at the end of client-server connection to a new
kind of network where all participants act both as clients and servers.
A more formal definition
There has been a lot of interest in the peer-to-peer field by the research community
resulting in a vast number of papers being published in the recent years. In this
wide range of publications one can find somewhat different definitions for the term.
Shirky (2000) defines peer-to-peer as “a class of applications that takes advantage
of resources –storage, cycles, content, human presence– available at the edges of
the internet”. However, this is probably too broad.
Applications
A wide range of applications can be modified or build from scratch to run on a
peer-to-peer overlay. Currently, the widest use of current peer-to-peer networks are
for content distribution, more specifically file sharing applications.
Chosen Approach
The most often used classification of peer-to-peer systems is the one that divides the
field in structured and unstructured approaches. Risson and Moors (2004) discuss
how this approach is becoming less relevant for two reasons. Firstly, they argue
that most unstructured proposals are evolving to incorporate some kind of structure,
therefore making the distinction between the two different kinds less clear.
Secondly, the new schema based peer-to-peer designs -mentioned above- introduce
a new category that cannot be classified under structured or unstructured designs
Conclusion
Peer-to-peer systems introduce a new approach to network computing. It is a move
from the classic client-server model to a network of peers where all participants
act as both clients and servers. It is essentially a move to cooperation instead of
coordination, to incentives instead of control (Steinmetz and Wehrle, 2005).
The promise of peer-to-peer comes as a solution to the ever-growing demands of
future network applications. These self-organising networks provide us with the
ability to share resources such as storage, content and CPU cycles. Whether the
challenges presented in this report will be successfully faced, it remains to be seen.
Nevertheless, by studying the various aspects of such systems and looking at their
evolution, one can only have an optimistic view of what is yet to come.
[attachment=19429]
Introduction
It was Napster in 1999 that introduced to the public a new approach of distributed
computing known as peer-to-peer networking. Immediately attracting millions of
users, such systems came in the forefront of a range of technologies that were going
to change the way we see the Internet. This new approach in networking, very
soon got the attention of the internet community. Not unexpectedly, its apparent
success fuelled academic research which gave birth to new generations of systems
but also new challenges that need to be faced. Nowadays, the term peer-to-peer
is for distributed computing what the term “Web 2.0” has become for internet
applications.
But besides being a buzz-word, peer-to-peer’s promise has provoked considerable
activity in the academic community and also the industry. Many new applications
have appeared and Internet Service Providers report large proportions of their overall
traffic being caused by peer-to-peer implementations for file sharing. The academic
world has responded to this newly found network architecture with a vast amount of
publications since the beginning of the millennium. The initial wave was “re-active”
in that it mainly examined the existing systems. Following was a flourish of new
ideas that either improved extensively current ideas or proposed completely new
approaches.
Background
It was only very recently, around 2000, that people started becoming aware of peerto-
peer technologies. It was companies -or rather phenomena- like Napster that
brought peer-to-peer in the public’s eye. Napster started providing its services in
September 1999 and by 2001 it was eventually shut down by the music industry. By
then its network consisted of 50 million users (Barkai, 2001). The characteristics
of this technology, its popularity, potential and its effect on the field of distributed
computing has lead many to call it a “disruptive technology” (Christensen, 1997).
Definition
According to the Oxford English Dictionary a peer is “a person of the same age,
status or ability as another specified person”. Based on this definition we can have
a -rather simplistic- view of what is meant by a the term “peer-to-peer network” :
A network of equals. A step forward from existing architectures where the client is
nothing more than a dump terminal at the end of client-server connection to a new
kind of network where all participants act both as clients and servers.
A more formal definition
There has been a lot of interest in the peer-to-peer field by the research community
resulting in a vast number of papers being published in the recent years. In this
wide range of publications one can find somewhat different definitions for the term.
Shirky (2000) defines peer-to-peer as “a class of applications that takes advantage
of resources –storage, cycles, content, human presence– available at the edges of
the internet”. However, this is probably too broad.
Applications
A wide range of applications can be modified or build from scratch to run on a
peer-to-peer overlay. Currently, the widest use of current peer-to-peer networks are
for content distribution, more specifically file sharing applications.
Chosen Approach
The most often used classification of peer-to-peer systems is the one that divides the
field in structured and unstructured approaches. Risson and Moors (2004) discuss
how this approach is becoming less relevant for two reasons. Firstly, they argue
that most unstructured proposals are evolving to incorporate some kind of structure,
therefore making the distinction between the two different kinds less clear.
Secondly, the new schema based peer-to-peer designs -mentioned above- introduce
a new category that cannot be classified under structured or unstructured designs
Conclusion
Peer-to-peer systems introduce a new approach to network computing. It is a move
from the classic client-server model to a network of peers where all participants
act as both clients and servers. It is essentially a move to cooperation instead of
coordination, to incentives instead of control (Steinmetz and Wehrle, 2005).
The promise of peer-to-peer comes as a solution to the ever-growing demands of
future network applications. These self-organising networks provide us with the
ability to share resources such as storage, content and CPU cycles. Whether the
challenges presented in this report will be successfully faced, it remains to be seen.
Nevertheless, by studying the various aspects of such systems and looking at their
evolution, one can only have an optimistic view of what is yet to come.