P2P Applications - Peer-to-Peer Computing: Applications, Architecture, Protocols, and Challenges - page 5

Chapter 2
P2P Applications
2.1 Introduction
As in many computing technology breakthroughs, the advancements in
peer-to-peer (P2P) systems are largely brought about by applications' de-
mands. Indeed, evolving from the early file sharing systems to nowadays'
video streaming systems, many novel e cient solutions have been proposed
and implemented to satisfy various users' requirements. Thus, before we look
at these advancements in detail in later chapters, it is useful for us to review
the evolution of P2P applications in this chapter.
As shown in Figure 2.1, there are three main components in a typical P2P
system. The first one is a Web portal of the application, also known as a login
server, which is the point of getting access to the P2P service. Essentially,
the user first connects to this server in order to check out the availability of
services as well as peers. The second main component is usually referred to as
the tracker nowadays, which is essentially a directory server furnishing peers
availability information to a new peer. In the early days of P2P computing,
these two servers were typically implemented in a single system. Yet as user
population grows and the service becomes much more diversified (e.g., there
are numerous video channels available), the peers tracking system function has
to be overloaded to a separate server, namely the tracker. The third component
is of course the peers, which are autonomous client machines that join and
leave the system at will. The peers serve many useful or even critical functions.
For instance, one of the most important functions is that peers help each other
to get the necessary data packets.
Specifically, there are three key aspects governing the behaviors of a P2P
application.
Discovery. Upon entering the P2P system, the very first tasks a new peer
needs to carry out are the discoveries of services, data, and peers. The
peer first has to find out if a particular desired service (e.g., a certain
video channel or a specific file) is available. Then the peer needs to
determine the various pieces of meta-data about the service, such as the
location information and the size of the actual data. Finally, before the
actual data can be downloaded, the peer needs to obtain a list of peers,
from the corresponding tracker, for making data transfer connections.
5