P2P Applications - Peer-to-Peer Computing: Applications, Architecture, Protocols, and Challenges - page 6

Login Server

Tracker

Internet

Peer

Peer

Peer

Peer

FIGURE 2.1: A general architecture of a P2P application.

Location. Here, location is used in two different senses. First, a new peer
needs to obtain location information about the corresponding tracker
(e.g., the tracker's IP address), about the peers that own the needed
actual data. Second, the new peer also needs to report to the P2P servers
about its own location and the data it already possesses. Such location
information exchange is crucial for the tracker servers to keep accurate
data about the availability of actual data and peers.

Data Transfer. The uploading and downloading of the actual desired data
are obviously the ultimate important steps. These steps are also the as-
pects where different P2P systems take on different approaches. For one
thing, there are the so-called push and pull approaches for data exchange.
In a push-based approach, it is the data uploading peer who determines
the recipients of the data. In contrast, in a pull-based approach, it is the
data downloading peer who sends out transfer requests to a set of po-
tential data senders. Another important dimension about data transfer
is the network topology issue. Again there are two general approaches.
The first one is a structured approach, in which the connections among
peers are governed by a well-defined network topology such as using a
distributed hash table (DHT). The second approach is a so-called mesh
approach in which connections among peers are totally ad hoc and do
not follow any structured topology.

More about these different key aspects of a P2P application are further
explained when we describe the specific applications below.