In social system, people with similar interests gather
and create a community. This structure organizes people effectively and
makes sharing information easier. In sociology the behavior of such structures
has been investigated for a long time. Fortunately this structure can
be extended to Peer-to-Peer (P2P) systems. This is due to the fact that
peers in P2P systems usually have few interests like people in the real
world and they try to find other peers with similar interests. On the
other hand, the structure of the underlying models in P2P has a direct
effect on different aspect of such systems. In this study the performance
related parameters of a P2P system with social network characteristics
are measured by simulation. The result shows that using similar structure
as same as real world inside a community produces better performance.
In addition, flooding technique in such systems creates higher traffic
than random structured model; however a simple controlled flooding can
provide a satisfaction result.