NEIGHBOURING NODE FEATURE COMPUTATION SCHEME TO IMPROVE PEER-TO-PEER NETWORK EFFICIENCY

Naveen Vijay, Noor Saba, Silpi Dutta, Suprith S, Suma R

Abstract: This project focuses on improving the QoS provided by the searching mechanism in unstructured Peer-to-Peer network. In an unstructured Peer-to-Peer network, any node in the network act as both the server and a client, i.e., there is no centralised server to ping if any node requires a file which is available somewhere in the network. Consequently, a node will send a file query to its neighbouring node(s), and the neighbouring node(s) will forward the query to their neighbour(s), and the process repeats till the query reaches the node containing the required file, which is sent back to the querying node. Our projects aims at improving this system by introducing a probability based statistical selection algorithm, in contrast to the existing blind search methods. In this method of searching, a querying node will send the file query to only one selected neighbour, and a similar forwarding process takes place. The selection process is performed by calculating certain features of each neighbour and selecting the node which more likely has the required file, using standard deviation and probability. This helps improving transmission efficiency and reducing network congestion, with a reasonable increase in the processing overhead.

Keywords: Peer-To-Peer Network, Neighbouring Node Feature Computation Scheme, Search Mechanism

DOI: https://doi.org/10.15623/ijret.2016.0516030