An Innovative Approach to Content Search Across P2P Inter-Networks
Files
Date
2009-02-25
Journal Title
Journal ISSN
Volume Title
Publisher
INFLIBNET Center
Abstract
Peer-to-Peer (P2P) computing has emerged as a popular model aiming at further utilizing Internet
information and resources. Flooding is the basic method of searching in unstructured P2P networks;
however, the blind flooding based search mechanism causes a large volume of unnecessary traffic,
and greatly limits the performance of P2P systems. Our study shows that a large amount of this
unwanted traffic is divinable and can be avoided while searching in P2P networks. In this paper, we
aim at reducing the volume of unnecessary traffic, by proposing An Unnecessary Message Prediction
based Searching mechanism (UMPS). UMPS is a pure distributed scheme, it is based on the distributed
neighbor list where neighbors within two hops are stored in. UMPS cuts down unwanted traffic by
terminating unnecessary flooding that has been predicted in advance, and the techniques related to
distributed neighbor list are also discussed. Simulation results show that more than 60% unnecessary
messages can be reduced by UMPS, and the query coverage range is retained at the same time, and
more unwanted messages can be reduced if a peer has larger degree, therefore load balancing is
achieved also.
Description
Keywords
UMPS, P2P Network