Font Size: a A A

A Gnutella Search Method Based On GPath-Tree

Posted on:2006-07-07Degree:MasterType:Thesis
Country:ChinaCandidate:L ZhuFull Text:PDF
GTID:2168360152982867Subject:Computer applications
Abstract/Summary:PDF Full Text Request
In today's architecture of networks, P2P is a potential and promising technique that shares resources and services of computer by exchanging data directly. It has been widely applied in Peer-to-Peer computing, coordinated operation, search engine and file exchange with an unexampled superiority to traditional Client/Server model. Gnutella is a representative of decentralized P2P network. Because of its character of entirely distributed and highly dynamic, it can effectively avoid isolation nodes bottleneck, intelligently finding peers, acquire valid information and make network much more robust. While flooding is used to broadcast query across Gnutella networks, it generates exponentially increasing redundant messages, so that searching efficiency is decreased rapidly and the growth and scalability of P2P system is hindered. In these years, P2P is a worldwide hotspot in research field. Some relative works are reported in China, focusing on structured P2P network, such as CAN, Tapestry, Chord, Pastry, but the research on unstructured P2P-Gnutella is yet deficient. This paper takes efficiency improvement of flooding as topic, then makes an extensibility discussion on it, and mainly analyzes the characteristic and deficiency of this mechanism. Attempting to realize the efficiency enhancement, "he cost and redundancy reduction, a new technique called GPath-Tree is presented, and meanwhile, GPSA is given on the evaluation of Gpath. The experiment shows that GPath-Tree is available, and it could increase flooding efficiency, reduce redundant messages, eliminate shot-circuiting effect, and adjust itself to the dynamic change of P2P system causing by constant peer arrivals and departures.
Keywords/Search Tags:Gnutella, flooding, GPath-Tree, GPath, GPSA
PDF Full Text Request
Related items