Improving Search Performance Using Selective Query in Peer to Peer Networks
| Autor: | Chih Feng Hsieh, 謝志峰 |
|---|---|
| Rok vydání: | 2004 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 93 Peer-to-Peer (P2P) Networking is a new infrastructure used for distributed network applications. The most popular P2P file sharing system is Gnutella that uses flooding to query all peers. Although flooding is a simple approach, it is inefficient due to its increase of message overhead and latency time. In this paper, we propose “selective query” to improve search performance in Gnutella. In our approach, query messages are restrictedly forwarded to selected peers rather than to be flooded to all peers. We incorporate interest-based locality to limit querying scope. Search latency was reduced by establishing a shortcut-based tree based on the reputation scores of peers. In any event, if a peer has a higher reputation, it is allocated to the upper level of the tree. Thus the peer is queried earlier in order. That improves the efficiency of query and decreases the amount of query messages. We evaluate our simulation by five diverse content distribution applications. Simulation results show that shortcut-based trees increase the query success rate, and the proposed method takes less time to obtain the query results. |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|