隨著資訊的進步以及網路越來越普及化之下,有效率的交換彼此的 資訊也因此變的格外的重要,傳統的client-server架構無法滿足用戶 間大量的資訊交流,因此P2P(Peer-to-Peer) network在近年來變成最 受歡迎的網際網路應用之一。如何提升P2P系統的搜尋效能變成一個很 重要的議題。在非結構式的P2P網路之下,Gnutella 是最為知名的一套 P2P軟體,但是他的主要搜尋方式是採用flooding的方式來進行尋找, 此方法會造成網路整體上多出太多不必要的流量浪費,在此論文中針對 這點進行在各peer的directory cache的應用,以及加入預測索引內容 機制,以達到更有效率的搜尋和減輕網路流量負擔的目的。此外本論文 也將負載平衡作為路徑搜尋時的考量因素之一,以避免出現某些peer負 載過重的情形。
As the improvement of information technology and Internet becoming more popular, it is very important that Internet users exchange information efficiently with each other. The traditional client-server model can not satisfy a vest amount of information exchange between Internet users. Accordingly, P2P (Peer-to-Peer) network become one of the most popular Internet applications in recent years. To improve the performance of searching information in P2P networks becomes a very important topic. In unstructured P2P networks, Gnutella is the most famous P2P system. Gnutella would waste a lot of bandwidth, since it uses flooding to locate and search resource. In this thesis, we use a directory cache (DC) and predict its index to improve search efficiency. In addition this thesis also addresses load balancing issue in searching paths to avoid some peers overloading.