부산대학교 도서관

Since P2P computing was firstly proposed to overcome the restrictions of client/server model, many researchers have proposed various P2P systems. Recent P2P systems are mostly based on DHT. But it is difficult to support range search efficiently in these systems because DHT makes difficult to search keys by ordered. In order to support range search this thesis propose to use the additional indices ordered by key. These indices are distributed uniformly into the peers in the system.Also this thesis proposes to use RDF ontology in order to fix the semantic gap between keyword and its results. RDF ontology is general method describing resources on the world wide web. The proposed system in this thesis has descriptions for the resources described by RDF ontology and uses them in order to fix the semantic gap.For the improvement of P2P system, it is very important to distribute indices evenly among peers. This thesis proposes to use histogram in order to distibute indices. The whole system was divided into some parts. S_node, which is superior to the other nodes in the part, checks a dense information and flatten the information in its histogram.Lastly, this thesis implements the prototype system using proposed methods and verifies that the proposed method supports range query well. Also this thesis shows that the proposed system properly solves the semantic gap problem by using RDF ontology.