Hierarchical Resource Classification Overlay Network (HRCON) with Efficient Search in DHT-based Structure P2P Networks

• Main Authors: Chuan-Chi Weng, 翁傳奇
• Other Authors: Ching-Wen Chen
• Format: Others
• Language: zh-TW
• Published: 2006
• Online Access: http://ndltd.ncl.edu.tw/handle/02358941570327551044

碩士 === 朝陽科技大學 === 資訊工程系碩士班 === 94 === The most significant challenge in designing a P2P file sharing system is providing a keyword search mechanism that allows users to efficiently locate desired files with short duration. DHT-based systems are highly scalable for locating data items with unique item identifier; however, DHTs do not support keyword search. While unstructured P2P systems like Gnutella are flexible enough to support keyword search, however, this system achieve poor scalability due to query flooding. As described above, structured and unstructured P2P networks are lack with the ability to achieve both scalability and keyword search capability simultaneously. In this paper, we proposed a Hierarchical Resource Classification Overlay Networks (HRCON) on top of DHT where each overlay is responsible for particular category of files. In our scheme, when a peer need a file that type related to belonging group, desired files could be obtained from the peers inside the group, which are topologically close. Clearly, content caching inside groups can further increase the efficiency of peer to peer communications by reduce the number of messages that need to get out of the group. Since there are often a large number of possibilities in keyword search, HRCON can greatly increase the performance by narrowed down the searching area.