A Trustworthy and Churn-Resilient Journal Distribution and Retrieval System in P2P Networks.

• Main Authors: Li, Feng-Wei, 李豐緯
• Other Authors: Chuang, Yung-Ting
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/q4684g

碩士 === 國立中正大學 === 資訊管理系醫療資訊管理研究所 === 107 === Due to the internet becoming worldwide recently, people prefer to get information through the internet. The peer-to-peer network offers a platform that let lots of users share their files to each other. However, there is still no such study that can completely solve these trustworthy and efficient issues. Some existing systems try to combine the search mechanism with trust management mechanism, but however, these systems still require too much overhead. In this paper, we propose a trustworthy and churn-resilient journal distribution and retrieval system in P2P networks, named TCR. This system can be used in the following situations: 1. Maintain a high search successful rate for the presence of network churn ; 2. Categorize users according to research topics, to improve search efficiency with low maintaining cost; 3. Use trust management mechanism TCRTrust to make sure the target files can still be found even in the presence of malicious attacks; 4. The algorithm of this study will balance the advantages and disadvantages of structured and unstructured peer-to-peer networks with hybrid peer-to-peer network architecture. The most common problem with peer-to-peer networks is query efficiency. This study will propose a solution to the problem. First, we compare many peer-to-peer network models and develop a trustworthy and academic retrieval system that can be applied to the peer-to-peer networks with resistance to churn. Then, for the malicious nodes in the network, we use the trust management mechanism, TCRTrust provided in this study to let each node evaluate others in the network, so that each retrieval can be implemented successfully. This study compares the proposed algorithm with other algorithms, under the same system environment. We simulate the real-world network environment by crawling the information on the academic community website, and then compare their retrieval rate and retrieval cost. The experimental results confirm that our architecture can perform better when comparing with other existing systems.