Applying a peer-to-peer based mechanism to the performance and loading-balance optimization in multi-server application system.

• Main Authors: Kai-Ho Chang, 張凱賀
• Other Authors: Wei-Ming Chen
• Format: Others
• Language: zh-TW
• Published: 2010
• Online Access: http://ndltd.ncl.edu.tw/handle/83843001609538108950

碩士 === 國立宜蘭大學 === 資訊工程研究所碩士班 === 98 === As the Internet developed, bandwidth upgrade, and the diversity of Internet application services, network users and network applications are increasing, network applications and services has become a part of life. Web Application Service System to provide services to the public, naturally to be able to load a large number of users, multi-server systems and multi-layered structure of the program is common. However, System overload situation frequently. Therefore, this research proposes a hybrid Peer to Peer based mechanisms to improve system performance and load capacity. The system data flow is divided into query and transaction data flow. Generally, query data size is more than transaction data and less change. We can use Peer to Peer to share the query data between users. The system process the transaction data flow, its data size is small then the query data. Thereby reduce the system's loading. Furthermore, in hybrid Peer to Peer Group, deliver the transaction data through by the controller in the group, to reduce connections to the server. This way can also improve the performance of the system; and make relevant information security issues, to protect data in transmission security. This research design and program two simulation systems, a multi-server architecture and a Peer to Peer mechanism simulation system architecture for performance assessment. By analyzing the experimental result, the structure of this study is feasible, and the performance is better, reading or writing data on the system response time significantly faster than the multi-server architecture. The future of cloud technology is moving towards more web services and a lot more users, so the system load and performance requirements more so towards the future can study the cloud structure.