Using Hadoop Distributed Computing Architecture to Implement a System of Journal Submission Retrieval and Cloud Management

• Main Authors: Chung Wei Pu, 浦仲緯
• Other Authors: Jyi Ta Chen
• Format: Others
• Language: zh-TW
• Published: 103
• Online Access: http://ndltd.ncl.edu.tw/handle/41183849256607685691

碩士 === 南台科技大學 === 資訊管理系 === 102 === As time progresses, the huge amount of data increase rapidly, cloud computing becomes a hot issue which was discussed by everyone. Nowadays, more and more applications are used on the cloud system. On the other hand, the general application which was developed before are gradually replaced by the powerful Cloud Computing environment. Scholars and people are eager to develop a Cloud Computing environment to take the preemptive opportunities of business. Now the development platform largely applied in several areas is Hadoop platform. Hadoop is an opened-source software with low hardware limit to build a cloud system with distributed computing ability. With many software supporting mutually to enforce the operating efficiency, Hadoop occupies an important role in the cloud computing market. The issue we will discuss in this research is how to increase the operation efficiency of the system by planning and assembling the Hadoop Cloud Computing environment. However, not all the application used on the cloud system can improve the system performance. For example, the application need few data processing and without immediate updating support, would not function efficiently; it will even cause the waste of resource if it transfer to cloud system from the client/server system. Therefore, we suppose a journal submission system is the best application to develop under cloud computing environment. This system needs to process huge amount of journal data by sentences analysis and keyword matching for journal retrieval. Besides, the daily increasing data about journal need to be updated and change all the time. The above problems cause the system to be overloaded consequently, and also stall the processing time, even lower the working efficiency. To prevent above problems, we propose a system with the Hadoop Cloud Computing Cluster to provide a high efficient journal submitting system by utilizing a distributed structure environment to reach a perfect loading balance. Moreover, considering the convenient control for system administrator, we build a simple and easily used managing platform to optimize the system maintenance. This system combines all the advantages of cloud systems. It’s available for crossing platform which can managing data from different application of cloud systems in one screen, also it provides good suggestions and more opportunities when someone wants to submit a paper to some journals.