Interactive web application system development based on Hadoop cluster - A Case Study of Online Examination

• Main Authors: YANG, CHENG-XI, 楊承羲
• Other Authors: CHEN, PE-DE
• Format: Others
• Language: zh-TW
• Published: 2019
• Online Access: http://ndltd.ncl.edu.tw/handle/xk4m9z

碩士 === 中國科技大學 === 資訊工程系資訊科技應用碩士在職專班 === 107 === The traditional three-tier (3-Tier) architecture is the most commonly used architecture for Web applications: including presentation layer, application layer, and data layer. If this architecture is applied to the cloud computing platform, it has the following disadvantages: increasing time for system developers to learn big data analysis tools; application development suite and cloud computing platform projects are prone to conflict; replacing or upgrading components directly affects the performance of the application system. This study develops an interactive web application system combined with Hadoop cloud computing platform (hereinafter referred to as: Hadoop), and uses Docker container technology to build and manage Hadoop, compared to the physical machine or traditional virtualization technology, the study can reduce the time of the development, testing, and deployment, Based on the multi-tier architecture (N-Tier), a five-tier architecture with Hadoop as the core is proposed. From top to bottom, it is Load Balancing Tier, UI Tier, Application Tier, Data Access Tier and Data Lake Tier. As mentioned above, the study develop an interactive web application system based on Hadoop, taking online examination system as an example, and reduce the cost of Hadoop. This study uses the HDFS file browsing interface to simplify the user's complicated operation process in Hadoop and perform big data analysis tools; Pig, Hive, HBase, and Phoenix. Finally, the system architecture of this study, front-end, back-end, and Hadoop can be independently developed and deployed, and has with good scalability.