| With the deepening of cloud storage, cloud computing, big data and other a new generation of information technologies in the field of education. Wisdom Education was given a new connotation and characteristics, it’s a new cloud service model. With the deepening of the construction of campus informatization, some serious problems exposed gradually, such as education info island, lacking of data exchange standards in the departments of education system, data sharing is inconsistent, data utilization is not high etc, leading to high cost and low quality of teaching. As the same time, the existing collective preparation system are mostly single computer-assited instruction (CAI) courseware making, lacking of collaboration among teachers, teachers could not complete the collective lesson planning tasks as scheduled with quality. In the education system, a large number of users will read these education data resources with massed, diversified, decentralized, fragmented and other characteristics frequently. Due to a large number of small files will affect scalability and performance of Hadoop (the distributed system infrastructure that developed by Apache Foundation). So far, there is no suitable and universial solution for any scene to solve the problem of HDFS small files storage. Based on the above formulate questions, we research and develope the Collective Lessoon Preparation System based on Hadoop Distributed File System (HDFS) Cloud Stroage which can integrate and use the related data resources generated by a lot of teaching and learning activities effectively and scientificly. No matter seeing from the user information needs or from the future wisdom education development direction indicated by the national policy, it has important theoretical and practical significance, the research contents include as following:1) Through the study of computer-assited instruction(CAI) and the e-xisting systems of collective lesson preparation, taking advantages of design patterns and object-oriented technologies and use Java, Jsp, ActionScript and other languages. Eclipse and Flash Buider work as the development of tools, the mainstream framework Structs+IBatis work as the core. We develope the Collective Lesson Preparation System and use the scientific preparation process to standardize preparation, set up a teaching and research evaluation system, break the barriers of time and space between teachers. So, teachers can conduct collaborative lesson preparations. Meanwhile, we focus on the protection of great thoughts of teachers.2) Through the study of HDFS and the comparison of several com-mon small HDFS file storage solutions, we propose a small file storage solution applied to the Collective Lesson Preparation System. The solution considers the actual enviroment and needs of the system, the particular scenario with file-related, access locality and other characterisitics. It uses the storage and access efficiency as evaluation criterion and mainly studies from the three aspects as:file merge, file mapping, file prefetching in detail.3) Based on the Hadoop cloud storage architecture, we bulid a clou-d storage system based on computing resources of public facilities for large numbers of preparation document data resources, collaborative handwritten image data resources, collaborative audio and video resources. Then, we use these massive education data for efficient storage management and intelligent data mining.The cloud storage system runs on the HDFS distributed file system implemented by Hadoop cluster.The achievements of this paper all have passed rigorous tests, the test results show that the system could be reliable and stable. |
