| As the Internet and mobile phone networks continue to evolve, more and more media platform, and the corresponding media platform to support more and more media formats. The format of the content on different platforms completely different requirements, including the encoding format / size, resolution, frame rate and other parameters change, so the encoding format of multimedia content has become essential for the conversion. The traditional centralized data transcoding system inadequate to meet the large volume of data, more than the number of transcoding needs, and with the growing number of video data, centralized data transcoding system is difficult to achieve the capacity increases linearly with the requirements of data . Based on the distributed platform Hadoop, designed and implemented a distributed transcoding system.This paper describes the development of cloud computing technology and the Hadoop platform focuses on the design of distributed file system HDFS and MapReduce programming ideological principles. Then introduced a video compression-related knowledge. Distributed in the understanding of the knowledge platform and video compression, based on a distributed transcoding system implementation framework. Transcoding framework gives each of the major modules, including video processing module, the file transfer module, job queue management module, database module, MapReduce parallel tasks module, principles and implementation.Finally, distributed transcoding system for some initial testing and analysis, the test showed that the Hadoop cluster for video transcoding is feasible and can reduce the video processing time overhead. The successful implementation of the Hadoop cluster video transcoding application to expand the scope of the Hadoop system, while, Hadoop cluster is built on a normal PC machine, the computing power and storage capacity can be increased as demand on the basis of freedom in the original expansion has high practical significance and application value. |
PDF Full Text Request