Design And Implementation Of Transcoding Server Based On Cloud Computing Platform

In recent years, the Internet and new media technology has been rapid development, many applications based on new media in the software market are more and more abundant, especially the video service. In the current market, due to the terminal device type diversification, all media video format and network environment differences, which requires a solution to the real-time and fast processing, massive video transcoding, video technology. The cloud platform through its powerful parallel computing capability, high fault tolerance and good scalability characteristics, and cloud technology is becoming more and more mature, so you can explore the video transcoding tasks transferred to cloud platform for. In many of the cloud platform,currently is the most widely used Hadoop cloud computing platform, it is an open source framework also supports distributed computing, massive data processing and storage. Therefore, Hadoop open source framework, which solve the problem of real-time video transcoding for a variety of terminal equipment, a variety of video formats, multiple stream requirements become possible.This paper relies on the research department of "new media broadcast" project,proposed and implemented a computing platform based on Hadoop cloud transcoding scheme. In this paper, the main technology is using Hadoop cloud computing platform and ffmpeg open source transcoding tool distributed transcoding of video, the main contents are two core research and Analysis on the Hadoop platform: HDFS(Hadoop distributed file system) and Map Reduce parallel computing model, by means of the video file encapsulation principle, video format, and the video codec technology by based on GOP(group of pictures) continuous frame points technique, the original video file is divided into a plurality of video clips, combining with the Map Reduce programming design thought and the realization of the mapper and reducer task. At last, the paper for the system of distributed transcoding function were tested and.Experiments show that the implementation of the cloud transcoding system can improve the efficiency of video transcoding to a certain extent, and enhance the ability of the project platform. The scheme not only make full use of the cloud platform cluster parallel computing ability, solve the lab before project single transcoding performance bottlenecks, but also meet the requirement of the demand of the development of the laboratory of new media interactive broadcast services, has a strong practical significance.