The Design And Implementation Of A Distributed Video Transcoding System

With the network infrastructure and video technology development andimprovement of Internet streaming media applications in the gradual increase in theproportion of applications. The main video encoding formats are MPEG, H264, DivX,WMA, RM, etc., packaging formats are mainly avi, ps, ts, mov, mkv, mpg, etc., andthe player supports different formats for different, so different terminal videostreaming format also has special needs. Improve transcoding performance hasbecome an important topic of streaming media technology.Taking into account the existence of streaming media applications mostlycalculated pressure, time-consuming problem, we designed a distributed computingframework for streaming media to form a complete set of distributed transcodingsolutions, transcoding with high efficiency, high robustness characteristics.The system uses the Java language, based on the current most populardistributed computing framework Hadoop development, the design is divided intoseven modules: transcoding module, transmission module, memory module,exception handling module, load balancing module, video split merge module,background interface module. Transcoding module package ffmpeg and mencodertranscoding function to achieve all video formats conversion, when a large numberof video transcoding tasks submitted to the cluster, the video split module willfollow a certain strategy to split the video automatically and through thetransmission module assignment on different nodes in the cluster, and then use thecluster system and the existing computing power of open source software videotranscoding multiple concurrent video transcoding. Cluster transcoding taskallocation and exception handling each node separately from the load balancingmodule and exception handling module is responsible for processing, to ensureefficiency and stability of the system.After testing, it’s proved that the system has completed the desired real-timeand efficiently video transcoding tasks, and has better robustness, encapsulation. Theimplementation of video transcoding is transparent to external and it is highscalability and the system reduce the time required for video transcoding effectively.