Real-time distributed video transcoding on a heterogeneous computing platform

The requirements for real-time video transcoding systems have been significantly increased due to the easy and widely available access to high resolution video streams and large-scale applications in recent years. A real-time distributed video transcoding system working in a heterogeneous environment has been proposed to tackle the high requirements of such applications. The proposed approach emphasizes throughput of video data aggregation which involves the continuous input video stream and outcomes of transcoded video streams that are accessible on-the-fly in contrast to the batch-oriented approach such as the MapReduce framework, where output latencies can be significant.