Hadoop scalable Video Transcoding technique in cloud environment

Abstract

Nowadays video is being produced and consumed in more component representation formats, more device types and over a variety of networks than ever. Transmission of video through network takes more time. So Video Transcoding is a very important factor when the video is moved between various heterogeneous clients in the cloud environment. Transcoding is a process of translating one coded form of video into another. However most of the time transcoding becomes computationally intensive and time consuming process. This proposed cloud system has four of the video compression standards such as Low Quality Encoding , Standard Quality Encoding , High Quality Encoding and High Definition. These standards achieve better compression performance with Quality. In general compression process takes more time. Map Reduce is used for managing a work in a considerable short period of time. This in turn helps in faster and efficient video transcoding. These Four standards are embedded into the Hadoop Distributed File System implementation and trial runs were done. Using the HDFS Map Reduce functionality, the video is splited using 64 MB blocks (Segments of Streams) and processed separately for maintaining efficiency in a time based aspect. Index Terms-Cloud Computing, Video Encoding, Hadoop, Map Reduce, FFMPEG 978-1-4799-6480-2/15/$31.00

MapReduce-Based Distributed Video Encoding Using Content-Aware Video Segmentation and Scheduling

“…The study determined that the number of Mappers and the size and duration of video files are significant factors affecting the overall performance, and the relationships between these factors were described. Reference [22] proposed a video block split-up scheme that efficiently splits videos for Hadoop-based transcoding. This study calculated the histogram values of the forward and backward videos to split videos into 64 MB blocks, the default HDFS block size.…”