Video Conversion In Different Format Using MapReduce On Hadoop

Abstract There is great success in analyzing and processing structured data in Hadoop where as analyzing unstructured data like digital video is always difficult. Videos are available in different format. For the portability usage and to access video on different digital devices, videos are necessary to convert in specific format acceptable by digital devices. However, the traditional approach of video format conversion takes lots of time. Therefore, we propose a Hadoop MapReduce based video conversion module in order to reduce the burden of computing power. It consists of two parts: a storage system, i.e., Hadoop distributed file system (HDFS) for video data and a MapReduce program for video transcoding. It can process video data in distributed manner and parallel conversion, thereby minimizing the computing infrastructure overhead. The existing conversion tools such as format factory, total video convertor, any to any convertor provide the quality in conversion but they consumes lots of time. Single node consumes more amount of time for conversion of huge amount of video file (eg. GBs). Instead of that a big size video splitted into multiple blocks and parallel execution is performed and finally merging of all blocks takes place which result into less amount of conversion time.