期刊名称：International Journal of Electrical and Computer Engineering
电子版ISSN：2088-8708
出版年度：2015
卷号：5
期号：4
页码：821-831
DOI：10.11591/ijece.v5i4.pp821-831
语种：English
出版社：Institute of Advanced Engineering and Science (IAES)
摘要：The main contribution of this paper is to present the Platform-as-a-Service(PaaS) Environment for Social Multimedia Service (PESMS), derived fromthe Social Media Cloud Computing Service Environment. The main role ofour PESMS is to support the development of social networking services thatinclude audio, image, and video formats. In this paper, we focus in particular on the design and implementation of PESMS, including the transcoding function for processing large amounts of social media in a parallel and distributed manner. PESMS is designed to improve the quality and speed of multimedia conversions by incorporating a multimedia conversion module based on Hadoop, consisting of Hadoop Distributed File System for storing large quantities of social data and MapReduce for distributed parallel processing of these data. In this way, our PESMS has the prospect of exponentially reducing the encoding time for transcoding large numbers of image files into specific formats. To test system performance for the transcoding function, we measured the image transcoding time under a variety of experimental conditions. Based on experiments performed on a 28-node cluster, we found that our system delivered excellent performance in the image transcoding function.
其他摘要：The main contribution of this paper is to present the Platform-as-a-Service (PaaS) Environment for Social Multimedia Service (PESMS), derived from the Social Media Cloud Computing Service Environment. The main role of our PESMS is to support the development of social networking services that include audio, image, and video formats. In this paper, we focus in particular on the design and implementation of PESMS, including the transcoding function for processing large amounts of social media in a parallel and distributed manner. PESMS is designed to improve the quality and speed of multimedia conversions by incorporating a multimedia conversion module based on Hadoop, consisting of Hadoop Distributed File System for storing large quantities of social data and MapReduce for distributed parallel processing of these data. In this way, our PESMS has the prospect of exponentially reducing the encoding time for transcoding large numbers of image files into specific formats. To test system performance for the transcoding function, we measured the image transcoding time under a variety of experimental conditions. Based on experiments performed on a 28-node cluster, we found that our system delivered excellent performance in the image transcoding function.
