The main contribution of this paper is to present the Platform-as-a-Service<br />(PaaS) Environment for Social Multimedia Service (PESMS), derived from<br />the Social Media Cloud Computing Service Environment. The main role of<br />our PESMS is to support the development of social networking services that<br />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

more »

... f 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.