Design and Implementation of a Cloud-based Distributed Multimedia Streaming Service (CloudDMSS) System
With the recent appearance of various heterogeneous smart devices and the expansion of social network services, services that support the production and sharing of social media data are actively provided. The stable provision of such services requires a transcoding technology supporting the N-Screen service and distributed streaming technologies for providing a quality of service (QoS)-based rich media streaming service. However, the transcoding and streaming services through the existing distributed computing environment places a substantial burden on the Internet infrastructure and computing resources by requesting vast computing resources. In addition, in the current situation where system expandability and stability are required, there are intrinsic difficulties such as the requirement of a variety of distributed management and operation policies and techniques including a data recovery policy in the event of content loss, an automatic recovery policy, an automatic content storage technique, and a cluster relocation technique. Therefore, in order to overcome such difficulties, in this paper, we have proposed a cloud-based distributed multimedia streaming service (CloudDMSS). Using the MapReduce framework, a core element technology of cloud computing, the proposed service enables high-speed transcoding and provides a stable streaming service by building a streaming service that is applied with a distributed streaming task distribution technique. In addition, with the application of the Hadoop policies and the Hadoop distributed file system (HDFS), the proposed service secures system expandability and stability.
Keywords- Streaming Service, Distributed Transcoding, Cloud Computing, Hadoop