On high efficient cloud video transcoding

Abstract

A multimedia server has to provide bandwidth compatible best QoS perception for end users. It has to perform transcoding to serve users under different network environments and with different devices. In this paper, we proposed a stability-driven hierarchical scheduling algorithm for a cloud-based video streaming system to speed up the video transcoding process for real-time service applications. It helps to coordinate the system operations and dynamically adjust the number of slots so that the cloud clusters can finish the process more efficiently. Experimental results showed that the proposed scheduling control methods help to maintain good system load balancing. The resource utilization of the entire cloud clusters can be as high as 98%, and the total transcoding time can be reduced by 10%-19%.