Distributed media processing on multiple computing servers

Abstract

We propose distributed media computing method for high performance and massive computing of media processing. By dividing media streams temporally and spatially, and processing them, the method reduces the elapsed time of media processing and improves the system scalability. The experiment results show the method increases the elapsed time. The performance degradation could be mitigated using high-speed interconnection networking technology. Furthermore, we could utilize distributed and parallel processing approach by distributing media processing into remote servers.