QoS concept for scalable MPEG-4 video object decoding on multimedia (NoC) chips

Scalable implementations of multimedia applications offer increased flexibility in mapping those applications onto the executing platform used in a consumer product. In this paper, we describe a hierarchical quality-of-service (QoS) model for managing multimedia applications running on a multiprocessor systems-on-chip (SoC). First, we present the possible scalability of an MPEG-4 arbitrary-shaped video decoder with respect to computational and communicational resources. Second, we provide a novel model for QoS management based on the principles of predictable mapping and run-time information on the resource utilization. We demonstrate the QoS framework by mapping of an MPEG-4 arbitrary-shaped decoder on a NoC, employing eight ARM cores with specific monitoring features in the network (e.g. Ethereal NoC). The scalable implementation results in lowering the computational requirements by 26% and communication by 43%. Experiments revealed that the combination results in more than 85% decoded frames of higher quality than in a QoS approach based on the predictable mapping only