Multimedia processing environment is a hybrid environment that manipulates stream processing tasks and event-driven tasks. Each stream tasks is a soft real-time task, which manipulates continuous media such as video and audio, has timing constraints according to characteristics of media. On the other hand, event-driven task is a non real-time task which manipulates discrete media with no timing constraint, and requires better response time. We have been an adaptive scheduling policy called ADM (Adaptive Deadline Modification) which dynamically adaptable to various processing environments. In this report, we apply ADM to scheduling of stream processing tasks and event-driven tasks in the hybrid environment. Namely we propose scheduling policy which meets timing constraints for stream processing tasks and provide better response time for even-driven tasks.
展开▼