A scheduling technique for real-time systems with end-to-end timing constraints

This paper presents a scheduling technique for guaranteeing end-to-end timing constraints. Applications are structured as a set of tasks which are the sequence of actions to produce an output. Task is decomposed into subtasks and shares subtasks with other tasks. According to the proposed scheduling technique, for each instance of a task, the sequence of subtasks is determined off-line and then called by the task sequentially at runtime. This technique avoids unnecessary execution of shared tasks and satisfies precedence constraints between subtasks even in overload situations.