Resource Allocation Optimization through Task Based Scheduling Algorithms in Distributed Real Time Embedded Systems

Distributed embedded system is a type of distributed system, which consists of a large number of nodes, each node having lower computational power when compared to a node of a regular distributed system (like a cluster). A real time system is the one where every task has an associated dead line and the system works with a continuous stream of data supplied in real time.Such systems find wide applications in various fields such as automobile industry as fly-by-wire,brake-by-wire and steer-by-wire systems. Scheduling and efficient allocation of resources is extremely important in such systems because a distributed embedded real time system must deliver its output within a certain time frame, failing which the output becomes useless.In this paper, we have taken up processing unit number as a resource and have optimized the allocation of it to the various tasks.We use techniques such as model-based redundancy,heartbeat monitoring and check-pointing for fault detection and failure recovery.Our fault tolerance framework uses an existing list-based scheduling algorithm for task scheduling.This helps in diagnosis and shutting down of faulty actuators before the system becomes unsafe. The framework is designed and tested using a new simulation model consisting of virtual nodes working on a message passing system