Optimized Deadline Assignment and Schedulability Analysis for Distributed Real-Time Systems with Local EDF Scheduling

The assignment of scheduling parameters under the Earliest Deadline First (EDF) scheduling policy is trivial in single processor systems because deadlines are used directly. However, it is still difficult to find a feasible deadline assignment for EDF distributed systems when the utilization levels of the CPUs and communication networks are pushed near to their limits. Most distributed applications specify end-to-end deadlines for each transaction and there are no individual deadlines assigned to their tasks or messages. This paper presents a new heuristic algorithm, called HOSDA (Heuristic Optimized Scheduling Deadline Assignment), for optimizing the assignment of deadlines to tasks and messages in distributed hard real-time systems. The algorithm is based on HOPA (Heuristic Optimized Priority Assignment), a previous method for the assignment of priorities in fixed priority distributed systems. The results of the proposed algorithm are compared with two other algorithms that exist for solving the same problem, and show that a utilization increase of up to 18% is possible. The paper also proposes a new schedulability analysis technique for EDF distributed systems with local scheduling deadlines.