Hard Aperiodic Scheduling in Fault Tolerant Cruise System – Comparative Evaluation with Dual Redundancy

Abstract

A scheduling algorithm for hard real-time systems with aperiodic task arrivals and fault tolerance can basically meet the needs of applications in the automotive or avionics domain. In these applications, weight, size and power requirements are crucial. Any resource augmentation technique to satisfy this and ensuring safe functionality under faults can bring in a paradigm shift in the design. This paper is based on a strategy for fault tolerance with task level criticality on dual processor system. An application with parallelizable task set has been used to advantage for resource augmentation under fault free condition. A processor failure (permanent fault) leads to safe recovery mechanism with graceful degradation. This paper deals with fault tolerant periodic task scheduling with arrivals of hard aperiodic events. An algorithm for aperiodic scheduling with admission control plan for hard and soft aperiodic tasks is developed and implemented on LPC2148 processors for the cruise system. A comparison is made with a traditional dual redundant system with appropriate performance metrics for evaluation.