Title :
Compositional Real-Time Scheduling of Mixed-Criticality Real-Time Systems
Author :
Ali, Ahmad ; Kyong Hoon Kim
Author_Institution :
Dept. of Inf., Gyeongsang Nat. Univ., Jinju, South Korea
Abstract :
Compositional real-time scheduling frameworks provide ways for building large and complex real-time embedded systems from independent sub-systems. This paper introduces a novel scheme for scheduling a component of Mixed-Criticality (MC) sporadic tasks on uniprocessor system. It characterizes the mixed criticality resource allocations, and the interface model for abstracting the real-time requirements of such a component. The demand bound function and supply bound function for the low and high criticality modes are derived to determine the schedulability of a component in both modes. The execution times of jobs are monitored in this approach. If it is detected that the behavior of system is inconsistent with low criticality level, then tasks of low criticalities are discarded and the priorities of high criticality tasks are re-ordered in high criticality mode to guarantee the schedulabilty of high criticality tasks. We propose a new algorithm, which schedules a component of mixed criticality jobs in low mode through MC Rate Monotonic (MC-RM) algorithm by applying period transformation, and guarantees the schedulabilty by using MC Earliest Deadline First (MC-EDF) algorithm in high mode.
Keywords :
embedded systems; processor scheduling; resource allocation; safety-critical software; MC earliest deadline first algorithm; MC rate monotonic algorithm; MC-EDF algorithm; MC-RM algorithm; component schedulability; compositional real-time scheduling; mixed-criticality real-time systems; period transformation; real-time embedded systems; resource allocations; uniprocessor system; Optimal scheduling; Real-time systems; Resource management; Scheduling; Scheduling algorithms; Switches;
Conference_Titel :
Dependable, Autonomic and Secure Computing (DASC), 2014 IEEE 12th International Conference on
Conference_Location :
Dalian
Print_ISBN :
978-1-4799-5078-2
DOI :
10.1109/DASC.2014.89