Fault-Tolerant Rate-Monotonic Compact-Factor-Driven Scheduling in Hard-Real-Time Systems

摘要

Considering the disadvantage of first-fit strategy in fault-tolerant rate-monotonic first-fit (FTRMFF) algorithm, we analyze the slack time of processors and the schedulability of periodic tasks in rate-monotonic (RM) algorithm. Then, the RM-based idleness factor and compact factor are presented to quantify the compact degree of tasks assigned to the same processor. In this paper, the novel fault-tolerant rate-monotonic compact-factor-driven (FTRMCFD) algorithm, which follows the principle of compact factor maximal when allocating the processors for tasks, is proposed. FTRMCFD algorithm makes every processor contain more tasks and get higher utilization to increase the schedulability performance of distributed systems. The simulation experiments reveal that FTRMCFD can reduce the number of required processors by up to 11.5% (with an average of 5.3%).