现有的实时容错调度算法没有将实时任务的动态安全需求与其可调度性结合起来考虑.针对这一问题,文章展开以下研究工作:基于安全分级思想,构建了一个适应动态安全需求的实时调度模型,该模型详细地描述了实时任务、系统安全服务、任务容错等实时调度过程涉及的关键要素.以此模型为基础,提出了一种自适应实时容错调度算法(AFTS),该算法支持优先级抢占式调度策略,以牺牲普通任务的运行为代价来保证关键任务的可调度性,并采用主副本备份技术实现了关键任务的容错功能.当系统安全级别被动态调整时,该算法能够为实时任务选择满足当前安全需求的最优安全策略.仿真实验表明,文中提出的算法与同类算法相比,在系统动态安全需求的适应性,以及关键任务的可调度性和容错能力等方面有较好的表现.%Current fault-tolerant scheduling algorithms have considered either the dynamic security requirement or the schedulability one of real-time task, but, to our best knowledge, not both together. Sections 1 and 2 of the full paper explain our AFTS algorithm, which we believe is better than existing ones and whose core consists of; "Firstly , it builds a real-time scheduling model, which is adaptive to dynamic security requirement, and the model describes in detail the scheduling process, including real-time tasks, system security services, and task fault-toleran-tance. Based on the model, it proposes a new AFTS algorithm which supports priority preemptive scheduling policy , and guarantees the schedulability of critical tasks at the cost of deferring normal tasks and adopts primary/backup copy technique to ensure fault-tolerance of critical tasks. The algorithm selects the best-fit security policy for schedulable task under updated system security level. ". Simulation results, presented in Fig. 1, and their analysis show preliminarily that the improvements in the adaptability to dynamic security level, the schedulability, and the fault-tolerance of critical task can indeed be achieved by using our AFTS algorithm.
展开▼
