基于动态异构模型的非周期性任务容错调度

Fault Tolerant Scheduling for Aperiodic Tasks Based on Dynamic Heterogeneous Model

拟态技术可有效解决实时系统中的安全问题,但其异构冗余的特性会增加系统时延。为此,以拟态工控处理机架构为基础,在动态异构的多模冗余场景下,结合具体表决策略并执行清洗切换任务,提出一种硬实时非周期性任务容错调度算法。仿真结果表明,相比静态异构模型下的DRFTS算法,该算法在保证硬实时的条件下,能够提高调度成功率。

Mimic defense techniques can effectively solve the security problems in real-time systems,but its heterogeneous redundancy will increase the system delay.To solve this problem,based on the architecture of the mimic processor,a hard time and aperiodic tasks fault tolerant scheduling algorithm is proposed in the dynamic heterogeneous multi-mode redundancy scenario,combining the specific voting strategy and implementing the cleaning handover tasks.Simulation results show that compared with the static-heterogeneous-model-based DRFTS algorithm,this algorithm can improve the guarantee ratio under the condition of hard real-time.