面向混合任务的自适应调度策略研究

Adaptive Scheduling Strategy Oriented to Hybrid Tasks

针对实时系统负载动态变化的问题,提出一种面向软实时的基于资源预留的反馈调度模型(Feedback Scheduling M odel based on Resource Reservation,FSM-RR),当负载发生变化时调整服务器的CPU带宽.接着针对混合任务提出一种自适应分层调度框架(Adaptive Hierarchical Scheduling Framework,AHSF),对不同类型任务采用相应的调度算法,保证硬实时任务在其截止期之前完成,同时尽可能的降低软实时以及非实时任务的截止期错失率.在RTSim平台上进行仿真实验,并与传统调度算法进行对比分析,实验结果表明本文提出调度算法具有较好的性能.

Aim at the problem that the workloads of real-time system dynamically change over time, this paper presents a FeedbackScheduling Model based on Resource Reservation （FSM-RR）oriented to soft real-time, which adapts the server＇s bandwidth whenworkloads change. Then, an Adaptive Hierarchical Scheduling Framework （AHSF） oriented to hybrid tasks is proposed, which providesappropriate scheduling algrothms for different types of tasks. This framework is designed to reduce the Deadline Miss Ratio（DMR） ofsoft real-time and non real-time tasks without jeopardizing the guarantee of hard real-time deadlines. Simulation experiment is imple-mented on RTSim, which shows that the framework we propose in this paper has better performance than the traditional scheduling al-gorithm.