星载计算机系统高优先级恢复容错调度算法

Fault-Tolerant Scheduling Algorithm with High Priority Recovery for On-Board Computer Systems

为了在星载异构容错计算机上调度星载任务，在给出了星载计算机系统的容错调度模型的基础上，提出了一种星载计算机系统高优先级恢复容错调度算法（OHFSA），该算法通过引进可靠性代价量化了容错系统在维护容错调度时所需要的开销，并在调度过程中充分考虑了容错调度的实时性和低可靠性代价．同时，采用高优先级恢复机制可及时响应紧急时限下的恢复任务，从而提高了星载任务的整体实时性能．实验验证表明，OHFSA随着计算时间的增加，其所需的可靠性代价也相应增大．仿真结果表明，在任务负载不断增加的情况下，OHFSA的执行时间比目前算法减少20%～30%，特别是在恢复任务优先级比其主任务提高1个级别时，响应时间可减少近8．7%．

To schedule the on-board task in the on-board heterogeneous computer, a fault-tolerant scheduling algorithm with high priority recovery （OHFSA） for on-board computer systems was proposed by giving the fault-tolerant scheduling model. The overhead required by maintaining the fault-tolerant scheduling of the system can be quantified by introducing the dependability cost. The real-time property and low dependability cost of the fault-tolerant scheduling are considered sufficiently in the scheduling process. Meanwhile, the recovery tasks are responded timely under emergent time limit by utilizing the mechanism of the high priority recovery and the entire realtime performance of the on-board tasks is increased. Simulation results show that the dependability cost of OHFSA is increased with increasing the computing time, and when constantly increasing the task load the executed time of OHFSA compared with the existing algorithm is always decreased by 20%-30%. Especially, when the priority of recovery task is one level higher than its main task, the response time is improved about by 8. 7%.