Transformation of Sporadic Tasks for Off-line Scheduling with Utilization and Response Time Trade-offs.
Transformation of Sporadic Tasks for Off-line Scheduling with Utilization and Response Time Trade-offs.
复制标题
通过利用率和响应时间权衡将零星任务转变为离线调度。
DOI：
--
复制
发表时间：
2011
期刊：
影响因子：
--
通讯作者：
G. Fohler
中科院分区：
文献类型：
--
作者：
J. Theis;G. Fohler
许多安全关键的实时系统遵循的时间触发的方法，使用离线调度方法的原因，确定性，简单的容错，certifiability等，由其本身的性质，离线时间表不能处理的活动是不知道完全在设计时。虽然已经提出了一些在运行时集成事件触发活动的方法，但这些方法需要在运行时修改简单且确定性的表查找调度。在本文中，我们提供了一个积极主动的方法来处理零星的任务与标准的离线调度时间触发系统在运行时不改变时间表。该方法是基于转换的零星任务的参数，以创建定期预约任务，包括在定期任务集的标准离线调度器。由此产生的调度表可以由常规的时间触发系统执行。转换需要比零星任务实际需要更多的保留利用率。我们使用这个不需要保留的利用率，以减少零星任务的最坏情况下的响应时间。本文还提出了一种保留利用率和零星任务最坏情况响应时间之间的权衡。我们提供了一个可行性分析，也考虑到交换开销的保留利用率和最坏情况下的响应时间的给定界限。通过这种分析，我们确定了可行的参数的保留任务，因此，提供了一个设计师的选择保留的利用率和最坏情况下的响应时间。
Many safety critical real-time systems follow the time triggered approach using off-line scheduling methods for reasons of determinism, simple fault tolerance, certifiability, etc. By their very nature, offline schedules cannot handle activities which are not known completely at design time. While some methods have been presented to integrate event-triggered activities at runtime, these require modification of the simple and deterministic table look up scheduling at runtime. In this paper, we provide a proactive method for handling sporadic tasks with standard off-line scheduling in time-triggered systems without changing the schedule at runtime. The method is based on transforming the parameters of sporadic tasks to create periodic reservation tasks, for including in regular task sets for standard off-line schedulers. The resulting scheduling tables can be executed by regular time-triggered systems. The transformation requires more reserved utilization than the sporadic tasks actually need. We use this not needed reserved utilization to reduce the worst case response time of the sporadic tasks. A trade-off between reserved utilization and worst case response time of sporadic tasks is presented in this paper, too. We provide a feasibility analysis also considering the switching overhead for given bounds on the reserved utilization and worst case response time. With this analysis, we determine feasible parameters of the reservation tasks and hence, provide a designer choice for the reserved utilization and the worst case response time.