Preemptive Behavior Analysis and Improvement of Priority Scheduling Algorithms
IPDPS '05 Proceedings of the 19th IEEE International Parallel and Distributed Processing Symposium (IPDPS'05) - Workshop 2 - Volume 03
Analysis and optimisation of hierarchically scheduled multiprocessor embedded systems
International Journal of Parallel Programming - Special Issue on Multiprocessor-based embedded systems
A component based real-time scheduling architecture
Architecting dependable systems
Resource augmentation for fault-tolerance feasibility of real-time tasks under error bursts
Proceedings of the 20th International Conference on Real-Time and Network Systems
| Hi-index | 0.00 |
Off-line scheduling and fixed priority scheduling (FPS)are often considered as complementing and incompatibleparadigms.A number of industrial applications demandtemporal properties (predictability, jitter constraints, end-to-end deadlines, etc.) Those are typically achieved by usingoff-line scheduling. The rigid off-line scheduling schemesused, however, do not provide for flexibility. FPS has beenwidely studied and used in a number of applications, mostlydue to its simple run-time scheduling, and small overhead.It can provide more flexibility, but is limited with respect topredictability, as actual start and completion times of executiondepend on run-time events.In this paper we show how off-line scheduling and FPSrun-time scheduling can be combined to get the advantagesof both - the capability to cope with complex timing constraints and flexibility. The paper assumes that a schedulefor a set of tasks with complex constraints has been constructedoff-line. It presents a method to analyze the off-lineschedule and derive an FPS task set with FPS attributespriority, offset, and period, such that the runtime FPS executionmatches the off-line schedule. It does so by analyzingthe schedule and setting up inequality relations for the prioritiesof the tasks under FPS. Integer linear programming(ILP) is then used to find a FPS priority assignment that fulfilsthe relations. In case the priority relations for the tasksof the off-line schedule are not solvable we split tasks intothe number of instances, to obtain a new task set with consistenttask attributes. Our schedule translation algorithmkeeps the number of newly generated artifact tasks minimal.