Elasticity of Workloads and Periods of Parallel Real-Time Tasks

Orr, James; Gill, Chris; Agrawal, Kunal; Baruah, Sanjoy; Cianfarani, Christian; Ang, Phyllis; Wong, Christopher

doi:10.1145/3273905.3273915

Cited by 10 publications

(5 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Tian et al [19] extended the modified problem to include a "Quality-of-Control" metric as a part of the objective function of the quadratic optimization problem. More Orr et al [20], [21] provided algorithms to schedule sequential elastic tasks on multiprocessor systems and further extended the concept of the elastic task to federated DAG-based parallel task systems in [22], [23]. Beccari et al [24], [25] provided alternative algorithms to schedule similar applications by expressing the task period ranges in a linear programming formulation.…”

Section: Related Workmentioning

confidence: 99%

Scheduling Elastic Applications in Compositional Real-Time Systems

Salman

Mubeen

Markovic

et al. 2021

2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )

View full text Add to dashboard Cite

Many real-time applications have functional behaviour that requires variability in timing properties at runtime. The elastic task model provides a convenient mechanism to specify and encapsulate such variability and enables the modification of an application's periods during run-time to keep the application schedulable. Additionally, reservation-based scheduling techniques were proposed for the same purpose of taming unpredictability of timing variations, but with a different solution, i.e., by providing the spatial and temporal isolation for executing independent applications on the same hardware.In this paper, we combine the two approaches by proposing a two-level adaptive scheduling framework which is based on the elastic task model and the compositional framework based on the periodic resource model. The proposed framework minimises the number of requests for bandwidth adaption at the reservation (system) level and primarily enables schedulability by accounting for the application's elasticity by adjusting the periods. The motivation for this design choice is to rather localise the effect of the modifications within the application, without necessarily affecting all the applications at the system level compared to the changes made at the application level. The evaluation results show that the local application changes may often be enough to solve the problem of variability, significantly reducing the number of bandwidth adjustments, and therefore reducing the potential negative impact on all the applications of a system.

show abstract

Section: Related Workmentioning

confidence: 99%

Scheduling Elastic Applications in Compositional Real-Time Systems

Salman

Mubeen

Markovic

et al. 2021

2021 26th IEEE International Conference on Emerging Technologies and Factory Automation (ETFA )

View full text Add to dashboard Cite

show abstract

“…Low-utilization tasks with utilization U i < 1 are treated as sequential tasks under Federated Scheduling and are scheduled on the pool of remaining processors. Later work by Orr et al [17,18] extended the elastic task model to include parallel real-time DAG tasks under Federated Scheduling. To keep parallel elastic scheduling as semantically equivalent to Buttazzo's original model as possible, the authors present an optimal scheduling algorithm that directly solves a minimization problem similar to that given in Equation 1:…”

Section: Each Task Is Formally Represented As τmentioning

confidence: 99%

“…To perform this experiment, we extended the parallel (continuous) elastic concurrency platform from [17], which is available as open-source [1]. The underlying system calls, concurrency mechanisms, and synchronization techniques remain unchanged, but we replaced the original scheduling algorithm with Algorithm 1.…”

Section: Adaptive Virtual Real-time Hybrid Simulation Experimentsmentioning

confidence: 99%

“…In this section we look at the efect that discretization of tasks' periods and workloads has on schedulability of example tasksets through loss of system-wide processor utilization compared to the continuous version. We begin by randomly generating 10, 000 continuous parallel elastic tasks in the manner described in [17]. Each task is either period-elastic or computationally-elastic, and we schedule these continuous tasks according to the optimal algorithm provided in [17,18], noting the overall system utilization and objective function value.…”

Section: Effects Of Taskset Discretizationmentioning

confidence: 99%

“…We begin by randomly generating 10, 000 continuous parallel elastic tasks in the manner described in [17]. Each task is either period-elastic or computationally-elastic, and we schedule these continuous tasks according to the optimal algorithm provided in [17,18], noting the overall system utilization and objective function value. We then create four discretized tasksets from each continuous one by assigning a discretization delta of 0.05, 0.1, 0.2, and 0.5, to each task, meaning we discretize each task in such a way that in the new tasksets, there is a candidate utilization every 5%, 10%, 20%, and 50% of the way between U (min) and U (max ) , plus the endpoints.…”

Section: Effects Of Taskset Discretizationmentioning

confidence: 99%

See 2 more Smart Citations

Elastic Scheduling of Parallel Real-Time Tasks with Discrete Utilizations

Orr

Uribe

Gill

et al. 2020

Proceedings of the 28th International Conference on Real-Time Networks and Systems

Self Cite

View full text Add to dashboard Cite

Elastic scheduling allows for online adaptation of real-time tasks' utilizations (via manipulation of each task's computational workload or period) in order to maintain system schedulability in case the utilization demand of one or more tasks changes. This is done currently by assigning each task a utilization (and therefore period or workload) from within a continuous range of acceptable values. While this works well for anytime tasks whose quality of service improves with duration or for tasks that can run at any rate within a given range, many computationally-elastic tasks have a speciic workload for each distinct mode of operation and therefore cannot perform arbitrary amounts of work. Similarly, some period-elastic tasks must run at speciic (e.g. harmonic) rates. Therefore, a discrete set of candidate utilizations per task must be accommodated in such cases. This paper provides a new elastic task model in which each task has a discrete set of possible utilizations (instead of a continuous range). This allows users to specify only relevant candidate periods and workloads for each task. The discrete nature of this model also allows each task to modify its workload and/or its period when changing mode of operation, instead of adapting in only one dimension of task utilization. Elastic tasks thus can exploit both period elasticity and computational elasticity. This greatly increases both the diversity of adaptations available to each task and the kinds of tasks relevant to elastic scheduling.

show abstract

Algorithms for implementing elastic tasks on multiprocessor platforms: a comparative evaluation

Orr

Baruah

2021

Real-Time Syst

View full text Add to dashboard Cite

Elasticity of Workloads and Periods of Parallel Real-Time Tasks

Cited by 10 publications

References 7 publications

Scheduling Elastic Applications in Compositional Real-Time Systems

Scheduling Elastic Applications in Compositional Real-Time Systems

Elastic Scheduling of Parallel Real-Time Tasks with Discrete Utilizations

Algorithms for implementing elastic tasks on multiprocessor platforms: a comparative evaluation

Contact Info

Product

Resources

About