Audrey Marchand

Dynamic Real-time Scheduling of Firm Periodic Tasks with Hard and Soft Aperiodic Tasks

In this paper, we address the problem of the dynamic scheduling of skippable periodic task sets (i.e., period tasks allowing occasional skips of instances), together with aperiodic tasks. Scheduling of tasks is handled thanks to the merging of two existing approaches: the Skip-Over task model and the EDL (Earliest Deadline as Late as possible) aperiodic task server. The objective is to provide two on-line scheduling algorithms, namely EDL-RTO and EDL-BWP, in order to minimize the average response time of soft aperiodic requests, while ensuring that the QoS (Quality of Service) of periodic tasks will never be less than a specified bound. We also extend our results to the acceptance of sporadic tasks (i.e., aperiodic tasks with deadlines). We show that these novel scheduling algorithms have better performance compared to related algorithms regarding aperiodic response time and acceptance ratio.

Memory Resource Management for Real-Time Systems

Dynamic memory storage has been widely used for years in computer science. However, its use in real-time systems has not been considered as an important issue, and memory management has not receive much consideration, whereas todays real-time applications are often characterized by highly fluctuating memory requirements. In this paper we present an approach to dynamic memory management for real-time systems. In response to application behavior and requests, the underlying memory management system adjusts resources to meet changing demands and user needs. The architectural framework that realizes this approach allows adaptive allocation of memory resources to applications involving both periodic or aperiodic tasks. Simulation results demonstrate the suitability of the proposed mechanism.

Dynamic scheduling of periodic skippable tasks in an overloaded real-time system

The need for supporting dynamic real-time environments where changes in workloads may occur requires a scheduling framework that explicitly addresses overload conditions, allows the system to achieve graceful degradation and supports a mechanism capable of determining the load to be shed from the system to handle the overload. In applications ranging from video reception to air-craft control, tasks enter periodically and have response time constraints, but missing a deadline is acceptable, provided most deadlines are met. Such tasks are said to be occasionally skippable and have an assigned skip parameter. We look at the problem of uniprocessor scheduling of skippable periodic tasks which consists in maximizing the robustness of the system defined as the global completion ratio. In this paper, we propose a novel scheduling Skip-over algorithm, called RLP/T, a variant of Earliest-Deadline First which adjusts the system workload such that tasks adhere to their timing and skip constraints and guarantees the best robustness.

RLP: enhanced QoS support for real-time applications

In this paper, we study the problem of scheduling periodic task sets defined under quality of service (QoS) constraints. In our approach, periodic tasks allow occasional skips of instances. A new algorithm, called RLP (red tasks as late as possible) based on the skip-over model and the EDL (earliest deadline as late as possible) scheduling strategy, is proposed to enhance the QoS observed for periodic tasks, i.e, the ratio of periodic tasks which complete before their deadline. We prove that our results are never worse than those obtained in previous work. Experimental results also show significant improvement achieved by our algorithm over RTO and BWP.

Dynamic Scheduling of Skippable Periodic Tasks in Weakly-Hard Real-Time Systems

This paper deals with dynamic scheduling in real-time systems that have quality of service requirements. We assume that tasks are periodic and may miss their deadlines, occasionally, as defined by the so-called Skip-Over model. In this paper, we present a dynamic scheduling algorithm, called RLP (Red as Late as possible), a variant of earliest deadline to make slack stealing and get better performance in terms of ratio of periodic task instances which complete before their deadline. Simulation results show that RLP outperforms the two conventional skip-over algorithms, namely RTO (Red Tasks Only) and BWP (Blue When Possible), introduced about ten years ago. Finally, we present the integration of these QoS scheduling services into CLEOPATRE, a free open-source library which offers selectable real-time facilities on shelves

Providing Memory QoS Guarantees for Real-Time Applications

Nowadays, systems often integrate a variety of applications whose service requirements are heterogeneous. Consequently, systems must be able to concurrently serve applications which rely on different constraints. This raises the problem of the dynamic distribution of the system resources (CPU, memory, network, etc.). Therefore, an integrated Quality of Service (QoS) management is needed so as to efficiently assign resources according to the various application demands. Within this paper, we focus on a dynamic approach of QoS management for memory resource allocation based on the Skip-Over model. We detail our solution and we show how it improves the service of task memory requests while providing them guarantees. Quantitative results using the TLSF allocator have been performed in order to evaluate the memory failure probability with and without memory QoS manager.