When executing soft real-time tasks in a shared processor, it is important to properly allocate the computational resources such that the quality of service requirements of each task are satisfied. In this paper we propose Adaptive Reservations, based on applying a feedback scheme to a reservation based scheduler. After providing a precise mathematical model of the scheduler, we describe how this model can be used for synthesising the controller by applying results from control theory. Finally, we show the effectiveness of our method by simulation and by experiments with an MPEG player running on a modified Linux kernel. 1.

In this paper, we present a general scheduling methodology for managing overruns in a real-time environment, where tasks may have different criticality and flexible timing constraints. The proposed method achieves isolation among tasks through a resource reservation mechanism which bounds the effects of task interference, but also performs efficient reclaiming of the unused computation times to relax the utilization constraints imposed by isolation. The enhancements achieved by the proposed approach resulted to be very effective with respect to classical reservation schemes. The performance has been evaluated by implementing the algorithm on a real-time kernel. The runtime overhead introduced by the scheduling mechanism has also been investigated with specific experiments, in order to be taken into account in the schedulability analysis. However, it resulted to be negligible in most practical cases.

Aframework for scheduling a number of different real-time applications on a single sharedpreemptable processor is proposed. This framework enforces complete isolation among the different applications, such that the behavior of each application is very similar to its behavior if it had been executing on a slower dedicatedprocessor. A scheduling algorithm that implements this framework is presentedandprovedcorrect. Keywords. Hard-real-time systems# Preemptive scheduling# Earliest deadline first# Inter-application isolation. 1.

... important computing platforms, which require low energy consumption while meeting the resource demands of a dynamic application workload. Most proposed dynamic voltage scaling (DVS) algorithms, targeting either best-effort or hard real-time systems, however, cannot be directly applied to such open mobile systems. This paper presents a framework to integrate DVS into soft real-time (SRT) scheduling for open mobile systems, achieving energy saving of DVS while preserving resource guarantees of SRT scheduling. The integrated framework makes three major contributions. First, multimedia applications reserve resource based on their average resource usage, without the knowledge of worst-case execution time, which is difficult to estimate in an open mobile environment. Second, the SRT scheduling ensures the correctness of reservation admission and enforcement in a variable speed context. Finally, the DVS manager reduces the processor energy consumption by utilizing the unallocated resource, reclaiming the allocated but unused resource, or avoiding the unused resource. Our extensive simulation results demonstrate that our framework is able to save 4% to 32% energy while slightly affecting application performance.

In this paper we present results obtained in the context of Quality of Service (QoS) control for soft real-time applications. The discussion addresses the issue of dynamically adjusting the bandwidth for a set of periodic tasks, when a reservation-based (RB) CPU scheduling policy is used. RB techniques are particularly suitable for this kind of applications since they allow an accurate mathematical modelling of the dynamic evolution of the QoS experienced by tasks. Based on this model, a control policy guaranteeing specified QoS levels for different tasks is illustrated, along with necessary and sufficient conditions for its existence. Moreover, the problem of steering a task QoS back into its nominal level is tackled, in response to deviations due to temporary overload conditions. Simulation results are reported, for the purpose of validating the approach.

In this paper, we address the problem of adaptively reserving the CPU to concurrent soft real-time tasks, in order to meet target Quality of Service requirements. First, we present two new techniques inspired to the idea of stochastic control. Then, we present a flexible and modular software architecture suitable for adaptive scheduling, realised as a minimally invasive set of modifications to the Linux Kernel. Finally, we show experimental results that validate our approach and prove its effectiveness in the context of multimedia applications.

An important class of soft real-time applications require dynamic allocation of computational resources in order to comply with their quality of service (QoS) requirements. These applications are characterised by large fluctuations in their computation time requirements. One of the biggest problems in such systems is how to assign the bandwidths to the software tasks so that every task meets its QoS requirements and computational resources are not wasted. In this paper, we present a novel feedback scheduling controller based on a scheduling strategy called resource reservation. First, we model the scheduler as a discrete time switching system; then, we present hybrid control techniques for the design of the feedback scheduler; finally, we report simulation results that show the effectiveness of our approach.

Reservation based (RB) scheduling is a class of scheduling algorithms that is well-suited for a large class of soft real-time applications. They are based on a “bandwidth ” abstraction, meaning that a task is given the illusion of executing on a dedicated slower processor. In this context, a crucial design issue is deciding the bandwidth that each task should receive. The point we advocate is that, in presence of large fluctuations on the computation requirements of the tasks, it can be a beneficial choice to dynamically adapt the bandwidth based on QoS measurements and on the subsequent application of feedback control (adaptive reservations). In this paper, we present two novel contributions to this research area. First, we propose three new control algorithms inspired to the ideas of stochastic control. Second, we present a flexible and modular software architecture for adaptive reservations. An important feature of this architecture is that it is realised by means of a minimally invasive set of modifications to the Linux kernel. 1.

Resource reservations are a very popular choice to schedule multimedia tasks. However, the high variability of the resource requirements hinders a static choice of the scheduling parameters. In this paper we address this problem by a combination of two strategies: adaptive reservations and resource reclaiming. The first one operates “locally” (using the information of a single task), the second one operates “globally ” distributing unused bandwidth between the tasks. In this paper, we show by analytical results and by extensive simulations that the two techniques can be safely and usefully combined. 1.

In this paper, we present a software architecture to support soft real-time applications, such as multimedia streaming and telecommunication systems, in open embedded systems. Examples of such systems are consumer electronic devices (as cellular phones, PDAs, etc.), as well as multimedia servers (video servers, VoIP servers, etc.) and telecommunication infrastructure devices. For such applications, it is important to keep under control the resource utilization of every task, otherwise the Quality of Service experienced by the users may be degraded. Our proposal is to combine a resource reservation scheduler (that allows us to partition the CPU time in a reliable way) and a feedback based mechanism for dynamically adjusting the CPU fraction (bandwidth) allocated to a tasks. In particular, our controller enables specified Quality of Service (QoS) levels for the application while keeping the allocated bandwidth close to its actual needs. The adaptation mechanism consists of the combination of a prediction and of a feedback correction that operates locally on each task. The consistency of the system is preserved by a supervisor component that manages overload conditions and enacts security policies. We implemented the framework in AQuOSA, a software architecture that runs on top of the Linux kernel. We provide extensive experimental validation of our results and offer evaluation of the introduced overhead, which is remarkably lower than the one introduced by other different solutions. I.