Sporadic Server Research Articles

Exact response time analysis of fixed priority systems based on sporadic servers

Abstract Real-time servers have been widely explored in the scheduling literature to predictably execute aperiodic activities, as well as to allow hierarchical scheduling settings. As they facilitate timing isolation between different software components, there is a renewed interest for the adoption of fixed priority real-time servers in the automotive domain, as a way to implement more efficient reservation mechanisms than TDMA-based methods. In this paper, we focus on the Sporadic Server, which is the only fixed priority server supported by the POSIX standard. Despite its popularity, we realized that only sufficient schedulability conditions exist for real-time systems scheduled with a Sporadic Server. Thus, we develop a formal characterization of an exact response time analysis for fixed priority systems based on Sporadic Servers in a multi-level scheduling setting under preemptive scheduling. We then provide an experimental characterization of the schedulabilty improvement that can be obtained with respect to existing sufficient schedulability tests, proving the effectiveness of the proposed exact analysis.

On the efficiency of sporadic servers on ethernet with FTT-SE

Ethernet is generating a growing interest as a network for real-time embedded systems such as airplanes and automobiles. In this realm, network reservations appear as important design elements that favor composability in the time domain. One Ethernet protocol that provides such reservations is FTT-SE. In this work-in-progress we make initial steps to assess the efficiency of one particular worst-case network delay analysis for sporadic reservations associated to asynchronous messages using extensive simulation. With over 1000 message sets we found that analytic values match the observations, on average, in 31% of the messages across all generated sets whereas for 85% of the cases, the analysis upper bounds are within 2.85 times the observed values.

Response Time Analysis for Sporadic Server Based Budget Scheduling in Real Time Virtualization Environments

Virtualization techniques for embedded real-time systems typically employ TDMA scheduling to achieve temporal isolation among different virtualized applications. Recent work already introduced sporadic server based solutions relying on budgets instead of a fixed TDMA schedule. While providing better average-case response times for IRQs and tasks, a formal response time analysis for the worst-case is still missing. In order to confirm the advantage of a sporadic server based budget scheduling, this paper provides a worst-case response time analysis. To improve the sporadic server based budget scheduling even more, we provide a background scheduling implementation which will also be covered by the formal analysis. We show correctness of the analysis approach and compare it against TDMA based systems. In addition to that, we provide response time measurements from a working hypervisor implementation on an ARM based development board.

Simulation and Evaluation of Multiple Sporadic Servers Rescheduling

This paper deals with task scheduling in heterogeneous distributed embedded computer environments of renewable and timing variable resources. Real-time application performance can be considerably influenced by unexpected events in the environment. The predictive-reactive rescheduling procedure can enable significant improvements. The rescheduling procedure based on multiple server tasks enables usage of one's own resources, but also a fast connection to nodes with appropriate performance that will soften the influence of disturbances. We developed a tool that enables simulation of application execution conditions for a given environment and extraction of better scheduling parameter settings. Numerous scheduling parameters are improved with an increase in the number of nodes available for rescheduling. We also investigate cost-effectiveness of the method based on rescheduling time cost. The results demonstrate a possibility of achieving performance improvement by tuning cost parameters.

On-line schedulability tests for adaptive reservations in fixed priority scheduling

Adaptive reservation is a real-time scheduling technique in which each application is associated a fraction of the computational resource (a reservation) that can be dynamically adapted to the varying requirements of the application by using appropriate feedback control algorithms. An adaptive reservation is typically implemented by using an aperiodic server (e.g. sporadic server) algorithm with fixed period and variable budget. When the feedback law demands an increase of the reservation budget, the system must run a schedulability test to check if there is enough spare bandwidth to accommodate such increase. The schedulability test must be very fast, as it may be performed at each budget update, i.e. potentially at each instance of a task; yet, it must be as efficient as possible, to maximize resource usage. In this paper, we tackle the problem of performing an efficient on-line schedulability test for adaptive resource reservations in fixed priority schedulers. In the literature, a number of algorithms have been proposed for on-line admission control in fixed priority systems. We describe four of these tests, with increasing complexity and performance. In addition, we propose a novel on-line test, called Spare-Pot algorithm, which has been specifically designed for the problem at hand, and which shows a good cost/performance ratio compared to the other tests.

A Hybrid Real-Time Scheduling Approach on Multi-Core Architectures

This paper proposes a hybrid scheduling approach for real-time system on homogeneous multi-core architecture. To make the best of the available parallelism in these systems, first an application is partitioned into some parallel tasks as much as possible. Then the parallel tasks are dispatched to different cores, so as to execute in parallel. In each core, real-time tasks can run concurrently with non-real-time tasks. The hybrid scheduling approach uses a two-level scheduling scheme. At the top level, a sporadic server is assigned to each scheduling policy. Each sporadic server is used to schedule the dispatched tasks according to its scheduling policy. At the bottom level, a rate-monotonic OS scheduler is adopted to maintain and schedule the top level sporadic servers. The schedulability test is also considered in this paper. The experimental results show that the hybrid scheme is an efficient scheduling scheme.

Multiple Servers and Capacity Sharing for Implementing Flexible Scheduling

The aperiodic server scheduling mechanism (like the Deferrable Server or Sporadic Server) is an adequate technique to provide service for soft and firm tasks in flexible environments. For handling multiple sources of events either a single server or multiple servers can be used. A single server minimizes the number of capacity exhaustions but provides a poor performance when the sources have different temporal requirements. This problem can be solved using multiple servers at different priorities, however this approach suffers from much more capacity exhaustions as the capacity has to be statically partitioned among them. We overcome these two problems by introducing the capacity sharing protocol in which a server can use the unused capacity of other servers. In this paper, we describe the protocol, evaluate its performance in comparison with single and multiple servers without capacity sharing and we analyze its implementation complexity. The results of the simulation analysis show that the capacity sharing protocol exhibits a better performance than single server and multiple servers without capacity sharing.

Joint scheduling of garbage collector and hard real-time tasks for embedded applications

Programs with complex data structures often require dynamic memory management based on automatic memory reclamation (garbage collection). A major problem in adopting garbage collection for embedded real-time systems is that it often causes unpredictable pauses and that, as a result of such delays, hard real-time tasks may miss their deadlines. In this paper, we propose a new real-time garbage collection technique for embedded applications. In our approach, the system jointly schedules garbage collector and hard real-time tasks using one of the aperiodic server approaches. Our study focuses on reducing memory requirements while guaranteeing the deadlines of hard real-time tasks. To achieve this objective, we model garbage collection requests as aperiodic hard real-time tasks, and schedule them using the sporadic server (SS). We also present an effective live-memory analysis to bound the worst-case garbage collection time. Performance analysis shows that the proposed approach considerably reduces the worst-case memory reservation compared with a background policy. The analytic results are verified by simulation based on trace-driven data.

Extending Ada's real-time systems annex with the POSIX scheduling services

In this paper we propose extending the scheduling model of the Ada 95 Real-Time Systems Annex with the services specified in the Real-Time POSIX standard. These services include a round robin within priorities scheduling policy, a sporadic server scheduling policy, and execution time clocks and timers. With these services the Ada Real-Time Annex will enable addressing a larger number of application requirements.

Scheduling garbage collector for embedded real-time systems

This paper proposes a new scheduling method for multiple mutators and a garbage collector running on embedded real-time systems with a single processor and no virtual memory. The hard real-time tasks should reserve a certain amount of heap memory to prevent memory starvation and/or deadline miss. Since the memory requirement depends on the worst-case response time of a garbage collector, the traditional approach in which garbage collection is performed in the background demands large memory space. The proposed scheduling algorithm is based on an aperiodic scheduling technique, sporadic server . This paper also presents a modified copying garbage collection algorithm with hardware support. In order to minimize the worst-case response time of a garbage collector thus reducing the memory requirement, the garbage collector runs as the highest priority task with a preset bandwidth. This paper also investigates the schedulability of a garbage collector and mutator tasks as well as the worst-case memory requirement. Performance analysis shows that the proposed algorithm can provide a considerable reduction in the worst-case memory requirement compared with the background policy. Simulation results demonstrate that the proposed algorithm can produce the feasible memory requirement comparable to the complex on-line scheduling algorithm such as slack stealing.

Minimizing the effects of jitter in distributed hard real-time systems

In this paper we present a study of the effects caused on distributed real-time systems by the presence of jitter in the activation of tasks and messages. We show that although jitter has usually a small impact on the schedulability of single-processor systems, in distributed architectures the worst-case response times are significantly delayed. Reducing or eliminating jitter in these systems can increase the schedulability of the system up to 50% more than when jitter is permitted. The delay effects caused by jitter can be prevented by using a bandwidth-preserving scheduling algorithm such as the sporadic server. Since this kind of scheduling policy is not designed for communication networks, we describe in this paper how to adapt and implement the sporadic server algorithm on communication networks. Using the sporadic server both in the processors and networks, we can build distributed systems with up to 100% utilization of the CPUs and communication resources, while still guaranteeing that hard real-time requirements are met.

Implementing analysable hard real-time sporadic tasks in Ada 9X

In this paper we illustrate how systems containing hard real-time sporadic tasks can be analysed for their worst case behaviour. In order to undertake this schedulability analysis, it is necessary to define the minimum inter-arrival time and/or maximum arrival frequency of sporadic tasks. Furthermore, at run-time it is essential to ensure that sporadic tasks are not invoked more often than has been guaranteed by the analysis. We assume that sporadics are invoked by interrupts and that interrupts can be masked under software control.Sporadic tasks are often analysed using the notion of bandwidth preserving sporadic servers within the Rate Monotonic Scheduling Analysis scheme. At run-time this requires the underlying kernel to support complex execution time monitoring mechanisms. Unfortunately such mechanisms are not generally supported by Ada 9X. This paper shows that by using Deadline Monotonic Scheduling Analysis there is no need to resort to bandwidth preserving sporadic servers, and the facilities available in Ada 9X can be used.

Aperiodic task scheduling for Hard-Real-Time systems

This thesis develops the Sporadic Server (SS) algorithm for scheduling aperiodic tasks in real-time systems. The SS algorithm is an extension of the rate monotonic algorithm which was designed to schedule periodic tasks. This thesis demonstrates that the SS algorithm is able to guarantee deadlines for hard-deadline aperiodic tasks and provide good responsiveness for soft-deadline aperiodic tasks while avoiding the schedulability penalty and implementation complexity of previous aperiodic service algorithms. It is also proven that the aperiodic servers created by the SS algorithm can be treated as equivalently-sized periodic tasks when assessing schedulability. This allows all the scheduling theories developed for the rate monotonic algorithm to be used to schedule aperiodic tasks. For scheduling aperiodic and periodic tasks that share data, this thesis defines the interactions and schedulability impact of using the SS algorithm with the priority inheritance protocols. For scheduling hard-deadline tasks with short deadlines, an extension of the rate monotonic algorithm and analysis is developed. To predict performance of the SS algorithm, this thesis develops models and equations that allow the use of standard queueing theory models to predict the average response time of soft-deadline aperiodic tasks serviced with a high-priority sporadic server. Implementation methods are also developed to support the SS algorithm in Ada and on the Futurebus+.