In recent years, there has been a rapid and wide spread of non-traditional computing platforms, especially mobile and portable computing devices. As applications become increasingly sophisticated and processing power increases, the most serious limitation on these devices is the available battery life. Dynamic Voltage Scaling (DVS) has been a key technique in exploiting the hardware characteristics of processors to reduce energy dissipation by lowering the supply voltage and operating frequency. The DVS algorithms are shown to be able to make dramatic energy savings while providing the necessary peak computation power in general-purpose systems. However, for a large class of applications in embedded real-time systems like cellular phones and camcorders, the variable operating frequency interferes with their deadline guarantee mechanisms, and DVS in this context, despite its growing importance, is largely overlooked/under-developed. To provide real-time guarantees, DVS must consider deadlines and periodicity of real-time tasks, requiring integration with the real-time scheduler. In this paper, we present a class of novel algorithms called real-time DVS (RT-DVS) that modify the OS's real-time scheduler and task management service to provide significant energy savings while maintaining real-time deadline guarantees. We show through simulations and a working prototype implementation that these RT-DVS algorithms closely approach the theoretical lower bound on energy consumption, and can easily reduce energy consumption 20% to 40% in an embedded real-time system.

/pdf/real-time-dynamic-voltage-scaling-for-low-power-embedded-39fxxz5wqc.pdf

Real-time dynamic voltage scaling for low-power embedded operating systems

This paper focuses on the problem of providing efficient run-time support to multimedia applications in a real-time system, where two types of tasks can coexist simultaneously: multimedia soft real-time tasks and hard real-time tasks. Hard tasks are guaranteed based on worst case execution times and minimum interarrival times, whereas multimedia and soft tasks are served based on mean parameters. The paper describes a server-based mechanism for scheduling soft and multimedia tasks without jeopardizing the a priori guarantee of hard real-time activities. The performance of the proposed method is compared with that of similar service mechanisms through extensive simulation experiments and several multimedia applications have been implemented on the HARTIK kernel.

/pdf/integrating-multimedia-applications-in-hard-real-time-4yc0jax79v.pdf

Integrating multimedia applications in hard real-time systems

Method and apparatus providing connection-oriented services for packet switched data communications networks. Directory services include distributed discovery of MAC addresses and protocol alias addresses. Topology services include a link state topology exchange among switches, which provides each switch with a complete topology graph of the network. This enables an access switch receiving a data packet to determine a complete path from a source end system to a destination end system. Another service includes resolution of broadcast frames to unicast frames, in order to reduce the amount of broadcast traffic. Policy restrictions may be applied prior to connection setup. Path determination services enable multiple paths from a source to a destination. Connection management includes source routed mapping of connections on the desired path. A distributed call rerouting service is provided wherein if a link on an active path fails, each switch receives a topology change notification and unmaps any connection involving the failed link. A broadcast/unknown service provides restricted flooding of nonresolvable packets. Furthermore, connection-oriented switching is provided based on the source and destination MAC addresses as a connection identifier. Still further, resolution of networks outside the switch domain is enabled by access switches listening for network and server route advertisements and maintaining best routes to said networks and servers. The best route metrics may be combined with best path metrics to determine a path from a first access switch to an egress switch connected to the external network.

Distributed connection-oriented services for switched communication networks

Cyber-physical systems (CPSs) are the next generation of engineered systems in which computing, communication, and control technologies are tightly integrated. Research on CPSs is fundamentally important for engineered systems in many important application domains such as transportation, energy, and medical systems. We overview CPS research from both a historical point of view in terms of technologies developed for early generations of control systems, as well as recent results on CPSs in many relevant research domains such as networked control, hybrid systems, real-time computing, real-time networking, wireless sensor networks, security, and model-driven development. We outline the potential for CPSs in many societally important application domains.

http://cesg.tamu.edu/wp-content/uploads/2012/03/ps_files/12-02-10-CPS-Centennial.pdf

Cyber–Physical Systems: A Perspective at the Centennial

We develop Feedback Control real-time Scheduling (FCS) as a unified framework to provide Quality of Service (QoS) guarantees in unpredictable environments (such as e-business servers on the Internet). FCS includes four major components. First, novel scheduling architectures provide performance control to a new category of QoS critical systems that cannot be addressed by traditional open loop scheduling paradigms. Second, we derive dynamic models for computing systems for the purpose of performance control. These models provide a theoretical foundation for adaptive performance control. Third, we apply established control methodology to design scheduling algorithms with proven performance guarantees, which is in contrast with existing heuristics-based solutions relying on laborious design/tuning/testing iterations. Fourth, a set of control-based performance specifications characterizes the efficiency, accuracy, and robustness of QoS guarantees. 
The generality and strength of FCS are demonstrated by its instantiations in three important applications with significantly different characteristics. First, we develop real-time CPU scheduling algorithms that guarantees low deadline miss ratios in systems where task execution times may deviate from estimations at run-time. We solve the saturation problems of real-time CPU scheduling systems with a novel integrated control structure. Second, we develop an adaptive web server architecture to provide relative and absolute delay guarantees to different service classes with unpredictable workloads. The adaptive architecture has been implemented by modifying an Apache web server. Evaluation experiments on a testbed of networked Linux PC's demonstrate that our server provides robust relative/absolute delay guarantees despite of instantaneous changes in the user population. Third, we develop a data migration executor for networked storage systems that migrate data on-line while guaranteeing specified I/O throughput of concurrent applications.

https://www.cse.wustl.edu/~lu/papers/thesis.pdf

Feedback Control Real-Time Scheduling: Framework, Modeling, and Algorithms*

Many real-time systems rely on static scheduling algorithms This includes cyclic scheduling, rate monotonic scheduling and fixed schedules created by off-line scheduling techniques such as dynamic programming, heuristic search, and simulated annealing However, for many real-time systems, static scheduling algorithms are quite restrictive and inflexible For example, highly automated agile manufacturing, command, control and communications, and distributed real-time multimedia applications all operate over long lifetimes and in highly non-deterministic environments Dynamic real-time scheduling algorithms are more appropriate for these systems and are used in such systems Many of these algorithms are based on earliest deadline first (EDF) policies There exists a wealth of literature on EDF-based scheduling with many extensions to deal with sophisticated issues such as precedence constraints, resource requirements, system overload, multi-processors, and distributed systems Deadline Scheduling for Real-Time Systems: EDF and Related Algorithms aims at collecting a significant body of knowledge on EDF scheduling for real-time systems, but it does not try to be all-inclusive (the literature is too extensive) The book primarily presents the algorithms and associated analysis, but guidelines, rules, and implementation considerations are also discussed, especially for the more complicated situations where mathematical analysis is difficult In general, it is very difficult to codify and taxonomize scheduling knowledge because there are many performance metrics, task characteristics, and system configurations Also, adding to the complexity is the fact that a variety of algorithms have been designed for different combinations of these considerations In spite of the recent advances there are still gaps in the solution space and there is a need to integrate the available solutions For example, a list of issues to consider includes: preemptive versus non-preemptive tasks, uni-processors versus multi-processors, using EDF at dispatch time versus EDF-based planning, precedence constraints among tasks, resource constraints, periodic versus aperiodic versus sporadic tasks, scheduling during overload, fault tolerance requirements, and providing guarantees and levels of guarantees (meeting quality of service requirements) Deadline Scheduling for Real-Time Systems: EDF and Related Algorithms should be of interest to researchers, real-time system designers, and instructors and students, either as a focussed course on deadline-based scheduling for real-time systems, or, more likely, as part of a more general course on real-time computing The book serves as an invaluable reference in this fast-moving field

Deadline Scheduling for Real-Time Systems: EDF and Related Algorithms

Knowledge of complexity, fundamental limits and performance bounds-well known for many scheduling problems-helps real time designers choose a good design and algorithm and avoid poor ones. The scheduling problem has so many dimensions that it has no accepted taxonomy. We divide scheduling theory between uniprocessor and multiprocessor results. In the uniprocessor section, we begin with independent tasks and then consider shared resources and overload. In the multiprocessor section, we divide the work between static and dynamic algorithms. >

/pdf/implications-of-classical-scheduling-results-for-real-time-3o9vxozxh5.pdf

Implications of classical scheduling results for real-time systems

In this paper we present five new on-line algorithms for servicing soft aperiodic requests in realtime systems, where a set of hard periodic tasks is scheduled using the Earliest Deadline First (EDF) algorithm. All the proposed solutions can achieve full processor utilization and enhance aperiodic responsiveness, still guaranteeing the execution of the periodic tasks. Operation of the algorithms, performance, schedulability analysis, and implementation complexity are discussed and compared with classical alternative solutions, such as background and polling service. Extensive simulations show that algorithms with contained run-time overhead present nearly optimal responsiveness. A valuable contribution of this work is to provide the real-time system designer with a wide range of practical solutions which allow to balance efficiency against implementation complexity.

Scheduling Aperiodic Tasks in Dynamic Priority Systems

https://www.springer.com/productFlyer_978-0-7923-8269-0.pdf?SGWID=0-0-1297-33320917-0

Deadline Scheduling for Real-Time Systems

We present a comparative study among scheduling algorithms which use different priority assignments and different guarantee mechanisms to improve the performance of a real-time system during overload conditions. In order to enhance the quality of service, we assume that tasks are characterized not only by a deadline, but also by an importance value. The performance of the scheduling algorithm is then evaluated by computing the cumulative value gained on a task set, i.e. the sum of the values of those tasks that completed by their deadline. The purpose of this simulation study was twofold. Firstly, we wanted to discover which priority assignment is able to achieve the best performance in overload conditions. Secondly, we were interested in understanding how the pessimistic assumptions made in the guarantee test affect the performance of the scheduling algorithms, and how much a reclaiming mechanism can compensate this degradation. Simulation results show that, without any admission control, value-density scheduling performs best. Simple admission control based on worst case estimates of the load worsen the performance of all value based algorithms. EDF scheduling performs best if admission control is used along with a reclaiming mechanism that takes advantage of early completions. Finally, scheduling by deadline before overload and by value during overload works best in most practical conditions.

/pdf/value-vs-deadline-scheduling-in-overload-conditions-5eqoqhg86h.pdf

Marco Spuri

Papers

Deadline Scheduling for Real-Time Systems: EDF and Related Algorithms

Implications of classical scheduling results for real-time systems

Scheduling Aperiodic Tasks in Dynamic Priority Systems

Deadline Scheduling for Real-Time Systems

Value vs. deadline scheduling in overload conditions