By Topic

Real-Time Systems, 2004. ECRTS 2004. Proceedings. 16th Euromicro Conference on

Date 30 June-2 July 2004

Filter Results

Displaying Results 1 - 25 of 37
  • Optimized slowdown in real-time task systems

    Page(s): 155 - 164
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (383 KB) |  | HTML iconHTML  

    Slowdown factors determine the extent of slowdown a computing system can experience based on functional and performance requirements. Dynamic voltage scaling (DVS) of a processor based on slowdown factors can lead to considerable energy savings. We address the problem of computing slowdown factors for dynamically scheduled tasks with specified deadlines. We present an algorithm to compute a near optimal constant slowdown factor based on the bisection method. As a further generalization, for the case of tasks with varying power characteristics, we present the computation of near optimal slowdown factors as a solution to convex optimization problem using the ellipsoid method. The algorithms are practically fast and have the same time complexity as the algorithms to compute the feasibility of a task set. Our simulation results show on an average 20% energy gains over known slowdown techniques using static slowdown factors and 40% gains with dynamic slowdown. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Biasing effects in schedulability measures

    Page(s): 196 - 203
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (441 KB) |  | HTML iconHTML  

    The performance of a schedulabilty test is typically evaluated by generating a huge number of synthetic task sets and then computing the fraction of those that pass the test with respect to the total number of feasible ones. The resulting ratio, however, depends on the metrics used for evaluating the performance and on the method for generating random task parameters. In particular, an important factor that affects the overall result of the simulation is the probability density function of the random variables used to generate the task set parameters. In this paper we discuss and compare three different metrics that can be used for evaluating the performance of schedulability tests. Then, we investigate how the random generation procedure can bias the simulation results of some specific scheduling algorithm. Finally, we present an efficient method for generating task sets with uniform distribution in a given space, and show how some intuitive solutions typically used for task set generation can bias the simulation results. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Fixed priority timing analysis of real-time systems with limited parallelism

    Page(s): 231 - 238
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (364 KB) |  | HTML iconHTML  

    Conventional real-time system implementations assume a von Neumann processor-memory architecture with at most one software process executing at any time. Scheduling approaches for such systems multiplex the processor among runnable processes. Fixed priority scheduling, notably, utilises timing analysis to determine offline the run-time timing behaviour of the system. Recent alternative real-time system implementations combine processor-memory architectures with reconfigurable hardware for speedup. Such reconfigurable hardware can execute many functions in parallel. This leads to a limited parallel system where a single software process can execute on the processor at any time, in parallel with a number of functions implemented in hardware. Conventional fixed priority timing analysis is unsuitable for this model, as fundamental assumptions are compromised, e.g. the notion of a critical instant. This paper proposes new timing analysis aimed at such limited parallel systems, illustrated by an example system utilising field programmable gate arrays as the reconfigurable hardware resource. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Reducing the number of preemptions in fixed priority scheduling

    Page(s): 144 - 152
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (317 KB) |  | HTML iconHTML  

    Fixed priority scheduling (FPS) has been widely studied and used in a number of applications, mostly due to its flexibility, simple run-time mechanism and small overhead. However, preemption related overhead in FPS may cause undesired high processor utilization, high energy consumption, or, in some cases, even infeasibility. In this paper, we propose a method to reduce the number of preemptions in legacy FPS systems consisting of tasks with priorities, periods and offsets. Unlike other approaches, our algorithm does not require modification of the basic FPS mechanism. Our method analyzes offline a set of periodic tasks scheduled by FPS, detects the maximum number of preemptions that can occur at run-time, and reassigns task attributes such that the tasks are schedulable by the same scheduling mechanism, while achieving a significantly lower number of preemptions. In some cases, there is a cost to pay for a lower number of preemptions in terms of increased amount of tasks and/or reduced task execution flexibility. Our method provides for the ability to choose a user-defined number of preemptions with respect to the cost to pay. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Low-latency hard real-time communication over switched Ethernet

    Page(s): 13 - 22
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (311 KB) |  | HTML iconHTML  

    Ethernet, the most widely used commodity network, increasingly moves toward switches as implementation technology thus replacing buses. This allows to use traffic shaping techniques to implement hard real-time distributed systems on commodity networks. However, because Ethernet switches lack build-in policing features, nodes connected by switched Ethernet need to be cooperative. Although the theory behind traffic shaping for real-time communication is known for some time, it has not been considered for Ethernet so far. In this paper, we present the implementation of traffic shaping on switched Ethernet technology. We make thorough experiments to understand the cost and practical limits of using fast and gigabit Ethernet for hard real-time communication. We do measurements to analyze properties of switches and delays that we can achieve using these switches. We further analyze the influences of non real-time Linux nodes sharing the network. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Managing quality-of-control performance under overload conditions

    Page(s): 53 - 60
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (359 KB) |  | HTML iconHTML  

    A common method for dealing with overload conditions in periodic task systems is to reduce the load by enlarging activation periods. When a periodic task implements a digital controller, however, the variation applied on the task period also affects the control law, which needs to be recomputed for the new activation rate. If computing a new control law requires too much time to be performed at runtime, a set of controllers has to be designed offline for different rates and the system has to switch to the proper controller in the presence of an overload condition. In this paper, we present a method for reducing the number of controllers to be designed offline, while still guaranteeing a given control performance. The proposed approach has been integrated with the elastic scheduling theory to promptly react to overload conditions. The effectiveness of the proposed approach has been verified through extensive simulation experiments performed on an inverted pendulum. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Generalized pre-scheduler

    Page(s): 127 - 134
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (383 KB)  

    A preschedule is a list of "executives" generated offline for a set of time-driven jobs. If in addition there exists event-driven workload competing for resources, sufficient slacks must be embedded in a valid preschedule. A prescheduler was defined in Wang et al. (2003) with the following assumptions: the competing workload is modeled as sporadic tasks and the time-driven jobs to be prescheduled are independent. In this paper, the prescheduling problem is generalized. First, the event-driven workload is no longer limited to sporadic tasks or any specific task model. Instead, a concept called "prescheduling space" is introduced so that the time-driven jobs to be prescheduled are inside the space, the competing event-driven workload is outside the space, and resource competition is abstracted to the space bounded by a supply function and a supply contract. Second, precedence constraints between the time-driven jobs to be prescheduled are allowed. The generalized prescheduler presented in this paper produces a valid preschedule if and only if one exists under the constraints of execution time bounds, valid scopes, precedence requirements and supply contract; otherwise, it returns a notice of failure. The generalized prescheduler runs in time polynomial to the number of time-driven jobs. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Collaborative resource allocation in wireless sensor networks

    Page(s): 35 - 44
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (932 KB) |  | HTML iconHTML  

    Traditional real-time resource allocation algorithms assume that the available resources in a system such as total CPU and network bandwidth do not change over time. However, in wireless sensor networks, the amount of available resources on the devices and the communication channel may not be constant for all times: for instance, a node can be turned off in some time intervals to increase its battery lifetime. Since sensor networks have limited network capacity and computational capabilities, it is crucial to optimally assign the available resources among all the active tasks. In this paper, we propose a fast online resource allocation algorithm (CoRAl) to dynamically reconfigure a sensor network whenever a new hot spot occurs (e.g., a new intruder is detected) or a node's activity changes (i.e., sleep vs. active mode). Our experimental results show that CoRAl provides always near-optimal resource allocation while keeping its online overhead low. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Improved prediction for Web server delay control

    Page(s): 61 - 68
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (405 KB) |  | HTML iconHTML  

    Control methods are being used increasingly for uncertainty management and QoS in modern Web server systems. Previous approaches have suggested combined feedforward and feedback control strategies, using queuing theory for feedforward delay prediction. While queuing theory allows one to predict delay as a function of arrival and service rates, the prediction applies only to long-term averages, and is therefore insensitive to sudden load changes. Unfortunately, Internet load is very bursty, leaving room for predictor improvement. The main contribution of this paper is an extension of the combined feedforward/feedback framework in which the queuing model is replaced with a predictor that instead uses instantaneous measurements to predict future delays. The proposed strategy is evaluated in simulation and by experiments on an Apache Web server. It is shown that the approach performs better than the combined queuing model based feedforward and feedback control presented in earlier papers. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Comparing real-time communication under electromagnetic interference

    Page(s): 45 - 52
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (281 KB) |  | HTML iconHTML  

    The contribution of this paper is threefold. First, an improvement to a previously published paper on the timing analysis of controller area network (CAN) in the presence of transient network faults is presented. A probabilistic fault model is considered, where random faults from electromagnetic interference occur according to a Poisson distribution. The analysis provides worst case response times for message frames, not as a single value, but as a probability distribution. Secondly, a similar result is produced for time-triggered CAN (TTCAN), a version of CAN based on time-driven schedule. Thirdly, these analyses are applied to an example message set and used to discuss the dependability of event-triggered and time-triggered communication in the presence of electromagnetic interference. The results indicate that, an event-triggered bus can generally provide a higher probability of timely-delivery of data than a time-triggered bus. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Integrating model-based design and preemptive scheduling in mixed time- and event-triggered systems

    Page(s): 119 - 126
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (258 KB) |  | HTML iconHTML  

    Model-based design is advocated as the method of choice when dealing with critical systems as well as high quality systems. However, it often abstracts implementation details such as execution times. This can be a problem when dealing with urgent events whose implementation requires preemptive scheduling. In this paper, we propose an intertask communication mechanism on top of a fixed-priority deadline monotonic preemptive execution scheme, which preserves the ordering of computations validated in a "zero-time" synchronous framework, and formally show its correctness. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • A variable rate execution model

    Page(s): 135 - 143
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (457 KB) |  | HTML iconHTML  

    We present a task model for adaptive real-time tasks in which a task's execution rate requirements are allowed to change at any time. The model, variable rate execution (VRE), is an extension of the rate-based execution (RBE) model. We relax the constant execution rate assumption of canonical real-time task models by allowing both the worst case execution time (WCET) and the period to be variable. The VRE model also supports tasks joining and leaving the system at any time. A schedulability condition for the VRE task model is presented that can be used as an online admission control test for the acceptance of new tasks or rate changes. Finally, a VRE scheduler was implemented in Linux as a loadable module, and several experiments demonstrate its correctness and analyze the overhead. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Multiprocessor energy-efficient scheduling with task migration considerations

    Page(s): 101 - 108
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (407 KB) |  | HTML iconHTML  

    This paper targets energy-efficient scheduling of tasks over multiple processors, where tasks share a common deadline. Distinct from many research results on heuristics-based energy-efficient scheduling, we propose approximation algorithms with different approximation bounds for processors with/without constraints on the maximum processor speed, where no task migration is allowed. When there is no constraint on processor speeds, we propose an approximation algorithm for two-processor scheduling to provide trade-offs among the specified error, the running time, the approximation ratio, and the memory space complexity. An approximation algorithm with a 1.13-approximation ratio for M-processor systems is also derived (M > 2). When there is an upper bound on processor speeds, an artificial-bound approach is taken to minimize the energy consumption with a 1.13-approximation ratio. An optimal scheduling algorithm is then proposed in the minimization of the energy consumption when task migration is allowed. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • On energy-constrained real-time scheduling

    Page(s): 165 - 174
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (316 KB) |  | HTML iconHTML  

    In this paper, we explore the feasibility and performance optimization problems for real-time systems that must remain functional during an operation/mission with a fixed, initial energy budget. We show that the feasibility problem is NP-hard in the context of systems with dynamic voltage scaling (DVS) capability and discrete speed levels. Then, we focus on energy-constrained periodic task systems where the available energy budget is not sufficient to meet all the deadlines. We propose techniques to maximize the total number of deadlines met and the total reward (utility) while guaranteeing the completion of the mission and a minimum performance for each task. We consider separately: (i) systems with or without DVS capability, and (ii) off-line (static) and on-line (dynamic) solutions to select most valuable jobs for execution. We also discuss the tractability of the involved optimization problems. Our on-line algorithms combine job promotion, job demotion and speed reduction techniques to maximize the system performance while guaranteeing the completion of the mission. We evaluate our schemes through simulations and show that the on-line schemes can yield significant performance improvements over static solutions. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Quality aware MPEG-2 stream adaptation in resource constrained systems

    Page(s): 23 - 32
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (337 KB) |  | HTML iconHTML  

    A number of algorithms have been presented for handling software decoding of MPEG-2 streams based on buffering or rate adjustment focusing on providing good average quality. The potentially arising drops in quality are tolerated, e.g., in transmissions over the Internet; they cannot be accepted in high quality consumer products: these mandate real-time methods. When resources, such as processing power or network bandwidth, are limited and not all frames can be handled, best effort decoders incur unnecessary quality decrease while wasting resources. In this paper, we present a method for quality aware frame selection for MPEG decoding under limited resources, based on realistic timing constraints for the decoding of MPEG streams. Given that not all frames can be processed, it selects those which provide the best picture quality while matching the available resources, starting only such decoding, which is guaranteed to be completed. We formulate the method as real-time scheduling problem and present its application in an example scheduling algorithm. Results from study based on realistic MPEG-2 video underline the effectiveness of our approach. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Pessimistic concurrency control and versioning to support database pointers in real-time databases

    Page(s): 261 - 270
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (318 KB) |  | HTML iconHTML  

    In this paper we present a concurrency control algorithm that allows co-existence of soft real-time, relational database transactions, and hard real-time database pointer transactions in real-time database management systems. The algorithm uses traditional pessimistic concurrency-control (i.e. locking) for soft transactions and versioning for hard transactions to allow them to execute regardless of any database lock. We provide formal proof that the algorithm is deadlock free and formally verify that transactions have atomic semantics. We also present an evaluation that demonstrates significant benefits for both soft and hard transactions when our algorithm is used. The proposed algorithm is suited for resource-constrained safety critical, real-time systems that have a mix of hard real-time control applications and soft real-time management, maintenance, or user-interface applications. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Executing aperiodic jobs in a multiprocessor constant-bandwidth server implementation

    Page(s): 109 - 116
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (799 KB) |  | HTML iconHTML  

    The constant bandwidth server (CBS) framework can be implemented on a preemptive uniprocessor platform to make full use of the computing capacity of the platform; when implemented upon a preemptive multiprocessor platform, however, it is known that the schedulable utilization is strictly less than the capacity of the platform. The issue of using the excess processing capacity is addressed here, and an algorithm is presented, and proven correct, that uses this excess capacity to provide guaranteed real-time service to aperiodic jobs. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Stabilizing execution time of user processes by bottom half scheduling in Linux

    Page(s): 71 - 78
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (568 KB) |  | HTML iconHTML  

    The CPU time allocated to user processes is rendered inaccurate by an unexpectedly and frequently occurring interrupt and a bottom half that consumes most interrupt processing time. Additionally, when the time consumed in the kernel mode greatly fluctuates with interrupt processing, the scheduler cannot distribute CPU time to user processes normally. This problem can dramatically distort the stable execution time of user processes. In addition, such time-sensitive applications as multimedia players cannot provide consistent quality. To overcome this stolen-time problem, we propose a bottom half scheduling approach that dynamically restricts the maximum time consumed by bottom halves. In this paper, we implement our proposed scheme in Linux 2.4. In addition, we show that the fluctuation of CPU time allocated to user processes by stolen-time can be shrunk with our proposed scheme by means of experiments using a multimedia application. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Schedulability-driven partitioning and mapping for multi-cluster real-time systems

    Page(s): 91 - 100
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (418 KB)  

    We present an approach to partitioning and mapping for multicluster embedded systems consisting of time-triggered and event-triggered clusters, interconnected via gateways. We have proposed a schedulability analysis for such systems, including a worst-case queuing delay analysis for the gateways responsible for routing intercluster traffic. Based on this analysis, we address design problems characteristic to multiclusters: partitioning of the system functionality into time-triggered and event-triggered domains, and mapping of processes onto architecture nodes. We present a branch-and-bound algorithm for solving these problems. Our algorithm is able to find schedulable implementations under limited resources, achieving an efficient utilization of the system. The developed algorithms are evaluated using extensive experiments and a real-life example. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • An event stream driven approximation for the analysis of real-time systems

    Page(s): 187 - 195
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (295 KB) |  | HTML iconHTML  

    This paper presents a new approach to understand the event stream model. Additionally a new approximation algorithm for the feasibility test of the sporadic and the generalized multiframe task system scheduled by earliest deadline first is presented. The new algorithm has polynomial complexity to solve the problem of schedulability analysis. The approximation error of the algorithm is bounded. In contrary to earlier work, where the error depends on the different deadlines of the tasks, the error of our algorithm depends only on the capacity of the chosen processor. It guarantees the acceptances on a processor with slightly higher capacity than the unknown optimal processor. While the algorithm is scalable and the run-time depends on the chosen error, a trade-off between running time and error is possible. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • QoS control strategies for high-quality video processing

    Page(s): 3 - 12
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (400 KB) |  | HTML iconHTML  

    Video processing in software is often characterized by highly fluctuating, content-dependent processing times, and a limited tolerance for deadline misses. We present an approach that allows close-to-average-case resource allocation to a single video processing task, based on asynchronous, scalable processing, and QoS adaptation. The QoS adaptation balances different QoS parameters that can be tuned by user-perception experiments: picture quality, deadline misses, and quality changes. We model the balancing problem as a discrete stochastic decision problem, and propose two closely related solution strategies, for which the processing-time statistics are determined offline and at run time, respectively. We enhance both strategies with a compensation for structural (non-stochastic) load fluctuations. Finally, we validate our approach by means of simulation experiments, and conclude that both enhanced strategies perform close to the theoretical optimum. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Maintaining data freshness in distributed real-time databases

    Page(s): 251 - 260
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (358 KB) |  | HTML iconHTML  

    Many real-time systems need to maintain fresh views which are derived from shared data that are distributed among multiple sites. When a base data item changes, all derived views that are based on it need to be recomputed. There are two major derived data re-computation strategies - immediate update and on-demand update. However, they both have their advantages and limitations. In this paper, we study the performance of derived data update using immediate and on-demand strategies in distributed real-time databases and identify several criteria for choosing proper update policies. Based on these criteria, we propose a derived data update algorithm. In our algorithm, the update policy of a particular derived data item is determined dynamically by its access frequency, current transaction miss ratio and the system utilization. A thorough simulation study shows that our algorithm outperforms immediate and on-demand update in most cases. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Worst-case execution time analysis for dynamic branch predictors

    Page(s): 215 - 222
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (292 KB) |  | HTML iconHTML  

    Branch prediction mechanisms are becoming commonplace within modern microprocessors. For developers of real-time control systems, the mechanisms present predictability problems. The reasons are they increase the difficulty in analysing software for its worst-case execution time without introducing unmanageable pessimism and they increase the variability of the software's execution times. In this paper, we improve upon existing branch prediction analysis by taking into account the semantic context of the branches in the source code in order to classify them as either being easy-to-predict or hard-to-predict. Based on this classification we provide a static analysis approach for bimodal and global-history branch prediction schemes. The analysis is applied to a previously published example with the benefit that a more detailed explanation of its results is obtained. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Analysis of tree-shaped transactions in distributed real time systems

    Page(s): 239 - 248
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (344 KB) |  | HTML iconHTML  

    A worst case response time analysis exploiting precedence constraints between fixed priority scheduled tasks in tree-shaped transactions is presented. The algorithm extends existing methods for analysis of linear transactions allowing a wider range of systems, in which tasks may trigger more than one succeeding task on their completion, i.e. the transactions form trees. It also improves existing methods, producing tighter response time bounds for tasks in both linear and tree-shaped transactions. The improvement is due to better exploitation of the precedence relations between tasks, which also makes the analysis faster than earlier. Simulation results show a significant reduction of estimated worst case response times when compared to earlier analysis methods for both linear and tree-shaped transactions. Such improvement leads to less pessimistic schedulability tests for distributed fixed priority scheduled systems. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.
  • Static determination of probabilistic execution times

    Page(s): 223 - 230
    Save to Project icon | Request Permissions | Click to expandQuick Abstract | PDF file iconPDF (364 KB) |  | HTML iconHTML  

    Most previous research done in probabilistic schedulability analysis assumes a known distribution of execution times for each task of a real-time application. This is however not trivial to determine it with a high level of confidence. Methods based on measurements are often biased since not in general exhaustive on all the possible execution paths, whereas methods based on static analysis are mostly worst-case execution time - WCET-oriented. Using static analysis, this work proposes a method to obtain probabilistic distributions of execution times. It assumes that the given real time application is divided into multiple tasks, whose source code is known. Ignoring in this paper hardware considerations and based only on the source code of the tasks, the proposed technique allows designers to associate to any execution path an execution time and a probability to go through this path. A source code example is presented to illustrate the method. View full abstract»

    Full text access may be available. Click article title to sign in or learn about subscription options.