S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard RealTime Systems," Proceedings of the IEEE Real-Time Systems Symposium, pp. 166-174, 1986.  Home/Search   Document Not in Database   Summary   Related Articles   Check

.... management have focused on realtime systems that are static (in both the above senses) and perform device level, sampled data monitoring and regulatory control that is usually centralized [Bak91, LRT92, LSS87, RCF97, RTL93, SB96, SKG91, SLS88, SSL89, TLS96, XP90] but occasionally distributed [CSR86, HS92, Kao95, KDK89, RSZ89, Shi91, SR91, SRC85, Ver95, WSM95]. These techniques cannot be practically employed or adapted for systems that are dynamic [Jen99, Koob96, Sta96, SK97] Dynamic real time computer systems and their applications have workload characteristics that are inherently posteriori. Thus, such systems require adaptive resource management ....

S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proceedings of The IEEE Real-Time Systems Symposium, 1986.

....of each other. This is useful in heterogeneous systems where different scheduling strategies are employed in different processors. It also allows for parallel scheduling of sequential tasks, a feature that can aid in increasing the scheduling performance for systems with on line scheduling [13]. I2. The release jitter of each task as caused by precedence constraints can be eliminated. As discussed in, for example, 14, 15] uncontrolled release jitter can negatively affect the schedulability of real time applications. The deadline distribution problem is addressed in the context of ....

S. Cheng, J. A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proc. of the IEEE RealTime Systems Symposium, New Orleans, Louisiana, Dec. 2--4, 1986, pp. 166--174.

....prototypes of the major components depicted in Figure 3. An important goal is to evlove the infrastructure to address (1) quality of service assessment and (2) adaptive resource management for path based systems. 4. Related Work Resource management in real time systems has been addressed in [1 5,7 33] (a non exhaustive list) In much of this work, a task is defined as an independent program or a program component having a single thread of control, with timing constraints imposed on it. Schedule construction or schedulability analysis is performed statically using some heuristics fixed ....

....which are designed to handle anticipated peak demands leads to (1) inefficient resource utilization and (2) inadequate crisis response. The engineering of systems having the dynamic, large grain characteristics of the AAW system present many challenges to technologists. Can the existing paradigms [1 5,7 33] for real ime computing be extended to meet the challenges Is a new paradigm needed The answers to these questions will become clear as systems such as the Navy s 21 st Century Surface Combatant are engineered. 5. Conclusions This paper describes a representative shipboard control system, and ....

S. Cheng, J. Stankovic, and K. Ramamritham,"Dynamic Scheduling of Groups of Tasks With Precedence Constraints in Distributed Hard Real-Time Systems," Proceedings of The IEEE Real-Time Systems Symposium,1986. 10

....that meets the deadlines may exist without interleaving or overlapping. Using the timing information we have developed a compact task graph representation which can be used to generate schedules efficiently [8] We are currently developing extensions to algorithms for scheduling periodic tasks [2, 16] for generating interleaved schedules. We have also developed other notions of non intrusion, and static analysis algorithms for their detection, which are applicable to non real time applications [7] 17] ....

S. Cheng, J.A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proceedings of the 7th Real-Time Systems Symposium, pp.166-174, 1986.

....between the application s QoS specifications and the network resources. The V Gammanet provides a framework for flexible support of both real time and non real time communication requirements. B. Real Time Scheduling Many interesting results have been developed for timeline based schedulers [4, 19, 18] and rate based schedulers. The latter technique has been extensively studied for the case of periodic applications, including both the general case [14] and with more specific cases (priority inversion and other fixed priority based protocols [3, 17, 21] The algorithms that have considered the ....

S. Cheng, J. A. Stankovic, and K. Ramamritham. Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems. In Proc. IEEE Real-Time System Symposium, pages 175--180, 1986.

....scheduling policy decides to which node the task should be migrated. Scheduling precedence constrained tasks on Distributed Systems (H DY D PC) This section considers dynamic scheduling mechanisms for precedence constrained tasks. An algorithm derived by Cheng, Stankovic and Ramamritham [11] partitions groups of tasks which have the same deadline into subgroups, which are then scheduled to different nodes to run in parallel. This algorithm attempts to dynamically schedule groups of tasks which have arbitrary precedence constraints on a distributed system. The components of the node ....

Cheng, S-C., Stankovic, J.A., Ramamritham, K., Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real Time Systems. IEEE RealTime System Symposium, 1986,

.... both timing and precedence constraints, but their definition of timing constraints only in terms of periodic events (e.g. sampling rates) fixed execution times for events, and deadline or timeliness requirements provides too limited a model for the timing issues constraining spacecraft commands [2, 19, 20, 22]. Discrete event simulation or emulation models the states of the spacecraft and the events that cause a transition from one state to another [5, 7, 13] A simulation is run iteratively with a series of different combinations of process start times or fault injection times. Each simulation ....

Cheng, Shengchang, John A. Stankovic, and Krithivasan Ramamritham. "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems. " IEEE Real-Time System Symposium (1986): 166-174.

....the PSP and SSP strategies and study their performance benefit by applying them to a typical distributed application. Finally, we conclude our paper in Section 9. 2 Related Work The problem of scheduling real time distributed tasks has been studied mostly in a hard realtime environment (e.g. see [5, 11, 4, 12, 13, 3]) Some of these works concentrate on the taskallocation load balancing strategies, and on the schedulability of tasks. The communication overhead between subtasks with precedence relationship is considered when a task allocation decision is made. For example, subtasks that communicate a lot are ....

S. Cheng, J. A. Stankovic, and K. Ramaritham. Dynamic scheduling of groups of tasks with precedence constraints in distributed hard real-time systems. In Proceedings of IEEE Real-Time Systems Symposium, pages 166--174, 1986.

....for finding a schedule, additional tasks are split. In this incremental way, our algorithm increases in scheduling complexity gradually, if and only if the increase is definitely required. Other researchers have addressed the problem of scheduling multitask applications on a distributed system [1, 2, 7, 9]. In some cases the scheduling of independent tasks is considered [6] while others have attempted to treat the problem of scheduling interrelated tasks [1, 2, 9] Some of the techniques split the task graph into independent tasks [1, 2] In section 2 we introduce busy idle profiles and the CTG ....

....is definitely required. Other researchers have addressed the problem of scheduling multitask applications on a distributed system [1, 2, 7, 9] In some cases the scheduling of independent tasks is considered [6] while others have attempted to treat the problem of scheduling interrelated tasks [1, 2, 9]. Some of the techniques split the task graph into independent tasks [1, 2] In section 2 we introduce busy idle profiles and the CTG representation. In section 3 we present our multiprocessor scheduling algorithm based upon compact task graphs and present performance results obtained through ....

[Article contains additional citation context not shown here]

S. Cheng, J.A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proc. 7th Real-Time Syst. Symp., pp.166-174, 1986.

....910 Figure 1. Producer Consumer Example. 4 GUPTA, R. AND SPEZIALETTI, M. is that resulting ETG is large since its construction essentially leads to the unrolling of loops containing input or output statements. The larger the TGs, the greater is the complexity of scheduling. Thus, as discussed by Cheng, Stankovic, and Ramamritham (1986), clustering techniques are required to schedule groups of tasks in order to reduce the overhead of scheduling in situations where the schedule is being generated on line. An ETG represents one possible partitioning of tasks into smaller subtasks. There may be other partitionings that enable the ....

....schedule groups of subtasks from an ETG in a single step. It should be noted that although CTGs are labeled with greater amount of information than ETGs, the form in which this information is expressed does not increase the complexity of scheduling but rather simplifies the scheduling of clusters. Cheng, Stankovic, and Ramamritham (1986) presented a technique for clustering ordinary tasks into groups of tasks. The assignment of clusters to processing nodes is carried out for efficient on line scheduling. A distributed scheduling algorithm is used to distribute the clusters among the nodes with surplus processing power. In our ....

[Article contains additional citation context not shown here]

Cheng, S., J.A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proc. 7th IEEE Real-Time Systems Symposium, December, 1986, pp. 166--174.

.... The problem of scheduling real time distributed tasks has been studied mostly in a hard real time environment (e.g. on multiprocessor distributed systems [SRC85, RSS90, WRS92] flow shop [BL90, BL92] pipeline scheduling [L 87] and distributed tasks with precedence constraints [HS92, Ram90, CSR86, CL87] Some of these works concentrate on the task allocation load balancing strategies, and on the schedulability of tasks. For example, in [CSR86] the authors studied the problem of dynamic scheduling and allocation of groups of tasks with precedence constraints in distributed hard realtime ....

....[SRC85, RSS90, WRS92] flow shop [BL90, BL92] pipeline scheduling [L 87] and distributed tasks with precedence constraints [HS92, Ram90, CSR86, CL87] Some of these works concentrate on the task allocation load balancing strategies, and on the schedulability of tasks. For example, in [CSR86] the authors studied the problem of dynamic scheduling and allocation of groups of tasks with precedence constraints in distributed hard realtime systems. They consider a homogeneous system (i.e. tasks can execute at any node in the system) on which task groups are executed. A task group ....

S. Cheng, J. A. Stankovic, and K. Ramaritham. Dynamic scheduling of groups of tasks with precedence constraints in distributed hard real-time systems. In Proceedings of IEEE Real-Time Systems Symposium, pages 166--174, 1986.

....of each other. This is useful in heterogeneous systems where different scheduling strategies are employed in different processors. It also allows for parallel scheduling of sequential tasks, a feature that can be a means for increasing the scheduling performance in systems with on line scheduling [13]. I2. The release jitter of each task as caused by precedence constraints can be eliminated. As have been discussed in, for example, 14, 15] uncontrolled release jitter can negatively affect the schedulability of real time applications. The deadline distribution problem is addressed in the ....

S. Cheng, J. A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," Proc. of the IEEE Real-Time Systems Symposium, New Orleans, Louisiana, Dec. 2--4, 1986, pp. 166--174.

....for modeling the timing constraints on spacecraft. Similarly, many techniques that are currently available to model timing constraints tend to ignore precedence constraints. A wide variety of powerful formalisms exists to model the specifications and behavior of real time systems. See, e.g. [1, 4, 5, 6, 7, 13, 14, 15]. However, none of the available methods readily translates to the domain of validating error recovery on spacecraft. Some techniques consider both timing and precedence constraints but define timing constraints only in terms of periodic events (e.g. sampling rates) fixed execution times for ....

S. Cheng, J. A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," IEEE Real-Time System Symposium. New Orleans, LA: 1986, pp. 166-174.

....between the application s QoS specifications and the network resources. The V Gammanet provides a framework for flexible support of both realtime and non real time communication requirements. B. Real Time Scheduling Many interesting results have been developed for timeline based schedulers [4, 19, 18] and rate based schedulers. The latter technique has been extensively studied for the case of periodic applications, including both the general case [14] and with more specific cases (priority inversion and other priority based protocols [3, 17, 21] The algorithms that have considered the ....

S. Cheng, J. A. Stankovic, and K. Ramamritham. Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems. In Proc. IEEE Real-Time System Symposium, pages 175-- 180, 1986.

....about how busy they are (observations that are out of date when they are received) Communication between nodes takes non negligible amounts of time. Stankovic [Stankovic, 1985] also has discussed decentralized job scheduling using a Bayesian decision theoretic approach. Work by Cheng, et al. [Cheng et al. 1986] on dynamic, heuristic real time scheduling has more similarities. Like the GPGP DTT approach, Cheng s approach deals with groups of tasks (really executable methods in T MS terminology) that are interrelated. In Cheng s case these are precedence relationships only. All tasks in a group share a ....

S. Cheng, J.A. Stankovic, and K. Ramamritham. Dynamic scheduling of groups of tasks with precedence constraints in distributed hard real-time systems. In Real-time Systems Symposium, December 1986.

....in practical systems, since in real world applications intermittent overloads may occur due to exceptional situations, such as modifications in the environment, arrival of a burst of tasks, or cascades of system failures. A number of heuristic algorithms has been proposed to deal with overloads [2, 3, 4, 6, 7, 10, 11, 13, 15]. They all improve the performance of EDF, however, very few simulation studies have been done to evaluate the importance of the scheduling policy with respect to the guarantee mechanism used to deal with the overload. Baruah et al. 1] have shown that there exists an upper bound on the ....

S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," RealTime Systems Symposium, December 1986.

.... constraints, but their definition of timing constraints only in terms of periodic events (e.g. sampling rates) fixed execution times for events, and deadline or timeliness requirements provides too limited a model for the aperiodic and interval timing constraints on spacecraft commands [8, 35, 36, 40]. The work described here brings together the study of real time constraints with the study of precedence and data consistency constraints. A wide variety of powerful formalisms exists to model the specifications and behavior of real time concurrent systems. Many of these formalisms address to ....

S. Cheng, J. A. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems," IEEE Real-Time System Symposium. New Orleans, LA: 1986, pp. 166-174.

....in real time systems the precedence constraints imposed on tasks are often more general. A heuristic was proposed in the Spring project, where deadline and cost driven algorithms are combined together with rules to dynamically revise values and deadlines in accordance with the precedence relations [6]. A number of heuristic algorithms have also been proposed to deal with overloads [30] 13] which improve the performance of EDF. Baruah, et al. 3] have shown that there exists an upper bound on the performance of any on line (preemptive) algorithm working in overload conditions. The goodness ....

S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems, " Real-Time Systems Symposium, December 1986.

No context found.

S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard RealTime Systems," Proceedings of the IEEE Real-Time Systems Symposium, pp. 166-174, 1986.

No context found.

S. Cheng, J. Stankovic, and K. Ramamritham, "Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Realtime Systems," Proceedings of the IEEE Real-Time Systems Symposium, pages 166-174, 1986.

No context found.

S. Cheng, J.A. Stankovic, and K. Ramamritham, Dynamic Scheduling of Groups of Tasks with Precedence Constraints in Distributed Hard Real-Time Systems, IEEE Real-Time Systems Symposium 1986, pp. 166-174.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC