| T.-S. Tia and J. W.-S. Liu, "Assigning real-time tasks and resources to distributed systems," Internnational Journal of Minim and Microcomputers, vol. 17, no. 1, pp. 18--25, 1995. |
....allocation since the problem is NP hard. To overcome the scalability limitation, heuristic approaches [11 16] have been proposed for larger instances of the problem. Based on their performance measures, these approaches can be classified as schedulability based [12 14] or communication based [15, 16]. One common way to reduce the allocation search space is to cluster tasks into larger units of allocation, then allocate the resulting task clusters, not individual tasks, to available processors. Different flavors of this are proposed in [9, 15, 16] For the special case of fixed priority ....
....[12 14] or communication based [15, 16] One common way to reduce the allocation search space is to cluster tasks into larger units of allocation, then allocate the resulting task clusters, not individual tasks, to available processors. Different flavors of this are proposed in [9, 15, 16]. For the special case of fixed priority scheduling on homogeneous multiprocessor systems, rate monotonic analysis has been used to derive near optimal task clustering policies taking into consideration task periods [17, 18] In general, clustering heuristics, such as those in [9, 15] typically ....
[Article contains additional citation context not shown here]
T.-S. Tia and J. W.-S. Liu, "Assigning real-time tasks and resources to distributed systems," Internnational Journal of Minim and Microcomputers, vol. 17, no. 1, pp. 18--25, 1995.
....given task graph is invoked periodically under an end to end deadline. A task PENG et al. ASSIGNMENT AND SCHEDULING COMMUNICATING PERIODIC TASKS IN DISTRIBUTED REAL TIME SYSTEMS 747 allocation and message schedule are computed such that the end to end deadline is satisfied for each invocation. In [45], 46] 47] efficient methods are considered for allocating periodic tasks where different tasks may have different deadlines. To reduce the size of the problem, tasks are preclustered before allocation begins. Task clusters (instead of individual tasks) are then assigned to individual ....
....deadlines. To reduce the size of the problem, tasks are preclustered before allocation begins. Task clusters (instead of individual tasks) are then assigned to individual processors. Graph based heuristics, which attempt to minimize interprocessor communication, are used for task assignment in [45], 46] In contrast, the approach in [47] searches the space of all (cluster) assignments and schedules until either a feasible solution is found or the space of all solutions have been searched exhaustively. The search is directed by a heuristic that explores the most likely parts of the ....
T.-S. Tia and J.W.-S. Liu, "Assigning Real-Time Tasks and Resources to Distributed Systems," Int'l J. Minim and Microcomputers, vol. 17, no. 1, pp. 18-25, 1995.
....characteristics of real time applications. For example, 123] considers an abstract problem where a given task graph is invoked periodically under an end to end deadline. A task allocation and message schedule are computed such that the end to end deadline is satisfied for each invocation. In [105, 135, 145] efficient methods are considered for allocating periodic tasks where different tasks may have different deadlines. Graph based heuristics, which attempt to minimize interprocessor communication, are used for task assignment in [135, 145] Analytic models for load sharing have been developed in ....
....deadline is satisfied for each invocation. In [105, 135, 145] efficient methods are considered for allocating periodic tasks where different tasks may have different deadlines. Graph based heuristics, which attempt to minimize interprocessor communication, are used for task assignment in [135, 145]. Analytic models for load sharing have been developed in [121] Off line schedulability analysis [6, 119, 146, 147] was used to verify that the reserved resources are sufficient to meet al..l timing constraints. Optimal solutions for task allocation problems in hard real time systems have been ....
T.-S. Tia and J. W.-S. Liu, "Assigning real-time tasks and resources to distributed systems," Internnational Journal of Minim and Microcomputers, vol. 17, no. 1, pp. 18--25, 1995.
....and schedulability oriented methods. For example, 44] considers an abstract problem where a given task graph is invoked periodically under an end to end deadline. A task allocation and message schedule are computed such that the end to end deadline is satisfied for each invocation. In [45 47] efficient methods are considered for allocating periodic tasks where different tasks may have different deadlines. To reduce the size of the problem, tasks are pre clustered before allocation begins. Task clusters (instead of individual tasks) are then assigned to individual processors. ....
....deadlines. To reduce the size of the problem, tasks are pre clustered before allocation begins. Task clusters (instead of individual tasks) are then assigned to individual processors. Graph based heuristics, which attempt to minimize interprocessor communication, are used for task assignment in [45, 46]. In contrast, the approach in [47] searches the space of all (cluster) assignments and schedules until either a feasible solution is found or the space of all solutions have been searched exhaustively. The search is directed by a heuristic that explores the most likely parts of the search space ....
T.-S. Tia and J. W.-S. Liu, "Assigning real-time tasks and resources to distributed systems," Internnational Journal of Minim and Microcomputers, vol. 17, no. 1, pp. 18--25, 1995.
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