The Experts below are selected from a list of 408 Experts worldwide ranked by ideXlab platform
Daniel Mosse - One of the best experts on this subject based on the ideXlab platform.
-
IPPS - Fault-tolerant scheduling on a hard real-time multiprocessor system
Proceedings of 8th International Parallel Processing Symposium, 1Co-Authors: Sunondo Ghosh, Rami Melhem, Daniel MosseAbstract:Fault-tolerance is an important issue in hard real-time systems due to the critical nature of the supported tasks. One way of providing fault-tolerance is to schedule multiple copies of a task on different processors. If the primary copy of a task cannot be completed due to a fault, the Scheduled Backup copy is run and the task is completed. In this paper, we propose a new algorithm for fault-tolerant scheduling on multiprocessor systems. The algorithm guarantees the completion of a Scheduled task before its deadline in the presence of processor failures. Our algorithm schedules several Backup tasks overlapping one another and dynamically deallocates the Backups as soon as the original tasks complete executions, thus increasing the utilization of processors. Simulation results show that our method achieves higher task schedulability compared to using a spare processor as a Backup to be invoked in the event of a failure. Further, we show that the cost, in terms of schedulability, of guaranteeing fault tolerance for dynamic systems is quite low. >
Sunondo Ghosh - One of the best experts on this subject based on the ideXlab platform.
-
Fault-Tolerant Scheduling on a Hard Real-Time Multiprocessor System
1994Co-Authors: Sunondo Ghosh, Rami MelhemAbstract:Fault-tolerance is an important issue in hard realtime systems due to the critical nature of the supported tasks. One way of providing fault-tolerance is to schedule multiple copies of a task on different processors. If the primary copy of a task cannot be completed due to a fault, the Scheduled Backup copy is run and the task is completed. In this paper, we propose a new algorithm for fault-tolerant scheduling on multiprocessor systems. The algorithm guarantees the completion of a Scheduled task before its deadline in the presence of processor failures. Our algorithm schedules several Backup tasks overlapping one another and dynamically deallocates the Backups as soon as the original tasks complete executions, thus increasing the utilization of processors. Simulation results show that our method achieves higher task schedulability compared to using a spare processor as a Backup to be invoked in the event of a failure. Further, we show that the cost, in terms of schedulability, of guar..
-
IPPS - Fault-tolerant scheduling on a hard real-time multiprocessor system
Proceedings of 8th International Parallel Processing Symposium, 1Co-Authors: Sunondo Ghosh, Rami Melhem, Daniel MosseAbstract:Fault-tolerance is an important issue in hard real-time systems due to the critical nature of the supported tasks. One way of providing fault-tolerance is to schedule multiple copies of a task on different processors. If the primary copy of a task cannot be completed due to a fault, the Scheduled Backup copy is run and the task is completed. In this paper, we propose a new algorithm for fault-tolerant scheduling on multiprocessor systems. The algorithm guarantees the completion of a Scheduled task before its deadline in the presence of processor failures. Our algorithm schedules several Backup tasks overlapping one another and dynamically deallocates the Backups as soon as the original tasks complete executions, thus increasing the utilization of processors. Simulation results show that our method achieves higher task schedulability compared to using a spare processor as a Backup to be invoked in the event of a failure. Further, we show that the cost, in terms of schedulability, of guaranteeing fault tolerance for dynamic systems is quite low. >
Rami Melhem - One of the best experts on this subject based on the ideXlab platform.
-
Fault-Tolerant Scheduling on a Hard Real-Time Multiprocessor System
1994Co-Authors: Sunondo Ghosh, Rami MelhemAbstract:Fault-tolerance is an important issue in hard realtime systems due to the critical nature of the supported tasks. One way of providing fault-tolerance is to schedule multiple copies of a task on different processors. If the primary copy of a task cannot be completed due to a fault, the Scheduled Backup copy is run and the task is completed. In this paper, we propose a new algorithm for fault-tolerant scheduling on multiprocessor systems. The algorithm guarantees the completion of a Scheduled task before its deadline in the presence of processor failures. Our algorithm schedules several Backup tasks overlapping one another and dynamically deallocates the Backups as soon as the original tasks complete executions, thus increasing the utilization of processors. Simulation results show that our method achieves higher task schedulability compared to using a spare processor as a Backup to be invoked in the event of a failure. Further, we show that the cost, in terms of schedulability, of guar..
-
IPPS - Fault-tolerant scheduling on a hard real-time multiprocessor system
Proceedings of 8th International Parallel Processing Symposium, 1Co-Authors: Sunondo Ghosh, Rami Melhem, Daniel MosseAbstract:Fault-tolerance is an important issue in hard real-time systems due to the critical nature of the supported tasks. One way of providing fault-tolerance is to schedule multiple copies of a task on different processors. If the primary copy of a task cannot be completed due to a fault, the Scheduled Backup copy is run and the task is completed. In this paper, we propose a new algorithm for fault-tolerant scheduling on multiprocessor systems. The algorithm guarantees the completion of a Scheduled task before its deadline in the presence of processor failures. Our algorithm schedules several Backup tasks overlapping one another and dynamically deallocates the Backups as soon as the original tasks complete executions, thus increasing the utilization of processors. Simulation results show that our method achieves higher task schedulability compared to using a spare processor as a Backup to be invoked in the event of a failure. Further, we show that the cost, in terms of schedulability, of guaranteeing fault tolerance for dynamic systems is quite low. >
Shaohong Fang - One of the best experts on this subject based on the ideXlab platform.
-
DISTRIBUTED REAL-TIME FAULT-TOLERANCE
2005Co-Authors: Albert M. K. Cheng, Shaohong FangAbstract:The goal of this project is to study and simulate a distributed real-time fault-tolerance web monitoring system. The method of providing fault-tolerance is to schedule multiple copies of a task on different computer nodes in a distributed computing system. A fault-tolerant system automatically recovers from a specified number of failures. If the primary task cannot be completed due to a fault, the Scheduled Backup task is run and all tasks are assured to complete. We use the web to monitor the fault-tolerant behavior of a distributed system. A web monitoring system is a convenient way to monitor remote tasks, both primary and Backup. Our simulation results show that it is possible to set up a distributed real-time fault-tolerance web monitoring system. To achieve our goal quickly, we use the existing the Ganglia network and the RRDTool technology. We found that we can use the Ganglia system to set up a distributed real-time fault-tolerance web monitoring system. The RRDTool is only for simulation purposes. It is used to show the results of the simulation graphically, but it can’t accurately store the status data of the tasks. We will use the MySQL to store the status of the real time tasks instead of the RRDTool in further research and to yield more accurate results within the time constraints
Albert M. K. Cheng - One of the best experts on this subject based on the ideXlab platform.
-
DISTRIBUTED REAL-TIME FAULT-TOLERANCE
2005Co-Authors: Albert M. K. Cheng, Shaohong FangAbstract:The goal of this project is to study and simulate a distributed real-time fault-tolerance web monitoring system. The method of providing fault-tolerance is to schedule multiple copies of a task on different computer nodes in a distributed computing system. A fault-tolerant system automatically recovers from a specified number of failures. If the primary task cannot be completed due to a fault, the Scheduled Backup task is run and all tasks are assured to complete. We use the web to monitor the fault-tolerant behavior of a distributed system. A web monitoring system is a convenient way to monitor remote tasks, both primary and Backup. Our simulation results show that it is possible to set up a distributed real-time fault-tolerance web monitoring system. To achieve our goal quickly, we use the existing the Ganglia network and the RRDTool technology. We found that we can use the Ganglia system to set up a distributed real-time fault-tolerance web monitoring system. The RRDTool is only for simulation purposes. It is used to show the results of the simulation graphically, but it can’t accurately store the status data of the tasks. We will use the MySQL to store the status of the real time tasks instead of the RRDTool in further research and to yield more accurate results within the time constraints
