{"title":"RTEM系统上应用程序任务的开始和完成时间界限","authors":"Dar-Tzen Peng","doi":"10.1109/RTA.1993.263116","DOIUrl":null,"url":null,"abstract":"The author studies the start and completion times of application tasks on RTEM system, a descendant of MAFT developed at AlliedSignal. The RTEM workload consists of a set of periodic tasks, each of which is executed redundantly on the system for fault-tolerance. A priority based nonpreemptive list scheduling algorithm is used for task scheduling. The distributed agreement mechanism used to ensure fault-tolerance adds a nonzero, constant delay to the possible start of the tasks immediately succeeding the task being terminated. Since task execution times may vary between their maximum and minimum values, the start (and thus the completion) time of each copy of a task is difficult to predict unless the task execution order is fixed. He derives both lower and upper bounds of the start time for each task when all tasks are of the same periods, each node executes the entire task set, no branch conditions are allowed, and the task execution order need not be fixed.<<ETX>>","PeriodicalId":293622,"journal":{"name":"[1993] Proceedings of the IEEE Workshop on Real-Time Applications","volume":"40 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"1993-05-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Bounds of start and completion times of application tasks on the RTEM system\",\"authors\":\"Dar-Tzen Peng\",\"doi\":\"10.1109/RTA.1993.263116\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"The author studies the start and completion times of application tasks on RTEM system, a descendant of MAFT developed at AlliedSignal. The RTEM workload consists of a set of periodic tasks, each of which is executed redundantly on the system for fault-tolerance. A priority based nonpreemptive list scheduling algorithm is used for task scheduling. The distributed agreement mechanism used to ensure fault-tolerance adds a nonzero, constant delay to the possible start of the tasks immediately succeeding the task being terminated. Since task execution times may vary between their maximum and minimum values, the start (and thus the completion) time of each copy of a task is difficult to predict unless the task execution order is fixed. He derives both lower and upper bounds of the start time for each task when all tasks are of the same periods, each node executes the entire task set, no branch conditions are allowed, and the task execution order need not be fixed.<<ETX>>\",\"PeriodicalId\":293622,\"journal\":{\"name\":\"[1993] Proceedings of the IEEE Workshop on Real-Time Applications\",\"volume\":\"40 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"1993-05-13\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"[1993] Proceedings of the IEEE Workshop on Real-Time Applications\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1109/RTA.1993.263116\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"[1993] Proceedings of the IEEE Workshop on Real-Time Applications","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1109/RTA.1993.263116","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
Bounds of start and completion times of application tasks on the RTEM system
The author studies the start and completion times of application tasks on RTEM system, a descendant of MAFT developed at AlliedSignal. The RTEM workload consists of a set of periodic tasks, each of which is executed redundantly on the system for fault-tolerance. A priority based nonpreemptive list scheduling algorithm is used for task scheduling. The distributed agreement mechanism used to ensure fault-tolerance adds a nonzero, constant delay to the possible start of the tasks immediately succeeding the task being terminated. Since task execution times may vary between their maximum and minimum values, the start (and thus the completion) time of each copy of a task is difficult to predict unless the task execution order is fixed. He derives both lower and upper bounds of the start time for each task when all tasks are of the same periods, each node executes the entire task set, no branch conditions are allowed, and the task execution order need not be fixed.<>