R. Ha. Validating Timing Constraints in Multiprocessor and Distributed Systems. PhD thesis, University of Illinois at Urbana-Champaign, Aug. 1995. Available as Technical Report UIUCDCS-R-95-1907.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....the energy consumed by a system as a rstclass concept, on par with logical and temporal correctness [8, 12, 17, 20, 22, 27, 28, 29, 31, 32, 33] Multiprocessor real time scheduling. Another topic of growing interest within the real time systems community is multiprocessor scheduling theory [1, 3, 4, 6, 5, 7, 9, 11, 16, 18, 24, 25, 26]. This research has been motivated both by the advent of reasonably priced multiprocessor systems, and also by a growth in computation intensive real time applications that have pushed beyond the capabilities of single processor systems. In addition to o ering more computing power, Supported in ....

Ha, R., and Liu, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. In Proceedings of the 14th IEEE International Conference on Distributed Computing Systems (Los Alamitos, June 1994), IEEE Computer Society Press.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [6, 7, 5], who proposed the following de nition: De nition 1 (Priority driven algorithms [7] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some ....

....certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [6, 7, 5] who proposed the following de nition: De nition 1 (Priority driven algorithms [7]. A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some instant in time, then J i always has higher priority than J j . By this de nition, ....

[Article contains additional citation context not shown here]

Ha, R., and Liu, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. In Proceedings of the 14th IEEE International Conference on Distributed Computing Systems (Los Alamitos, June 1994), IEEE Computer Society Press.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [6, 7, 5], who proposed the following de nition: De nition 1 (Priority driven algorithms [7] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some ....

....which is provably superior to EDF. In Section 6, we summarize the results presented in this paper. 2 Background We brie y describe below some results in multiprocessor real time scheduling theory that will be used in the remainder of this paper. 2. 1 Predictable scheduling algorithms Ha and Liu [6, 7, 5] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Let us de ne a job J j = r j ; e j ; d j ) as being characterized by an arrival time r j , an execution requirement e j , and a deadline d j , with the interpretation ....

[Article contains additional citation context not shown here]

Ha, R., and Liu, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. Tech. Rep. UIUCDCS-R-93-1833, Department of Computer Science, University of Illinois at Urbana-Champaign, October 1993.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [6, 7, 5], who proposed the following de nition: De nition 1 (Priority driven algorithms [7] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some ....

....which is provably superior to EDF. In Section 6, we summarize the results presented in this paper. 2 Background We brie y describe below some results in multiprocessor real time scheduling theory that will be used in the remainder of this paper. 2. 1 Predictable scheduling algorithms Ha and Liu [6, 7, 5] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Let us de ne a job J j = r j ; e j ; d j ) as being characterized by an arrival time r j , an execution requirement e j , and a deadline d j , with the interpretation ....

[Article contains additional citation context not shown here]

Ha, R. Validating timing constraints in multiprocessor and distributed systems. PhD thesis, Department of Computer Science, University of Illinois at Urbana-Champaign, 1995. Available as Technical Report No. UIUCDCS-R-95-1907.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [5, 6, 4], who proposed the following de nition: De nition 3 (Priority driven algorithms [6] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some instant ....

....certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [5, 6, 4] who proposed the following de nition: De nition 3 (Priority driven algorithms [6]. A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some instant in time, then J i always has higher priority than J j . By this de nition, ....

Ha, R., and Liu, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. In Proceedings of the 14th IEEE International Conference on Distributed Computing Systems (Los Alamitos, June 1994), IEEE Computer Society Press.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [5, 6, 4], who proposed the following de nition: De nition 3 (Priority driven algorithms [6] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some instant ....

Ha, R., and Liu, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. Tech. Rep. UIUCDCS-R-93-1833, Department of Computer Science, University of Illinois at Urbana-Champaign, October 1993.

....to individual jobs by the algorithms. Some scheduling algorithms are observed to have certain desirable features in terms of ease (and eciency) of implementation, particularly upon multiprocessor platforms. Some of the important characteristics of such algorithms were studied by Ha and Liu [5, 6, 4], who proposed the following de nition: De nition 3 (Priority driven algorithms [6] A scheduling algorithm is said to be a priority driven scheduling algorithm if and only if it satis es the condition that for every pair of jobs J i and J j , if J i has higher priority than J j at some instant ....

Ha, R. Validating timing constraints in multiprocessor and distributed systems. PhD thesis, Department of Computer Science, University of Illinois at UrbanaChampaign,

....is the case that W (A, m, 2 1 m ) s, I, t) # W (A # ,m,s,I,t) 1) # That is, an m processor work conserving algorithm completes at least as much execution as any other algorithm, if provided processors that are (2 1 m) times as fast. Predictable scheduling algorithms. Ha and Liu [11] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Definition 1 (Predictability) Let A denote a scheduling algorithm, and I = J 1 ,J 2 , J n any set of n jobs, J j = r j ,e j ,d j ) Letf j denote the time at ....

.... staticpriority algorithms (e.g. notice that the earliest deadline first scheduling algorithm, which schedules at each instant the currently active job whose deadline is the smallest, is a priority driven algorithm, but is not a static priority algorithm) The result from the work of Ha and Liu [11] that we will be using can be stated as follows. Theorem 2 (Ha and Liu) Any priority driven scheduling algorithm is predictable. # 3 Algorithm RM US[m (3m 2) We now present Algorithm RM US[m (3m 2) a staticpriority global scheduling algorithm for scheduling periodic task systems, and derive ....

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed real-time systems. In Proc. of the IEEE Int'l Conf. on Distributed Computing Systems, pages 162--171, Poznan, Poland, June 1994.

....deadlines [43,125] In networks, the notion of enforcement or policing is much more common [33,98,128] The work on priority inheritance protocols [8,16,51,86,97,108] addresses some aspects of the coordination problem. The distributed real time scheduling problem is an active area of research [38]. The QOS problem is another active area of research. Some operating system research in this area focuses on best effort approaches [21] although other research emphasizes guarantees [46,53,78,81,126] 1.3 Resource reserves This dissertation defines a resource reservation model that provides an ....

Rhan Ha and J. W. S. Liu. Validating Timing Constraints in Multiprocessor and Distributed Real-Time Systems. In Proceedings of the 14th IEEE International Conference on Distributed Computing Systems, June 1994.

...., it is the case that W (A; m; 2 1 m ) s; I; t) W (A 0 ; m; s; I; t) 1) That is, an m processor work conserving algorithm completes at least as much execution as any other algorithm, if provided processors that are (2 1 m ) times as fast. Predictable scheduling algorithms. Ha and Liu [11, 12, 10] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Definition 1 (Predictability) Let A denote a scheduling algorithm, and I = fJ 1 ; J 2 ; J n g any set of n jobs, J j = r j ; e j ; d j ) Let f j denote the ....

.... static priority algorithms (e.g. notice that the earliest deadline first scheduling algorithm, which schedules at each instant the currently active job whose deadline is the smallest, is a priority driven algorithm, but is not a static priority algorithm) The result from the work of Ha and Liu [11, 12, 10] that we will be using can be stated as follows. Theorem 2 (Ha and Liu) Any priority driven scheduling algorithm is predictable. 3 Algorithm RM US[m (3m 2) In this section, we present Algorithm RM US[m (3m 2) a static priority global scheduling algorithm for scheduling periodic task systems, ....

HA, R., AND LIU, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. In Proceedings of the 14th IEEE International Conference on Distributed Computing Systems (Los Alamitos, June 1994), IEEE Computer Society Press.

...., it is the case that W (A; m; 2 1 m ) s; I; t) W (A 0 ; m; s; I; t) 1) That is, an m processor work conserving algorithm completes at least as much execution as any other algorithm, if provided processors that are (2 1 m ) times as fast. Predictable scheduling algorithms. Ha and Liu [11, 12, 10] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Definition 1 (Predictability) Let A denote a scheduling algorithm, and I = fJ 1 ; J 2 ; J n g any set of n jobs, J j = r j ; e j ; d j ) Let f j denote the ....

.... static priority algorithms (e.g. notice that the earliest deadline first scheduling algorithm, which schedules at each instant the currently active job whose deadline is the smallest, is a priority driven algorithm, but is not a static priority algorithm) The result from the work of Ha and Liu [11, 12, 10] that we will be using can be stated as follows. Theorem 2 (Ha and Liu) Any priority driven scheduling algorithm is predictable. 3 Algorithm RM US[m (3m 2) In this section, we present Algorithm RM US[m (3m 2) a static priority global scheduling algorithm for scheduling periodic task systems, ....

HA, R., AND LIU, J. W. S. Validating timing constraints in multiprocessor and distributed real-time systems. Tech. Rep. UIUCDCS-R-93-1833, Department of Computer Science, University of Illinois at Urbana-Champaign, October 1993.

...., it is the case that W (A; m; 2 1 m ) s; I; t) W (A 0 ; m; s; I; t) 1) That is, an m processor work conserving algorithm completes at least as much execution as any other algorithm, if provided processors that are (2 1 m ) times as fast. Predictable scheduling algorithms. Ha and Liu [11, 12, 10] have studied the issue of predictability in the multiprocessor scheduling of real time systems from the following perspective. Definition 1 (Predictability) Let A denote a scheduling algorithm, and I = fJ 1 ; J 2 ; J n g any set of n jobs, J j = r j ; e j ; d j ) Let f j denote the ....

.... static priority algorithms (e.g. notice that the earliest deadline first scheduling algorithm, which schedules at each instant the currently active job whose deadline is the smallest, is a priority driven algorithm, but is not a static priority algorithm) The result from the work of Ha and Liu [11, 12, 10] that we will be using can be stated as follows. Theorem 2 (Ha and Liu) Any priority driven scheduling algorithm is predictable. 3 Algorithm RM US[m (3m 2) In this section, we present Algorithm RM US[m (3m 2) a static priority global scheduling algorithm for scheduling periodic task systems, ....

HA, R. Validating timing constraints in multiprocessor and distributed systems. PhD thesis, Department of Computer Science, University of Illinois at Urbana-Champaign,

....variable execution times, examining the schedule where each task has its maximum execution time may not be sufficient to draw a conclusion about the schedulability of tasks. The general problem of predicting the completion times of task instances in the presence of anomalies has not been solved [40]. The end to end scheduling problem studied in this thesis resembles the classical job shop scheduling problem. In the job shop model, each task needs to execute on a set of processors in a certain order, and each task may require a different order. A similar but more restricted model, called the ....

R. Ha. Validating Timing Constraints in Multiprocessor and Distributed Systems. PhD thesis, University of Illinois, Urbana-Champaign, Department of Computer Science, 1995.

....algorithms, which make use of the information on job release times. The problem solved by our algorithm is also related to the validation problem, that verifies if all timing constraints are satisfied in a real time system. Both problems deal with a set of jobs with variable execution times. Ha [6] has studied the validation problem in multiprocessor or distributed systems. In her work, a system is predictable if the completion time of a job can be bounded by the completion times of the job in the maximum schedule and minimum schedule, where the maximum(minimum) schedule is obtained by ....

....Bounding the completion times of jobs is a reasonable approach to validating the timing constraints for this kind of systems. Our algorithms provide tighter bounds and, thus, more accurate conclusions on the satisfiability of timing constraints than the general bounds provided by algorithms in [6]. The rest of the paper is organized as follows. Section 2 formally defines the problem addressed and introduces the notations used in the paper. Section 3, 4, and 5 present Algorithm ERT, CJA and ITR, respectively. A simulation was performed to compare the performance of these three algorithms, ....

[Article contains additional citation context not shown here]

R. Ha. Validating Timing Constraints in Multiprocessor and Distributed Systems. PhD thesis, University of Illinois, Urbana-Champaign, Department of Computer Science, 1995.

....in Section VI, the performance of this algorithm is poor compared with the other two algorithms which make use of information on release times. The problem solved by our algorithms is also related to the validation problem. Both problems deal with a set of jobs with variable execution times. Ha [9] has studied the validation problem in multiprocessor or distributed systems. In her work, a system is predictable if the completion time of a job can be bounded by the completion times of the job in the maximum schedule and minimum schedule, where the maximum(minimum) schedule is obtained by ....

....Bounding the completion times of jobs is a reasonable approach to validating the timing constraints for these kinds of systems. Our algorithms provide tighter bounds and, thus, more accurate conclusions on the satisfiability of timing constraints than the general bounds provided by algorithms in [9]. The rest of the paper is organized as follows. Section II formally defines the problem addressed and introduces the notations used in the paper. Sections III, IV, and V present Algorithms ERT, CJA and ITR, respectively. Section VI presents the performance of the three algorithms obtained by ....

[Article contains additional citation context not shown here]

R. Ha, Validating Timing Constraints in Multiprocessor and Distributed Systems, Ph.D. thesis, University of Illinois at Urbana-Champaign, Department of Computer Science, 1995.

No context found.

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed real-time systems. Technical Report UIUCDCS-R-93-1833, University of Illinois at Urbana-Champaign, 1993.

No context found.

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed realtime systems. Technical Report UIUCDCSR -93-1833, University of Illinois at UrbanaChampaign, 1993.

No context found.

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed realtime systems. Technical Report UIUCDCS-R-931833, University of Illinois at Urbana-Champaign, 1993.

No context found.

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed real-time systems. Technical Report UIUCDCS-R-93-1833, University of Illinois at UrbanaChampaign, 1993.

....starting sequences, even though the two observable starting sequences are same. Sometimes in the actual schedule, a job may be preempted by a different job from the one in the maximal schedule, even though all the starting sequences are same. Examples illustrating these facts can be found in [16]. We note from these examples that the completion times of the jobs are in fact accurately predicted by the upper bounds given by Theorem 9.5.3. These examples cause us to ask whether the bounds in Theorem 9.5.3 are tight in some sense when there is preemption in the maximal schedule. This ....

....1 Processor 2 (a) b) c) Figure 9.4 An Example Illustrating Algorithm INNF The following theorem allows us to conclude that if H i can complete by the deadline d i of J i in the schedule generated by algorithm INNF , then J i always completes by d i . The proof of the theorem can be found in [16]. Theorem 9.6.1 The completion time F (J i ) of J i is no later than the completion time of the transformed job H i in the schedule of G i , L i Gamma1 , and H i generated by algorithm INNF . Because lower priority preemptable jobs do not block any job, we can use Algorithm INNF with very little ....

[Article contains additional citation context not shown here]

R. Ha and J. W. S. Liu. Validating timing constraints in multiprocessor and distributed real-time systems. Technical Report UIUCDCS-R-93-1833, University

No context found.

R. Ha. Validating Timing Constraints in Multiprocessor and Distributed Systems. PhD thesis, University of Illinois at Urbana-Champaign, Aug. 1995. Available as Technical Report UIUCDCS-R-95-1907.

No context found.

R. Ha and J. W.-S. Liu, "Validating timing constraints in multiprocessor and distributed real-time systems," Proc. of the IEEE Int'l Conf. on Distributed Computing Systems, Poznan, Poland, June 21--24, 1994, pp. 162--171.

No context found.

R. Ha, J.W.S. Liu, \Validating timing constraints in multiprocessor and distributed real-time systems", technical report UIUCDCS-R-93-1833 , Department of Computer Science, UniversityofIllinois at UrbanaChampaign, (October 1993.

No context found.

R. Ha, \Validating timing constraints in multiprocessor and distributed systems", Ph.D. thesis, technical report UIUCDCS-R-95-1907, Department of Computer Science, University of Illinois at Urbana-Champaign (1995).

No context found.

R. Ha and J. W.-S. Liu. Validating timing constraints in multiprocessor and distributed real-time systems. In Proc. of the IEEE Int'l Conf. on Distributed Computing Systems, pages 162--171, Poznan, Poland, June 21--24, 1994.

First 50 documents

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