In this paper, we proposed a scheduling framework considering the response time of aperiodic jobs in mixed task sets on multiprocessor systems. The periodic tasks will be assigned to fixed processors under different task assignment heuristic before run time. We compared the performance of three different task assignment heuristics, i.e., Worst Fit, Best Fit, and First Fit. The unused utilizations let by periodic tasks on processors will utilized to be the sizes of TB Servers. The TB Servers on each processor will share evenly the left capacity. The TB Servers determine the virtual deadline of each arrival aperiodic and influence its response time. The experiment results show that under different aperiodic workload, the BF heuristic has better average response time than the WF and FF heuristics. The BF heuristic makes processors load balancing. The WF and FF heuristics make larger workload difference among processors and cause that there could have tiny-size TB Servers. When the aperiodic workload was very light and the processor number was larger than four, the WF heuristic may make that on some processors there are no assigned tasks and the left capacities for TB Servers are equal to 1. Then, the WF heuristic had the performance similar to that of the BF heuristic.
When the aperiodic workload increases, TB Servers with larger sizes may be too busy such that more aperiodic jobs must be assigned to tiny-size TB Servers. The jobs assigned tiny-size TB Servers will have larger deadlines and then have worse response time. For the future work, to improve the average response time we want to consider that how to reduce the effect caused by tiny-size TB Servers under the WF heuristic. Besides, we want to improve response time of aperiodic jobs if job migration is permitted at run time. Finally, we also consider the resource
competition among tasks on multiprocessor systems.
Bibliography
[1] B. Andersson and J. Jonsson S. Baruah. Static-priority scheduling on multiprocessors.
In IEEE Real-Time Systems Symposium, pages 193–202, 2001.
[2] H. Aydin and Q. Yang. Energy-aware partitioning for multiprocessor real-time systems.
In International Parallel and Distributed Processing Symposium, 2003.
[3] T. P. Baker and S. K. Baruah. Schedulability analysis of multiprocessor sporadic task systems. In Handbook of Realtime and Embedded Systems, 2007.
[4] S. Baruah, J. Goossens, and G. Lipari. Implementing constant-bandwidth servers upon multiprocessor platforms. In Real-Time and Embedded Technology and Applications Symposium, pages 154–163, 2002.
[5] S. Baruah and G. Lipari. A multiprocessor implementation of the total bandwidth server.
In International Parallel and Distributed Processing Symposium, 2004.
[6] S. K. Baruah. Partitioning real-time tasks among heterogeneous multiprocessors. In International Conference on Parallel Processing, pages 467–474, 2004.
[7] A. Bastoni, B. B. Brandenburg, and J. H. Anderson. An empirical comparison of global, partitioned, and clustered multiprocessor edf schedulers. In IEEE Real-Time Systems Symposium, pages 14–24, 2010.
[8] E. G. Coffman, M. R. Garey, and D. S. Johnson. Approximation algorithms for bin packing: a survey. In Approximation Algorithms for NP-Hard Problems, PWS, Boston, 1997.
[9] S. Funk and S. Baruah. Task assignment on uniform heterogeneous multiprocessors. In Euromicro Conference on Real-Time Systems, pages 219–226, 2005.
[10] L.A. Gutierrez, C.A. Franco, R.R. Jacinto, and C.A. Gutierrez. Minimizing the response times of aperiodic tasks in hard real-time systems with edf. In Electronics, Robotics, and Automotive Mechanics Conference, pages 268–273, 2006.
[11] M. Karic, G. Martinovic, and Z. Hocenski. Implementation and simulation results of multiple total bandwidth server mechanism. In International Symposium on Intelligent Systems and Informatics, pages 423–427, 2010.
[12] S. Kato and N. Yamasaki. Portioned static-priority scheduling on multiprocessors. In IEEE International Symposium on Parallel and Distributed Processing, pages 1–12, 2008.
[13] S. Kato and N. Yamasaki. Scheduling aperiodic tasks using total bandwidth server on multiprocessors. In IEEE/IFIP International Conference on Embedded and Ubiquitous Computing, pages 82–89, 2008.
[14] H. Kim, M. Park, S. Cho, H.J. Yoo, J. Chae, and Y. Cho. A combined approach to improve the response time of soft aperiodic tasks. In Software Technologies for Future Dependable Distributed Systems, pages 215–219, 2009.
[15] S. Lauzac, R. Melhem, and D. Mosse. Comparison of global and partitioning schemes for scheduling rate monotonic tasks on a multiprocessor. In IEEE Real-Time Systems Symposium, pages 188–195, 1998.
[16] J. P. Lehoczky and S. Ramos-Thuel. An optimal algorithm for scheduling soft-aperiodic tasks in fixed-priority preemptive systems. In IEEE Real-Time Systems Symposium, 1992.
[17] J. P. Lehoczky, L. Sha, and J. K. Strosnider. Enhanced aperiodic responsiveness in hard real-time environments. In IEEE Real-Time Systems Symposium, pages 261–270, 1987.
[18] T.H. Lin and W. Tarng. Scheduling periodic and aperiodic tasks in hard real-time com-puting systems. In ACM SIGMETRICS, pages 31–38, 1991.
[19] C. L. Liu and J. W. Layland. Scheduling algorithms for multiprogramming in a hard-real-time environment. Journal of the ACM, 20(1):46–61, January 1973.
[20] J. Lopez, J. Diaz, M. Garcia, and D. Garcia. Worst-case utilization bound for edf schedul-ing on real-time multiprocessor systems. In the 12th Euromicro Workshop on Real-Time Systems, 2000.
[21] B. Sprunt, L. Sha, and J. P. Lehoczky. Aperiodic task scheduling for hard-real time systems. Journal of Real-Time Systems, 1(1):27–60, 1989.
[22] M. Spuri and G. Buttazzo. Scheduling aperiodic tasks in dynamic priority systems. Real-Time Systems, 10(2):179–210, 1996.
[23] C.Y. Yang, J.J. Chen, and T.W. Kuo. An approximation algorithm for energy-efficient scheduling on a chip multiprocessor. In DATE 2005, pages 468–473, 2005.
Appendix A:
Table 5.1: Notations and Abbreviations Symbol/Term Description
ck Worst case execution time ofJk
di Absolutely deadline ofτi
dj,x,y Virtual deadline of theythjob on TB ServerT Bj,x
fk Completion time ofJk
rk Arrival time ofJk
τi Periodic taskτi
usj,x TB Server size ofT Bj,x
Aj Periodic task set assigned toPj after task assignment Ci Worst case execution time ofτi
Di Relative deadline ofτi
Jk Aperiodic jobJk
Ri Phase time ofτi
Sj TB Server set of processorPj
Sj,x Thexth TB Server inSj
Ti Period ofτi
T Bj,x Thexth TB Server on processorPj
Up,j Total utilization of periodic tasks on processorPj