End-to-end optimization in heterogeneous distributed real-time systems
In this paper we address an end-to-end optimization problem in a distributed real-time system when a set of pipelined task chains are given. End-to-end deadlines and end-to-end jitter constraints are assumed to be given for task chains, in addition to an objective function to be optimized throughout the optimization process. The objective of the optimization process is to obtain local deadlines and other system parameters that not only satisfy all the given end-to-end constraints but also minimize a given objective function.
A separable programming technique is used to solve the resulting nonlinear programming problems. If an objective function and constraints satisfy a certain condition, it is shown that those problems can be solved by using a linear programming technique which already has well-established theories and results. Also, it is shown that the condition is general enough that a wide class of optimization problems can be solved in designing distributed real-time systems by using this technique.
Unable to display preview. Download preview PDF.
- 1.T. P. Baker. A Stack-Based Resource Allocation Policy for RealTime Processes. In Proceedings, IEEE Real-Time Systems Symposium, 1990.Google Scholar
- 2.M. Bazaraa. Nonlinear Programming: Theory and Algorithms, chapter 11. Separable Programming, pages 453–471. John Wiley & Sons, 1979.Google Scholar
- 3.R. Bettati and J. Liu. End-to-end scheduling to meet deadlines in distributed systems. In Proceedings, IEEE International Conference on Distributed Computing Systems, 1992.Google Scholar
- 4.S. Chatterjee and J. Strosnider. Distributed Pipeline Scheduling. The Computer Journal, 38.Google Scholar
- 6.R. Gerber, S. Hong, and M. Saksena. Guaranteeing End-to-End Timing Constraints by Calibrating Intermediate Processes. In Proceedings IEEE Real-Time Systems Symposium, 1994. Also available as University of Maryland CS-TR-3274, UMIACS-TR-94-58.Google Scholar
- 7.J. Huang and D. Du. Resource management for continuous mutimedia database applications. In Proceedings, IEEE Real-Time Systems Symposium, Dec. 1994.Google Scholar
- 8.K. Jeffay. The real-time producer/consumer paradigm. In ACM/SIGAPP Symposium on Applied Computing, Feb. 1993.Google Scholar
- 9.D. Kang, R. Gerber, and M. Saksena. Performance-based design of distributed real-time systems. In Proceedings, IEEE Real-Time Technology and Applications Symposium, June 1997.Google Scholar
- 10.D. Luenberger. Linear and Nonlinear Programming, chapter The Simplex Method, pages 30–84. Addison-Wesley, 1984.Google Scholar
- 11.N. Malcolm and W. Zhao. The timed-token protocol for real-time communication. In IEEE Computer, Jan. 1994.Google Scholar
- 12.S. Mukherjee, D. Saha, M. Saksena, and S. Tripathi. A bandwidth allocation scheme for time constrained message transmission on a slotted ring Ian. In Proceedings, IEEE Real-Time Systems Symposium, pages 44–53, Dec. 1993.Google Scholar
- 13.K. Murty. Linear and Combinatorial Programming, chapter Appendix 1, pages 541–552. John Wiley & Sons, 1976.Google Scholar
- 16.L. Sha and S. Sathaye. A systematic approach to designing distributed real-time systems. In IEEE Computer, Sep. 1993.Google Scholar
- 17.L. Sha, S. Sathaye, and J. Strosnider. Scheduling real-time communication on dual link networks. In Proceedings, IEEE Real-Time Systems Symposium, pages 188–197, Dec. 1992.Google Scholar
- 18.J. Sun, J. Liu, and R. Bettati. And End-to-End Approach to Scheduling Periodic Tasks with Shared Resources in Multiprocessor Systems. Technical report, University of Illinois at Urbana-Champaign, Department of Computer Science, 1994.Google Scholar
- 19.D. Verma, Hui Zhang, and D. Ferrari. Guaranteeing delay jitter bounds in packet switching networks. In Proceedings of Tricomm’91, April 1991.Google Scholar