skip to main content
article
Free Access

Experience with trace driven modeling

Published:01 July 1976Publication History
Skip Abstract Section

Abstract

Trace-driven modeling is a technique for optimizing a computer system by simulation studies, in which the job load data required by the simulator is taken directly from the existing system. Its advantages are 1) a full statistical representation of the job load, 2) potentially more accurate, system-independent and relevant data, and 3) the possibility of wider-ranging simulation studies without the necessity of collecting more particular data.

The variations of the trace-driven modeling technique that have occurred in practice are summarized. The above advantages of the technique and their limitations are then analyzed in the context of experience with it for the evaluation of a class of scheduling algorithms for an interactive system on the CDC 6400. Results of simulations showing the effects of interdependence of resource demand parameters are presented.

References

  1. 1 Cheng, P. S., "Trace-Driven System Modeling," IBM Systems Journal, Vol. 8 No. 4, (1969).Google ScholarGoogle ScholarDigital LibraryDigital Library
  2. 2 Estrin, G., Muntz, R. R. and Uzgalis, R. C., "Modelling, Measurement and Computer Power," Proceedings, Spring Joint Computer Conference, 1972.Google ScholarGoogle Scholar
  3. 3 Herring, L. A., A Trace Driven Simulation Study of Predictive Scheduling Algorithms. Department of Computer Sciences (Masters Thesis). University of Texas, August, 1976.Google ScholarGoogle Scholar
  4. 4 Howard, J. H., "A Large-Scale Dual Operating System," Proceedings ACM National Conference, 1973, 242-248. Google ScholarGoogle ScholarDigital LibraryDigital Library
  5. 5 Johnson, D. S., A Process-Oriented Model of Resource Demands in Large, Multiprocessing Computer Utilities, Computation Center and Department of Computer Sciences TSN-29, (Ph.D. Dissertation), The University of Texas at Austin, 1972. Google ScholarGoogle ScholarDigital LibraryDigital Library
  6. 6 Noe, J. D. and G. J. Nutt, "Validation of a Trace-Driven CDC 6400 Simulation," Proceedings Spring Joint Computer Conference (1972), 749-757.Google ScholarGoogle Scholar
  7. 7 Noetzel, A. S., "The Design of a Meta-System," Proceedings Spring Joint Computer Conference (1971), 415-424.Google ScholarGoogle Scholar
  8. 8 Noetzel, A. S., Simulation Studies of Predictive Scheduling, Technical Report No. 37, Department of Computer Sciences, The University of Texas, 1974. Google ScholarGoogle ScholarDigital LibraryDigital Library
  9. 9 Noetzel, A. S. and Haley, J. R., The Decomposition of a Multiprocessor Event Trace into User Program Resource Demand Patterns, Technical Report, No. 49, Department of Computer Sciences, The University of Texas, 1975. Google ScholarGoogle ScholarDigital LibraryDigital Library
  10. 10 Sherman, S. W., Trace-Driven Modeling Studies of the Performance of Computer Systems, Computation Center and Department of Computer Sciences, TSN-30, (Ph.D. Dissertation), The University of Texas at Austin, August, 1972. Google ScholarGoogle ScholarDigital LibraryDigital Library
  11. 11 Sherman, S. W., J. C. Browne and F. Baskett, "Trace Driven Modeling and Analysis of CPU Scheduling in a Multi-programming System," Proceedings of ACM Workshop on Performance Evaluation, Cambridge, Massachusetts, April, 1971, 173-199. Google ScholarGoogle ScholarDigital LibraryDigital Library
  12. 12 Schwetman, H. D. A Study of Resource Utilization and Performance Evaluation of Large-scale Computer Systems, Computation Center TSN-12, (Ph.D. Dissertation), The University of Texas at Austin, August, 1970. Google ScholarGoogle ScholarDigital LibraryDigital Library
  13. 13 Thornton, J. E., Design of a Computer: The CDC 6600, Scott, Foresman and Co., Glenview, Illinois, 1970. Google ScholarGoogle ScholarDigital LibraryDigital Library

Index Terms

  1. Experience with trace driven modeling

      Recommendations

      Comments

      Login options

      Check if you have access through your login credentials or your institution to get full access on this article.

      Sign in

      Full Access

      PDF Format

      View or Download as a PDF file.

      PDF

      eReader

      View online with eReader.

      eReader