On the construction of a representative synthetic workload

ABSTRACT

A general method of constructing a drive workload representative of a real workload is described. The real workload is characterized by its demands on the various system resources. These characteristics of the real workload are obtained from the system accounting data. The characteristics of the drive workload are determined by matching the joint probability density of the real workload with that of the drive workload. The drive workload is realized by using a synthetic program in which the characteristics can be varied by varying the appropriate parameters. Calibration experiments are conducted to determine expressions relating the synthetic program parameters with the workload characteristics. The general method is applied to the case of two variables, cpu seconds and number of I/O activities; and a synthetic workload with 88 jobs is constructed to represent a month's workload consisting of about 6000 jobs.

REFERENCES

• 1: Henry Lucas, Jr., Performance Evaluation and Monitoring, ACM Computing Surveys (CSUR), v.3 n.3, p.79-91, Sept. 1971 [doi>10.1145/356589.356590]
• 2: R. F. Hitti , E. O. Joslin, Evaluation and performance of computers: application benchmarks: the key to meaningful computer evaluations, Proceedings of the 1965 20th national conference, p.27-37, August 24-26, 1965, Cleveland, Ohio, United States [doi>10.1145/800197.806031]
• 3: Shope, W.L., Kashmark, K.L., Inghram, J.W., and Decker, W.F. System performance study. Proc. SHARE XXXIV, Vol. I, Colorado, Mar. 1970, pp. 439-530.
• 4: Wood, D.C., and Forman, E.H. Throughout measurement using a synthetic job stream. Proc. AFIPS 1971 FJCC, Vol. 39, AFIPS Press, Montvale, N.J., pp. 51-56.
• 5: Buchholz, W. A synthetic job for measuring system performance. IBM Syst. J. 8, 4 (1969), 309-318.
• 6: Henry C. Lucas, Jr., Synthetic program specifications for performance evaluation, Proceedings of the ACM annual conference, p.1041-1058, August 01-01, 1972, Boston, Massachusetts, United States [doi>10.1145/800194.805897]
• 7: Gamse, R.N. Approach plan for a standard benchmark library for use in computer system-selection. MITRE Tech. Rep,, MTR-2226, Sept. 1971.
• 8: Ferrari, D. Workload characterization and selection in computer performance measurement. Computer (July/Aug. 1972), 18-24.
• 9: B. W. Kernighan , P. A. Hamilton, Synthetically generated performance test loads for operating systems, Proceeding of the 1973 ACM SIGME symposium, p.121-126, January 1973 [doi>10.1145/800268.809343]
• 10: D. E. Morgan , J. A. Campbell, An answer to a user's plea?, Proceeding of the 1973 ACM SIGME symposium, p.112-120, January 1973 [doi>10.1145/800268.809342]
• 11: Hunt, E., Hiehr, G,, and Garnatz, D. Who are users?--An analysis of computer use in a university computer center. Proc. AFIPS 1971 SJCC, Vol. 38, AFIPS Press, Montvale, N.J., pp. 231-238.
• 12: FORTRAN version of Buchholz's synthetic program. Transl. by J. Maranzano and modified by T.E. Bell. Proc. 3rd Meeting of the Computer Performance Evaluation Users Committee, July 1971, Washington, D.C.
• 13: IBM System/370 Model 155--functional characteristics, pp. 40-44.

INDEX TERMS

Primary Classification:
  K. Computing Milieux
  K.6 MANAGEMENT OF COMPUTING AND INFORMATION SYSTEMS
      K.6.2 Installation Management
          Subjects: Benchmarks

Additional Classification:
  C. Computer Systems Organization

  I. Computing Methodologies
  I.6 SIMULATION AND MODELING

General Terms:
Experimentation, Standardization

Keywords:
benchmarks, calibration, drive workload, input to simulation, probability distribution, representative workload, synthetic workload, workload characteristics