Queueing Systems

, Volume 64, Issue 1, pp 5–48 | Cite as

On the inapproximability of M/G/K: why two moments of job size distribution are not enough

  • Varun GuptaEmail author
  • Mor Harchol-Balter
  • J. G. Dai
  • Bert Zwart


The M/G/K queueing system is one of the oldest models for multiserver systems and has been the topic of performance papers for almost half a century. However, even now, only coarse approximations exist for its mean waiting time. All the closed-form (nonnumerical) approximations in the literature are based on (at most) the first two moments of the job size distribution. In this paper we prove that no approximation based on only the first two moments can be accurate for all job size distributions, and we provide a lower bound on the inapproximability ratio, which we refer to as “the gap.” This is the first such result in the literature to address “the gap.” The proof technique behind this result is novel as well and combines mean value analysis, sample path techniques, scheduling, regenerative arguments, and asymptotic estimates. Finally, our work provides insight into the effect of higher moments of the job size distribution on the mean waiting time.


Multi-server systems M/G/K Two-moment approximation Inapproximability Higher moments 

Mathematics Subject Classification (2000)

60K25 68M20 90B22 


Unable to display preview. Download preview PDF.

Unable to display preview. Download preview PDF.


  1. 1.
    Adan, I.J.B.F., Resing, J.: Queueing Theory. Eindhoven University of Technology (2002) Google Scholar
  2. 2.
    Barford, P., Crovella, M.: Generating representative web workloads for network and server performance evaluation. In: Proceedings of ACM SIGMETRICS/Performance’98, pp. 151–160 (1998) Google Scholar
  3. 3.
    Bertsimas, D., Natarajan, K.: A semidefinite optimization approach to the steady-state analysis of queueing systems. Queueing Syst. 56(1), 27–39 (2007) CrossRefGoogle Scholar
  4. 4.
    Borovkov, A.A.: Stochastic Processes in Queueing Theory. Nauka, Moscow (1972) Google Scholar
  5. 5.
    Burman, D.Y., Smith, D.R.: A light-traffic theorem for multi-server queues. Math. Oper. Res. 8, 15–25 (1983) CrossRefGoogle Scholar
  6. 6.
    Cosmetatos, G.P.: Some approximate equilibrium results for the multiserver queue (M/G/r). Oper. Res. Q. 27, 615–620 (1976) CrossRefGoogle Scholar
  7. 7.
    Daley, D.J.: Some results for the mean waiting-time and workload in GI/GI/k queues. In: Dshalalow, J.H. (ed.) Frontiers in Queueing: Models and Applications in Science and Engineering, pp. 35–59. CRC Press, Boca Raton (1997) Google Scholar
  8. 8.
    Daley, D.J., Rolski, T.: Some comparability results for waiting times in single- and many-server queues. J. Appl. Probab. 21, 887–900 (1984) CrossRefGoogle Scholar
  9. 9.
    de Smit, J.H.A.: A numerical solution for the multiserver queue with hyper-exponential service times. Oper. Res. Lett. 2(5), 217–224 (1983) CrossRefGoogle Scholar
  10. 10.
    de Smit, J.H.A.: The queue GI/M/s with customers of different types or the queue GI/H m/s. Adv. Appl. Probab. 15(2), 392–419 (1983) CrossRefGoogle Scholar
  11. 11.
    de Smit, J.H.A.: The queue GI/H m/s in continuous time. J. Appl. Probab. 22(1), 214–222 (1985) CrossRefGoogle Scholar
  12. 12.
    Downy, A., Harchol-Balter, M.: Exploiting process lifetime distributions for dynamic load balancing. ACM Trans. Comput. Syst. 15(3), 253–285 (1997) CrossRefGoogle Scholar
  13. 13.
    Foss, S., Korshunov, D.: Heavy tails in multi-server queue. Queueing Syst. 52(1), 31–48 (2006) CrossRefGoogle Scholar
  14. 14.
    Gans, N., Koole, G., Mandelbaum, A.: Telephone call centers: tutorial, review, and research prospects. Manuf. Serv. Oper. Manag. 5, 79–141 (2003) CrossRefGoogle Scholar
  15. 15.
    Gupta, V., Dai, J.G., Harchol-Balter, M., Zwart, B.: The effect of higher moments of job size distribution on the performance of an M/G/K queueing system. Technical Report CMU-CS-08-106, School of Computer Science, Carnegie Mellon University (2008) Google Scholar
  16. 16.
    Gupta, V., Harchol-Balter, M., Scheller-Wolf, A., Yechiali, U.: Fundamental characteristics of queues with fluctuating load. In: Proceedings of ACM SIGMETRICS, pp. 203–215 (2006) Google Scholar
  17. 17.
    Harchol-Balter, M., Schroeder, B.: Evaluation of task assignment policies for supercomputing servers. In: Proceedings of 9th IEEE Symposium on High Performance Distributed Computing (HPDC ’00), (2001) Google Scholar
  18. 18.
    van Hoorn, M.H., Tijms, H.C., Federgruen, A.: Approximations for the steady-state probabilities in the M/G/c queue. Adv. Appl. Probab. 13, 186–206 (1981) CrossRefGoogle Scholar
  19. 19.
    Hokstad, P.: Approximations for the M/G/m queue. Oper. Res. 26(3), 510–523 (1978) CrossRefGoogle Scholar
  20. 20.
    Hokstad, P.: The steady state solution of the M/K 2/m queue. Adv. Appl. Probab. 12(3), 799–823 (1980) CrossRefGoogle Scholar
  21. 21.
    Kimura, T.: Diffusion approximation for an M/G/m queue. Oper. Res. 31, 304–321 (1983) CrossRefGoogle Scholar
  22. 22.
    Kimura, T.: Approximations for multi-server queues: system interpolations. Queueing Syst. 17(3–4), 347–382 (1994) CrossRefGoogle Scholar
  23. 23.
    Kingman, J.F.C.: Inequalities in the theory of queues. J. R. Stat. Soc. 32(1), 102–110 (1970) Google Scholar
  24. 24.
    Kleinrock, L.: Queueing Systems, Volume I: Theory. Wiley-Interscience, New York (1975) Google Scholar
  25. 25.
    Köllerström, J.: Heavy traffic theory for queues with several servers. I. J. Appl. Probab. 11, 544–552 (1974) CrossRefGoogle Scholar
  26. 26.
    Lee, A.M., Longton, P.A.: Queueing process associated with airline passenger check-in. Oper. Res. Q. 10, 56–71 (1959) CrossRefGoogle Scholar
  27. 27.
    Lee, H.L., Cohen, M.A.: A note on the convexity of performance measures of M/M/c queueing systems. J. Appl. Probab. 20(4), 920–923 (1983) CrossRefGoogle Scholar
  28. 28.
    Ma, B.N.W., Mark, J.W.: Approximation of the mean queue length of an M/G/c queueing system. Oper. Res. 43(1), 158–165 (1995) CrossRefGoogle Scholar
  29. 29.
    Miyazawa, M.: Approximation of the queue-length distribution of an M/GI/s queue by the basic equations. J. Appl. Probab. 23, 443–458 (1986) CrossRefGoogle Scholar
  30. 30.
    Müller, A., Stoyan, D.: Comparison Methods for Stochastic Models and Risks. Wiley Series in Probability and Statistics. Wiley, Chichester (2002) Google Scholar
  31. 31.
    Nozaki, S.A., Ross, S.M.: Approximations in finite-capacity multi-server queues with Poisson arrivals. J. Appl. Probab. 15(4), 826–834 (1978) CrossRefGoogle Scholar
  32. 32.
    Boxma, J.C.O., Huffels, N.: Approximations in the mean waiting time in an M/G/s queueing system. Oper. Res. 27, 1115–1127 (1979) CrossRefGoogle Scholar
  33. 33.
    Ross, S.M.: Stochastic Processes, 2nd edn. Wiley, New York (1996) Google Scholar
  34. 34.
    Scheller-Wolf, A., Sigman, K.: New bounds for expected delay in FIFO GI/GI/c queues. Queueing Syst. 26(1–2), 169–186 (1997) CrossRefGoogle Scholar
  35. 35.
    Scheller-Wolf, A., Vesilo, R.: Structural interpretation and derivation of necessary and sufficient conditions for delay moments in FIFO multiserver queues. Queueing Syst. 54(3), 221–232 (2006) CrossRefGoogle Scholar
  36. 36.
    Stoyan, D.: Approximations for M/G/s queues. Math. Operationsforsch. Stat. Ser. Optim. 7, 587–594 (1976) Google Scholar
  37. 37.
    Stoyan, D.: A continuity theorem for queue size. Bull. Acad. Sci. Polon. 21, 1143–1146 (1973) Google Scholar
  38. 38.
    Stoyan, D.: Comparison Methods for Queues and Other Stochastic Models. Wiley Series in Probability and Mathematical Statistics: Applied Probability and Statistics. Wiley, Chichester (1983). Translation from the German edited by Daryl J. Daley Google Scholar
  39. 39.
    Takagi, H.: Queueing Analysis, Vol. 1: Vacation and Priority Systems. North-Holland, Amsterdam (1991) Google Scholar
  40. 40.
    Takahashi, Y.: An approximation formula for the mean waiting time of an M/G/c queue. J. Oper. Res. Soc. Jpn. 20, 147–157 (1977) Google Scholar
  41. 41.
    Thorisson, H.: The queue GI/GI/k: finite moments of the cycle variables and uniform rates of convergence. Commun. Stat. Stoch. Models 1(2), 221–238 (1985) CrossRefGoogle Scholar
  42. 42.
    van Doorn, E.A., Regterschot, J.K.: Conditional PASTA. Oper. Res. Lett. 7, 229–232 (1988) CrossRefGoogle Scholar
  43. 43.
    Whitt, W.: A diffusion approximation for the G/GI/n/m queue. Oper. Res. 52, 922–941 (2004) CrossRefGoogle Scholar
  44. 44.
    Whitt, W.: The effect of variability in the GI/G/s queue. J. Appl. Probab. 17, 1062–1071 (1980) CrossRefGoogle Scholar
  45. 45.
    Whitt, W.: Comparison conjectures about the M/G/s queue. OR Lett. 2(5), 203–209 (1983) Google Scholar
  46. 46.
    Whitt, W.: On approximations for queues, I: Extremal distributions. AT&T Bell Lab. Tech. J. 63, 115–138 (1984) Google Scholar
  47. 47.
    Whitt, W.: Approximations for the GI/G/m queue. Prod. Oper. Manag. 2(2), 114–161 (1993) Google Scholar
  48. 48.
    Yao, D.D.: Refining the diffusion approximation for the M/G/m queue. Oper. Res. 33, 1266–1277 (1985) CrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2009

Authors and Affiliations

  • Varun Gupta
    • 1
    Email author
  • Mor Harchol-Balter
    • 1
  • J. G. Dai
    • 2
  • Bert Zwart
    • 3
  1. 1.Computer Science DepartmentCarnegie Mellon UniversityPittsburghUSA
  2. 2.School of Industrial and Systems EngineeringGeorgia Institute of TechnologyAtlantaUSA
  3. 3.CWIAmsterdamThe Netherlands

Personalised recommendations