Skip to main content

Advertisement

Log in

Periodic load balancing

  • Published:
Queueing Systems Aims and scope Submit manuscript

Abstract

Multiprocessor load balancing aims to improve performance by moving jobs from highly loaded processors to more lightly loaded processors. Some schemes allow only migration of new jobs upon arrival, while other schemes allow migration of jobs in progress. A difficulty with all these schemes, however, is that they require continuously maintaining detailed state information. In this paper we consider the alternative of periodic load balancing, in which the loads are balanced only at each T time units for some appropriate T. With periodic load balancing, state information is only needed at the balancing times. Moreover, it is often possible to use slightly stale information collected during the interval between balancing times. In this paper we study the performance of periodic load balancing. We consider multiple queues in parallel with unlimited waiting space to which jobs come either in separate independent streams or by assignment (either random or cyclic) from a single stream. Resource sharing is achieved by periodically redistributing the jobs or the work in the system among the queues. The performance of these systems of queues coupled by periodic load balancing depends on the transient behavior of a single queue. We focus on useful approximations obtained by considering a large number of homogeneous queues and a heavy load. When the number of queues is sufficiently large, the number of jobs or quantity of work at each queue immediately after redistribution tends to evolve deterministically, by the law of large numbers. The steady-state (limiting) value of this deterministic sequence is obtained as the solution of a fixed point equation, where the initial value is equal to the expected transient value over the interval between successive redistributions conditional on the initial value. A refined approximation based on the central limit theorem is a normal distribution, where the mean and variance are obtained by solving a pair of fixed-point equations. With higher loads, which is natural to consider when load balancing is performed, a heavy-traffic limit theorem shows that one-dimensional reflected Brownian motion can be used to approximately describe system performance, even with general arrival and service processes. With these approximations, we show how performance depends on the assumed arrival pattern of jobs and the model parameters. We do numerical calculations and conduct simulation experiments to show the accuracy of the approximations.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Subscribe and save

Springer+ Basic
$34.99 /Month
  • Get 10 units per month
  • Download Article/Chapter or eBook
  • 1 Unit = 1 Article or 1 Chapter
  • Cancel anytime
Subscribe now

Buy Now

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

References

  1. J. Abate and W. Whitt, Transient behavior of regulated Brownian motion, I and II, Adv. in Appl. Probab. 19 (1987) 560-631.

    Article  Google Scholar 

  2. J. Abate and W. Whitt, Transient behavior of the M/M/1 queue via Laplace transforms, Adv. in Appl. Probab. 20 (1988) 145-178.

    Article  Google Scholar 

  3. J. Abate and W. Whitt, Calculating time-dependent performance measures for the M/M/1 queue, IEEE Trans. Commun. 37 (1989) 1102-1104.

    Article  Google Scholar 

  4. J. Abate and W. Whitt, The Fourier-series method for inverting transforms of probability distributions, Queueing Systems 10 (1992) 5-88.

    Article  Google Scholar 

  5. J. Abate and W. Whitt, Numerical inversion of Laplace transforms of probability distributions, ORSA J. Comput. 7 (1995) 36-43.

    Google Scholar 

  6. M. Abramowitz and I.A. Stegun, Handbook of Mathematical Functions (Dover, New York, 1972).

    Google Scholar 

  7. F. Baccelli and P. Brémaud, Elements of Queueing Theory (Springer, New York, 1994).

    Google Scholar 

  8. A. Barak, G. Shai and R.G. Wheeler, The MOSIX Distributed Operating System: Load Balancing for UNIX (Springer, Berlin, 1993).

    Google Scholar 

  9. P. Billingsley, Convergence of Probability Measures (Wiley, New York, 1968).

    Google Scholar 

  10. A.A. Borovkov, Stochastic Processes in Queueing Theory (Springer, New York, 1976).

    Google Scholar 

  11. G.L. Choudhury, D.M. Lucantoni and W. Whitt, Multidimensional transform inversion with applications to the transient M/G/1 queue, Ann. Appl. Probab. 4 (1994) 719-740.

    Google Scholar 

  12. E. Çinlar, Superpositions of point processes, in: Stochastic Point Processes: Statistical Analysis, Theory and Applications, ed. P.A.W. Lewis (Wiley, New York, 1972) pp. 549-606.

    Google Scholar 

  13. J.L. Davis, W.A. Massey and W. Whitt, Sensitivity to the service-time distribution in the nonstationary Erlang loss model, Managm. Sci. 41 (1995) 1107-1116.

    Google Scholar 

  14. D.L. Eager, E.D. Lazowska and J. Zahorjan, Adaptive load balancing in homogeneous distributed systems, IEEE Trans. Software Engrg. 12 (1986) 662-675.

    Google Scholar 

  15. S.N. Ethier and T.G. Kurtz, Characterization and Approximation of Markov Processes (Wiley, New York, 1986).

    Google Scholar 

  16. W. Feller, An Introduction to Probability Theory and Its Applications, Vol. I (Wiley, New York, 1968).

    Google Scholar 

  17. K.W. Fendick, V.R. Saksena and W. Whitt, Dependence in packet queues, IEEE Trans. Commun. 37 (1989) 1173-1183.

    Article  Google Scholar 

  18. G.J. Foschini, Unobtrusive communication of status in a packet network in heavy traffic, AT&T Tech. J. 64 (1985) 463-479.

    Google Scholar 

  19. G.J. Foschini and J. Salz, A basic dynamic routing problem and diffusion, IEEE Trans. Commun. 26 (1978) 320-327.

    Article  Google Scholar 

  20. B. Hajek, Performance of global load balancing by local adjustment, IEEE Trans. Inform. Theory 36 (1990) 1398-1414.

    Article  Google Scholar 

  21. M. Harchol-Balter and A.B. Downey, Exploiting process lifetime distributions for dynamic load balancing, in: Proc. SIGMETRICS '96 (1996).

  22. G. Hjálmtýsson, Lightweight call setup — supporting connection and connectionless services, in: Teletraffic Contributions for the Information Age, Proc. of the 15th Internat. Teletraffic Congress, eds. V. Ramaswami and P.E. Wirth (Elsevier, Amsterdam, 1997) pp. 35-45.

    Google Scholar 

  23. G. Hjálmtýsson and K.K. Ramakrishnan, UNITE — An architecture for lightweight signalling in ATM networks, in: IEEE Infocom '98 (1998) pp. 832-840.

  24. D.L. Iglehart and W. Whitt, Multiple channel queues in heavy traffic, I and II, Adv. in Appl. Probab. 2 (1970) 150-177 and 355–369.

    Article  Google Scholar 

  25. O. Kella and W. Whitt, Diffusion approximations for queues with server vacations, Adv. in Appl. Probab. 22 (1990) 706-729.

    Article  Google Scholar 

  26. J. Köllerström, Heavy traffic theory for queues with several servers, I, J. Appl. Probab. 11 (1974) 544-552.

    Article  Google Scholar 

  27. C.N. Laws, Resource pooling in queueing networks with dynamic routing, Adv. in Appl. Probab. 24 (1992) 699-726.

    Article  Google Scholar 

  28. W.E. Leland and T.J. Ott, Load balancing heuristics and process behavior, Sigmetrics 86(14) (1986) 54-69.

    Google Scholar 

  29. A. Mandelbaum and M.I. Reiman, On pooling in queueing networks, Managm. Sci., to appear.

  30. W.A. Massey and W. Whitt, Peak congestion in multi-server service systems with slowly varying arrival rates, Queueing Systems 25 (1997) 157-172.

    Article  Google Scholar 

  31. S.P. Meyn and R.L. Tweedie, Markov Chains and Stochastic Stability (Springer, New York, 1993).

    Google Scholar 

  32. M.I. Reiman, Some diffusion approximate with state space collapse, in: Modelling and Performance Evaluation Methodology, eds. F. Baccelli and G. Fayolle (Springer, Berlin, 1984) pp. 209-240.

    Google Scholar 

  33. M.H. Rothkopf and P. Rech, Perspectives on queues: Combining queues is not always beneficial, Oper. Res. 35 (1987) 906-909.

    Article  Google Scholar 

  34. H. Sakasegawa, An approximate formula Lq = αβρ/(1 — ρ), Ann. Inst. Statist. Math. 29 (1977) 67-75.

    Google Scholar 

  35. R.F. Serfozo, Partitions of point processes: multivariate Poisson approximations, Stochastic Process. Appl. 20 (1985) 281-294.

    Article  Google Scholar 

  36. D.R. Smith and W. Whitt, Resource sharing for efficiency in traffic systems, Bell System Tech. J. 60 (1981) 39-55.

    Google Scholar 

  37. D. Stoyan, Comparison Methods for Queues and Other Stochastic Models (Wiley, New York, 1983).

    Google Scholar 

  38. M.R. Taaffe and K.L. Ong, Approximating Ph(t)/M(t)/S/C queueing systems, Ann. Oper. Res. 8 (1987) 103-116.

    Article  Google Scholar 

  39. L. Takács, Introduction to the Theory of Queues (Oxford Univ. Press, New York, 1962).

    Google Scholar 

  40. E. van Doorn, Stochastic Monotonicity and Queueing Applications of Birth-Death Processes (Springer, New York, 1981).

    Google Scholar 

  41. R.W. Weber, On the optimal assignment of customers to parallel servers, J. Appl. Probab. 15 (1978) 406-413.

    Article  Google Scholar 

  42. W. Whitt, Weak convergence theorems for priority queues: Preemptive-resume discipline, J. Appl. Probab. 8 (1971) 74-94.

    Article  Google Scholar 

  43. W. Whitt, Some useful functions for functional limit theorems, Math. Oper. Res. 5 (1980) 67-85.

    Google Scholar 

  44. W. Whitt, Comparing counting processes and queues, Adv. in Appl. Probab. 13 (1981) 207-220.

    Article  Google Scholar 

  45. W. Whitt, Deciding which queue to join: Some counterexamples, Oper. Res. 34 (1986) 55-62.

    Google Scholar 

  46. W. Whitt, Planning queueing simulations, Managm. Sci. 35 (1989) 1341-1366.

    Google Scholar 

  47. W. Whitt, Understanding the efficiency of multi-server service systems, Managm. Sci. 38 (1992) 708-723.

    Google Scholar 

  48. W. Whitt, Approximations for the GI/G/m queue, Production Oper. Managm. 2 (1993) 114-160.

    Article  Google Scholar 

  49. M.H. Willebeck-LeMair and A.P. Reeves, Strategies for dynamic load balancing on highly parallel computers, IEEE Trans. Parallel Distrib. Systems 9 (1993) 979-993.

    Article  Google Scholar 

  50. W. Winston, Optimality of the shortest line discipline, J. Appl. Probab. 14 (1977) 181-189.

    Article  Google Scholar 

  51. R.W. Wolff, An upper bound for multi-channel queues, J. Appl. Probab. 14 (1977) 884-888.

    Article  Google Scholar 

  52. H. Zhang, G. Hsu and R. Wang, Heavy traffic limit theorems for a sequence of shortest queueing systems, Queueing Systems 21 (1995) 217-238.

    Article  Google Scholar 

  53. S. Zhou, A trace-driven simulation study of dynamic load balancing, IEEE Trans. Software Engrg. 14 (1988) 1327-1341.

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Rights and permissions

Reprints and permissions

About this article

Cite this article

Hjálmtýsson, G., Whitt, W. Periodic load balancing. Queueing Systems 30, 203–250 (1998). https://doi.org/10.1023/A:1019165004692

Download citation

  • Issue Date:

  • DOI: https://doi.org/10.1023/A:1019165004692

Navigation