Hostname: page-component-78c5997874-8bhkd Total loading time: 0 Render date: 2024-11-03T01:56:01.215Z Has data issue: false hasContentIssue false

Using Distributed-Event Parallel Simulation to Study Departures from Many Queues in Series

Published online by Cambridge University Press:  27 July 2009

Albert G. Greenberg
Affiliation:
AT&T Bell Laboratories, Murray Hill, New Jersey 07974-0636
Otmar Schlunk
Affiliation:
Coordinated Science Laboratory, University of Illinois, Urbana, Illinois 61801
Ward Whitt
Affiliation:
AT&T Bell Laboratories, Murray Hill, New Jersey 07974-0636

Abstract

In this paper we describe an application of distributed-event parallel simulation to study the transient behavior of a large non-Markovian network of queues. In particular, we implemented the parallel-prefix-based algorithm of Greenberg, Lubachevsky, and Mitrani [13,14] on the 8,192-processor CM-2 Connection machine and the 16,384-processor MasPar computer to simulate the departure times D(k, n) of the kth customer from the nth queue in a long series of single-server queues. Each queue has unlimited waiting space and uses the first-in first-out discipline. The service times of all the customers at all the queues are i.i.d. with a general distribution, and the system starts out with k customers in the first queue and all other queues empty. Glynn and Whitt [11] established limit theorems for this model, but very little could be said about the limits themselves. The simulation results presented here describe the limits and the quality of the approximations resulting from using the limits for finite k and n. Indeed, the simulations suggest interesting conjectures. For this model speeding up a single long run is far superior to independent replications, because very long runs are required to obtain unbiased estimates of the desired quantities and the variance of the estimator at the end of the run is small. The achieved simulation rate was about 17 billion service completions per hour, which is a speedup by about a factor of 100 compared to simulation on a conventional single-processor machine. This speedup contributed greatly to performing the desired experiments.

Type
Research Article
Copyright
Copyright © Cambridge University Press 1993

Access options

Get access to the full version of this content by using one of the access options below. (Log in options will check for institutional or personal access. Content may require purchase if you do not have access.)

References

1.Akl, S.G. (1989). The design and analysis of parallel algorithms. Englewood Cliffs, NJ: Prentice-Hall.Google Scholar
2.Apostolico, A., Atallah, M.J., Larmore, L.L., & McFaddin, H.S. (1988). Efficient parallel al-gorithms for string editing and related problems. The Society for Computer Simulation, University of California, Irvine, pp. 5357.Google Scholar
3.Baccelli, F. & Canales, M. (1991). Parallel simulation of stochastic petri nets using recursive equations. INRI A Rapport de Recherche 1520.Google Scholar
4.Blank, T. (1990). The MasPar MP-1 architecture. Campcon, Spring 1990. San Francisco: IEEE Computer Society Press, pp. 2024.Google Scholar
5.Daley, D.J. & Vere-Jones, D. (1988). An introduction to the theory of point processes. New York: Springer-Verlag.Google Scholar
6.Feder, T., Greenberg, A.G., Ramachandran, V., Rausch, M., & Wang, L.-C. (1992). The telephone connect problem (preprint).Google Scholar
7.Frieze, A. (1991). On the length of the longest monotone subsequence in a random permutation. Annals of Applied Probability 1: 301305.CrossRefGoogle Scholar
8.Fujimoto, R.M. (1991). Parallel discrete event simulation. Communications of the ACM 33: 3153.Google Scholar
9.Glynn, P.W. & Heidelberger, P. (1990). Analysis of parallel, replicated simulations under a completion time constraint. IBM Research Report RC 15466 (#68774).Google Scholar
10.Glynn, P.W. & Heidelberger, P. (1990). Experiments with initial transient deletion for paral lel, replicated steady-state simulations. IBM Research Report RC 15700 (#70118).Google Scholar
11.Glynn, P.W. & Whitt, W. (1991). Departures from many queues in series. Annals of Applied Probability 1: 546572.CrossRefGoogle Scholar
12.Glynn, P.W. & Whitt, W. (1992). The asymptotic efficiency of simulation estimators. Operations Research 40: 505520.CrossRefGoogle Scholar
13.Greenberg, A.G., Lubachevsky, B.D., & Mitrani, I. (1990). Unboundedly parallel simulations via recurrence relations. 1990 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, pp. 112.CrossRefGoogle Scholar
14.Greenberg, A.G., Lubachevsky, B.D., & Mitrani, I. (1991). Algorithms for unboundedly parallel simulations. ACM Transactions on Computer Systems 9: 201221.CrossRefGoogle Scholar
15.Harrison, J.M. (1978). The diffusion approximation for tandem queues in heavy traffic. Advances in Applied Probability 10: 886905.CrossRefGoogle Scholar
16.Harrison, J.M. & Reiman, M.I. (1981). On the distribution of multidimensional reflected Brownian motion. SIAM Journal on Applied Mathematics 41: 345361.CrossRefGoogle Scholar
17.Harrison, J.M. & Reiman, M.I. (1981). Reflected Brownian motion on an orthant. Annals of Probability 9: 302308.CrossRefGoogle Scholar
18.Harrison, J.M. & Williams, R.J. (1987). Brownian models of open queueing networks with homogeneous customer populations. Stochastics 22:77115.CrossRefGoogle Scholar
19.Heidelberger, P. (1986). Discrete event simulations and parallel processing: Statistical properties. SIAM Journal on Scientific and Statistical Computing 9: 114132.Google Scholar
20.Heidelberger, P. (1986). Statistical analysis of parallel simulations. In Wilson, J. & Henriksen, J. (eds.), Proceedings of the 1986 Winter Simulation Conference. New York: IEEE Press, pp. 290295.Google Scholar
21.Heidelberger, P. & Stone, H. (1990). Parallel trace-driven cache simulation by time partitioning. In Baki, O., Sadowski, R.P., & Nance, R.E. (eds.), Proceedings of the 1990 Winter Simulation Conference. New York: IEEE Press, pp. 734737.CrossRefGoogle Scholar
22.Iglehart, D.L. & Whitt, W. (1970). Multiple channel queues in heavy traffic. 11: Sequences, networks and batches. Advances in Applied Probability 2: 355369.CrossRefGoogle Scholar
23.Kruskal, C.P., Rudolph, L., & Snir, M. (1985). The power of parallel prefix. IEEE Transactions on Computers, C-34, 10: 965968.CrossRefGoogle Scholar
24.Ladner, R.E. & Fischer, M.J. (1980). Parallel prefix computation. Journal of the ACM 27: 831838.CrossRefGoogle Scholar
25.Leighton, F.T. (1992). Introduction to parallel algorithms and architectures: Arrays, trees, hypercubes. San Mateo, CA: Morgan Kaufman.Google Scholar
26.Nicol, D., Greenberg, A.G., & Lubachevsky, B.D. (1992). Massively parallel algorithms for trace-driven cache simulations. Workshop on Parallel and Distributed Simulation.Google Scholar
27.Reiman, M.I. (1984). Open queueing networks in heavy traffic. Mathematics of Operations Research 9: 441458.CrossRefGoogle Scholar
28.Srinivasan, R. (1993). Queues in series via interacting particle systems. Mathematics of Operations Research 18: 3950.CrossRefGoogle Scholar
29.Stoyan, D. (1983). Comparison methods for queues and other stochastic models. Chichester: Wiley.Google Scholar
30.Suresh, S. & Whitt, W. (1990). The heavy-traffic bottleneck phenomenon in open queueing networks. Operations Research Letters 9: 335362.CrossRefGoogle Scholar
31.Thinking Machines. (1990). Connection machine model CM-2 technical summary. Thinking Machines Corp., Cambridge MA.Google Scholar
32.Wagner, R.A. & Fischer, M.J. (1973). The string to string editing problem. Journal of the ACM 21: 168173.CrossRefGoogle Scholar
33.Whitt, W. (1972). On the quality of Poisson approximations. Zeitschrift für Wahrscheinlich keitstheorie und verwandte Gebiete 28: 2336.CrossRefGoogle Scholar
34.Whitt, W. (1989). Planning queueing simulations. Management Science 35: 13411366.CrossRefGoogle Scholar
35.Whitt, W. (1991). The efficiency of one long run versus independent replications in steady-state simulation. Management Science 37: 645666.CrossRefGoogle Scholar