Demetres D. Kouvatsos
Abstract
A new “hybrid” analytic framework, based on the principle of maximum entropy, is used to approximate the queue length distribution of a G/G/1 finite buffer queue. Robust recursive relations are derived and asymptotic connections to the infinite capacity queue are established. Furthermore, “equivalence” principles are applied to analyse two-stage cyclic queues with general service times and favourable comparisons with global balance solutions are made. Numerical examples provide useful information on how critically system behaviour is affected by the distributional form of interarrival and service patterns. It is shown that the maximum entropy solution predicts the bottleneck “anomaly” and also it defines bounds on system performance. Comments on the implication of the work to the analysis and aggregation of computer systems are included.
- 1 Mitrani, I., Nonpriority multiprogramming system under heavy demand conditions - customer's viewpoint, JACM, Vol. 19, pp.445-452, 1972. Google Scholar
Digital Library
- 2 Reiser, M. and Kobayashi, H., The effects of service time distributions on system performance, Proc. IFIP Congress 74, Stockholm, Amsterdam, pp.230-234, 1974.Google Scholar
- 3 Sauer, C.H. and Chandy, K.M., Computer systems performance modelling, Prentice-Hall Inc., Englewood Cliffs, New Jersey, 1981.Google Scholar
- 4 Courtois, P.J., Decomposability: queueing and computer system applications, Academic Press Inc., New York, 1977.Google Scholar
- 5 Denning, P.J. and Buzen, J.P., The operational analysis of queueing network models, Computing Surveys, Vol. iO, No. 3, pp.225-261# 1978. Google ScholarDigital Library
- 6 Lavenberg, S.8., The steady-state queueing time distribution for the M/G/I finite capacity queue, Manag. Sci., Vol. 21, No. 5, pp.501-506, 1975.Google ScholarCross Ref
- 7 Courtois, P.J., The M/G/I finite capacity queue with delays, IEEE Trans. on Communications, Vol. Com-28, No. 2, 1980.Google Scholar
- 8 Ei-Affendi, M.A. and Kouvatsos, D.D., A maximum entropy analysis of the M/G/I and G/M/I queueing systems at equilibrium, Acta Info., Vol. 19, pp.339-855, 1983.Google ScholarCross Ref
- 9 Gelenbe, E. and Pulolle, G., The behaviour of a single queue in a general queueing network, Acta Info., Vol. 7, pp.128-160, 1976.Google Scholar
- 10 Buzen, J.P., Operational analysis- An alternative to stochastic modelling, in Performance of Computer Installations, D. Ferrari (Ed.), North-Holland Publishing Co., pp.175-194, 1978.Google Scholar
- 11 Buzen, J.P. and Denning, P.J., Measuring and calculating queue length distributions, Computer IEEE, pp.33-44, 1980.Google ScholarDigital Library
- 12 Jaynes, E.T., Prior probabilities, IEEE Trans. Syst. Sci. and Cyber., Vol. SSC-4, No. 4, pp.227-241, 1968.Google ScholarCross Ref
- 13 Shore, J.E. and Johnson, R.W., Axiomatic derivation of the principle of maximum entropy and the principle of minimum cross-entropy, IEEE Trans. on Info. Theory, Vol. IT-26, No. l, pp.26-37, 1980.Google ScholarDigital Library
- 14 Ferninand, A.E., A statistical mechanical approach to systems analysis, IBM J. Res. Develop., Vol. 14, pp.5#9-547, 1979.Google Scholar
- 15 Shore, J.E., Information theoretic approximations for M/G/I and G/G/I queueing systems, Acta Info., Vol. 17, pp.43-61, 1982.Google ScholarDigital Library
- 16 Bard, Y., A model of shared DASD and multipathing, CACM, Vol. 23, pp.56#-572, 1980. Google ScholarDigital Library
- 17 Cantor, J.L., Information theoretic analysis for a multi-server queueing system of equilibrium with applications to queues in tandem, M.Sc. Thesis, Electrical Engineering Dept., University of Maryland, U.S.A., April 198#.Google Scholar
- 18 Kouvatsos, D.D., Maximum entropy methods for general queueing networks, in Modelling Techniques and Tools for Performance Analysis, D. Poitier (Ed.), pp.589-608, North-Holland, 1985.Google Scholar
- 19 Walstra, R.J., Nonexponential networks of queues- A maximum entropy analysis, ACM Sigmetrics, pp.27-37, 1985. Google ScholarDigital Library
- 20 Kouvatsos, D.D., A universal maximum entropy algorithm for the analysis of general closed networks, Proc. Int. Seminar on Computer Networking and Performance Evaluation, Tokyo, Sept. 1985.Google Scholar
- 21 Johnson, R.W., Determining probability distributions by maximum entropy and minimum cross-entropy, Proc. APL79, ACM 0-8971-005, pp.24-29, 1979. Google ScholarDigital Library
- 22 Veran, M. and Poitier, D., QNAP-2: A portable environment for queueing network modelling, Proc. Int. Conf. on Modelling Techniques and Tools for Pemfor#nance Analysis, INRIA, May 1984.Google Scholar
- 23 Kouvatsos, D.D. and Ei-Affendi, M.A., On the analysis of the G/G/I queue, submitted for publication.Google Scholar
- 24 WhitY, W., On approximations for queues, III" Mixtures of exponential distributions, AT & T BLTJ 63, Vol. i, pp.163-175, 1984.Google Scholar
- 25 Bondi, A.B., Modelling the effect of local area networks contention on the performance of host computers, Performance 84, Gelenbe, E. (Ed.), pp.303-318, Dec. 1984. Google ScholarDigital Library
- 26 Marie, R.A., An approximate analytical method for general queueing networks, IEEE Trans. on Soft. Eng., SE-5# Vol. 5# 1979.Google Scholar
- 27 Kouvatsos, D.D. and Almond, J. ForthcominK publication.Google Scholar
- 28 Suri, R. and Diehl, C.W., A new "building block" for performance evaluation of queueing networks with finite buffers, ACM O-89791-141-5/84/008/O134, pp.134#142, 1984.Google Scholar
Index Terms
- A maximum entropy queue length distribution for the G/G/1 finite capacity queue
Recommendations
A maximum entropy queue length distribution for the G/G/1 finite capacity queue
SIGMETRICS '86/PERFORMANCE '86: Proceedings of the 1986 ACM SIGMETRICS joint international conference on Computer performance modelling, measurement and evaluationA new “hybrid” analytic framework, based on the principle of maximum entropy, is used to approximate the queue length distribution of a G/G/1 finite buffer queue. Robust recursive relations are derived and asymptotic connections to the infinite capacity ...
Tail asymptotics for the queue length in an M/G/1 retrial queue
In this paper, we study the tail behavior of the stationary queue length of an M/G/1 retrial queue. We show that the subexponential tail of the stationary queue length of an M/G/1 retrial queue is determined by that of the corresponding M/G/1 queue, and ...
The M/G/1 Queue with Finite Workload Capacity
We consider the M/G/1 queueing system in which customers whose admission to the system would increase the workload beyond a prespecified finite capacity limit are not accepted. Various results on the distribution of the workload are derived; in ...
Comments