[go: up one dir, main page]
More Web Proxy on the site http://driver.im/ skip to main content
article

Insensitivity in processor-sharing networks

Published: 01 September 2002 Publication History

Abstract

We consider an open network of processor-sharing nodes with state-dependent service capacities, i.e., the speed of each node may depend on the number of customers at any node. We demonstrate that the stationary distribution of the network state is insensitive to the distribution of service times if and only if the service capacities are balanced, i.e., the considered network is a Whittle network. The stationary distribution then has a closed-form expression and the expected sojourn time of a customer at any node is proportional to its required quantity of service. These results are extended to the cases of closed networks and state-dependent arrival rates and routing. Two simple examples illustrate the practical interest of these results in the context of communication networks.

References

[1]
{1} F. Baccelli, P. Brémaud, Palm probability and stationary queues, Lect. Notes Stat. 41 (1987).
[2]
{2} A.D. Barbour, Generalized semi-Markov schemes and open queueing networks, J. Appl. Probab. 19 (1982) 469-474.
[3]
{3} F. Baskett, K.M. Chandy, R.R. Muntz, F.G. Palacios, Open, closed and mixed networks of queues with different classes of customers, J. Assoc. Comput. Mach. 22 (1975) 248-260.
[4]
{4} S. Ben Fredj, T. Bonald, A. Proutière, G. Regnié, J.W. Roberts, Statistical bandwidth sharing: a study of congestion at flow level, in: Proceedings of the ACM SIGCOMM Conference, 2001.
[5]
{5} D. Bertsekas, R. Gallager, Data Networks, Prentice-Hall, Englewood Cliffs, NJ, 1987.
[6]
{6} T. Bonald, L. Massouliè, Impact of fairness on internet performance, in: Proceedings of the ACM SIGMETRICS Conference, 2001.
[7]
{7} T. Bonald, A. Proutière, Insensitive bandwidth sharing, in: Proceedings of the IEEE GLOBECOM Conference, 2002.
[8]
{8} T. Bonald, A. Proutière, G. Régnié, J.W. Roberts, Insensitivity results in statistical bandwidth sharing, in: Proceedings of the ITC 17 Conference, Elsevier, Amsterdam, 2001.
[9]
{9} D. Burman, Insensitivity in queueing systems, Adv. Appl. Probab. 13 (1981) 846-859.
[10]
{10} X. Chao, M. Miyazawa, R.F. Serfozo, H. Takada, Markov network processes with product form stationary distributions, Queueing Syst. 28 (1998) 377-401.
[11]
{11} J.W. Cohen, The multiple phase service network with generalized processor sharing, Acta Inform. 12 (1979) 245-284.
[12]
{12} W. Henderson, Insensitivity and reversed Markov processes, Adv. Appl. Probab. 15 (1983) 752-768.
[13]
{13} F.P. Kelly, Reversibility and Stochastic Networks, Wiley, New York, 1979.
[14]
{14} F.P. Kelly, Loss networks, Ann. Appl. Probab. 1 (1991) 319-378.
[15]
{15} L. Kleinrock, Queueing Systems, Vol. 2, Wiley, New York, 1975.
[16]
{16} L. Massoulié, J.W. Roberts, Bandwidth sharing and admission control for elastic traffic, Telecommun. Syst. 15 (2000) 185-201.
[17]
{17} M. Miyazawa, Insensitivity and product-form decomposability of reallocatable GSMP, Adv. Appl. Probab. 25 (1993) 415-437.
[18]
{18} M. Miyazawa, H. Takada, Traffic flows and product form solutions in stochastic transfer networks, Queueing Syst. 37 (2001) 199-232.
[19]
{19} S. Oueslati Boulahia, E. Oubagha, An approach to elastic flow routing, in: Proceedings of the ITC 16 Conference, Elsevier, Amsterdam, 1999.
[20]
{20} R. Schassberger, Insensitivity of steady state distributions of generalized semi-Markov processes with speeds, Adv. Appl. Probab. 10 (1978) 836-851.
[21]
{21} R. Schassberger, Two remarks on insensitive stochastic models, Adv. Appl. Probab. 18 (1986) 791-814.
[22]
{22} R.F. Serfozo, Markovian network processes: congestion-dependent routing and processing, Queueing Syst. 5 (1989) 5-36.
[23]
{23} R.F. Serfozo, Queueing networks with dependent nodes and concurrent movements, Queueing Syst. 13 (1993) 143-182.
[24]
{24} R.F. Serfozo, Introduction to Stochastic Networks, Springer, Berlin, 1999.
[25]
{25} P. Whittle, Partial balance and insensitivity, J. Appl. Probab. 22 (1985) 168-176.

Cited By

View all
  • (2020)Simple Near-Optimal Scheduling for the M/G/1Proceedings of the ACM on Measurement and Analysis of Computing Systems10.1145/33794774:1(1-29)Online publication date: 5-Jun-2020
  • (2019)The Mean-field Behavior of Processor Sharing Systems with General Job Lengths Under the SQ(d) PolicyACM SIGMETRICS Performance Evaluation Review10.1145/3308897.330892446:3(54-55)Online publication date: 25-Jan-2019
  • (2018)Optimal heavy-traffic queue length scaling in an incompletely saturated switchQueueing Systems: Theory and Applications10.1007/s11134-017-9562-x88:3-4(279-309)Online publication date: 1-Apr-2018
  • Show More Cited By

Recommendations

Reviews

James Speybroeck

An excellent example of scholarship in the affirmation of techniques addressing insensitivity in processor-sharing networks is presented in this paper. The authors clearly state the purpose of the paper: to show that the stationary distribution of the network state is insensitive to the distribution of service times, if and only if the service capabilities are balanced. After this introduction, the authors painstakingly present their research. The model and the balance property that characterize whittle networks are described in the first part of the paper. Insensitivity results and performance results are detailed. The paper ends with a presentation of two examples that illustrate the beneficial results of the research. The definitions and theorems presented are sophisticated. They deal with whittle networks, conditional expected sojourn time, closed networks, state-dependent arrival rates and routing, and application of the research to communication networks. This paper is certainly not bedtime reading, but if you are a professional concerned with processor-sharing networks, it will be of significant interest. Online Computing Reviews Service

Access critical reviews of Computing literature here

Become a reviewer for Computing Reviews.

Comments

Please enable JavaScript to view thecomments powered by Disqus.

Information & Contributors

Information

Published In

cover image Performance Evaluation
Performance Evaluation  Volume 49, Issue 1-4
September 2002
503 pages

Publisher

Elsevier Science Publishers B. V.

Netherlands

Publication History

Published: 01 September 2002

Author Tags

  1. Whittle networks
  2. balance
  3. insensitivity
  4. processor-sharing queue

Qualifiers

  • Article

Contributors

Other Metrics

Bibliometrics & Citations

Bibliometrics

Article Metrics

  • Downloads (Last 12 months)0
  • Downloads (Last 6 weeks)0
Reflects downloads up to 16 Dec 2024

Other Metrics

Citations

Cited By

View all
  • (2020)Simple Near-Optimal Scheduling for the M/G/1Proceedings of the ACM on Measurement and Analysis of Computing Systems10.1145/33794774:1(1-29)Online publication date: 5-Jun-2020
  • (2019)The Mean-field Behavior of Processor Sharing Systems with General Job Lengths Under the SQ(d) PolicyACM SIGMETRICS Performance Evaluation Review10.1145/3308897.330892446:3(54-55)Online publication date: 25-Jan-2019
  • (2018)Optimal heavy-traffic queue length scaling in an incompletely saturated switchQueueing Systems: Theory and Applications10.1007/s11134-017-9562-x88:3-4(279-309)Online publication date: 1-Apr-2018
  • (2017)Proportional Switching in First-in, First-out NetworksOperations Research10.1287/opre.2016.156565:2(496-513)Online publication date: 1-Apr-2017
  • (2017)Towards Optimality in Parallel SchedulingProceedings of the ACM on Measurement and Analysis of Computing Systems10.1145/31544991:2(1-30)Online publication date: 19-Dec-2017
  • (2017)Whittle networks with resetsProceedings of the 11th EAI International Conference on Performance Evaluation Methodologies and Tools10.1145/3150928.3150940(28-35)Online publication date: 5-Dec-2017
  • (2017)Balanced fair resource sharing in computer clustersPerformance Evaluation10.1016/j.peva.2017.08.006116:C(70-83)Online publication date: 1-Nov-2017
  • (2017)Whittle-networks with signalsPerformance Evaluation10.1016/j.peva.2016.12.001109:C(8-22)Online publication date: 1-Mar-2017
  • (2016)Optimal Heavy-Traffic Queue Length Scaling in an Incompletely Saturated SwitchACM SIGMETRICS Performance Evaluation Review10.1145/2964791.290146644:1(13-24)Online publication date: 14-Jun-2016
  • (2016)Optimal Heavy-Traffic Queue Length Scaling in an Incompletely Saturated SwitchProceedings of the 2016 ACM SIGMETRICS International Conference on Measurement and Modeling of Computer Science10.1145/2896377.2901466(13-24)Online publication date: 14-Jun-2016
  • Show More Cited By

View Options

View options

Media

Figures

Other

Tables

Share

Share

Share this Publication link

Share on social media