Hostname: page-component-89b8bd64d-rbxfs Total loading time: 0 Render date: 2026-05-12T15:13:51.369Z Has data issue: false hasContentIssue false
  • English
  • Français

Robust scheduling for flexible processing networks

Part of: Stochastic processes Operations research and management science Special processes

Published online by Cambridge University Press:  26 June 2017

Ramtin Pedarsani ,
Jean Walrand  and
Yuan Zhong
Ramtin Pedarsani*
Affiliation:
University of California, Berkeley
Jean Walrand*
Affiliation:
University of California, Berkeley
Yuan Zhong*
Affiliation:
Columbia University
*
*Current address: Department of Electrical and Computer Engineering, University of California Santa Barbara, Santa Barbara CA 93106, USA. Email address: ramtin@ece.ucsb.edu
** Current address: Department of Electrical Engineering and Computer Sciences, University of California Berkeley, 257 Cory Hall, Berkeley CA 94720, USA. Email address: walrand@berkeley.edu
*** Postal address: University of Chicago Booth School of Business, 360 Harper Center, 5807 South Woodlawn Avenue, Chicago, IL 60637, USA. Email address: yuan.zhong@chicagobooth.edu
Get access Rights & Permissions [Opens in a new window]

Abstract

Modern processing networks often consist of heterogeneous servers with widely varying capabilities, and process job flows with complex structure and requirements. A major challenge in designing efficient scheduling policies in these networks is the lack of reliable estimates of system parameters, and an attractive approach for addressing this challenge is to design robust policies, i.e. policies that do not use system parameters such as arrival and/or service rates for making scheduling decisions. In this paper we propose a general framework for the design of robust policies. The main technical novelty is the use of a stochastic gradient projection method that reacts to queue-length changes in order to find a balanced allocation of service resources to incoming tasks. We illustrate our approach on two broad classes of processing systems, namely the flexible fork-join networks and the flexible queueing networks, and prove the rate stability of our proposed policies for these networks under nonrestrictive assumptions.

Keywords

Robust schedulingflexible queueing networkstochastic gradient projection

MSC classification

Primary: 60K25: Queueing theory
Secondary: 90B15: Network models, stochastic 60G17: Sample path properties

Information

Type
Research Article
Information
Advances in Applied Probability , Volume 49 , Issue 2 , June 2017 , pp. 603 - 628
Copyright
Copyright © Applied Probability Trust 2017 

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.)

Article purchase

Temporarily unavailable

References

[1] Andradóttir, S., Ayhan, H. and Down, D. G. (2003). Dynamic server allocation for queueing networks with flexible servers. Operat. Res. 51, 952–968. CrossRefGoogle Scholar
[2] Atar, R., Mandelbaum, A. and Zviran, A. (2012). Control of fork-join networks in heavy traffic. In 50th Ann. Allerton Conf. on Communication, Control, and Computing, IEEE, pp. 823–830. Google Scholar
[3] Baccelli, F. and Makowski, A. (1985). Simple computable bounds for the fork-join queue. In Proc. 19th Ann. Conf. on Information Sciences and Systems, John Hopkins University, Baltimore, MD, pp. 436–441. Google Scholar
[4] Baccelli, F., Makowski, A. M. and Shwartz, A. (1989). The fork-join queue and related systems with synchronization constraints: stochastic ordering and computable bounds. Adv. Appl. Prob. 21, 629–660. CrossRefGoogle Scholar
[5] Baharian, G. and Tezcan, T. (2011). Stability analysis of parallel server systems under longest queue first. Math. Meth. Operat. Res. 74, 257–279. Google Scholar
[6] Bambos, N. and Walrand, J. (1991). On stability and performance of parallel processing systems. J. Assoc. Comput. Mach. 38, 429–452. CrossRefGoogle Scholar
[7] Borkar, V. S. (2008). Stochastic Approximation. Cambridge University Press. Google Scholar
[8] Boyd, S. and Vandenberghe, L. (2004). Convex Optimization. Cambridge University Press. Google Scholar
[9] Chow, Y. S. and Teicher, H. (1997). Probability Theory, 3rd edn. Springer, New York. CrossRefGoogle Scholar
[10] Dai, J. G. (1995). On positive Harris recurrence of multiclass queueing networks: a unified approach via fluid limit models. Ann. Appl. Prob. 5, 49–77. Google Scholar
[11] Dai, J. G. (1999). Stability of fluid and stochastic processing networks. MaPhySto Miscellanea Publication 9, University of Aarhus. Google Scholar
[12] Dean, J. and Ghemawat, S. (2008). Mapreduce: simplified data processing on large clusters. Commun. Assoc. Comput. Mach. 51, 107–113. Google Scholar
[13] Dimakis, A. and Walrand, J. (2006). Sufficient conditions for stability of longest-queue-first scheduling: second-order properties using fluid limits. Adv. Appl. Prob. 38, 505–521. Google Scholar
[14] Durrett, R. (2010). Probability: Theory and Examples (Camb. Ser. Statist. Prob. Math. 31), 4th edn. Cambridge University Press. Google Scholar
[15] Harrison, J. M. (2000). Brownian models of open processing networks: canonical representation of workload. Ann. Appl. Prob. 10, 75–103. Google Scholar
[16] Harrison, J. M. and Nguyen, V. (1993). Brownian models of multiclass queueing networks: current status and open problems. Queueing Systems 13, 5–40. Google Scholar
[17] Jiang, L. and Walrand, J. (2010). A distributed CSMA algorithm for throughput and utility maximization in wireless networks. IEEE/ACM Trans. Networking 18, 960–972. Google Scholar
[18] Kandula, S. et al. (2009). The nature of data center traffic: measurements and analysis. In Proc. 9th ACM SIGCOMM Conf. Internet Measurement Conference, ACM, New York, pp. 202–208. Google Scholar
[19] Konstantopoulos, P. and Walrand, J. (1989). Stationarity and stability of fork-join networks. J. Appl. Prob. 26, 604–614. Google Scholar
[20] Mandelbaum, A. and Stolyar, A. L. (2004). Scheduling flexible servers with convex delay costs: heavy-traffic optimality of the generalized cμ-rule. Operat. Res. 52, 836–855. CrossRefGoogle Scholar
[21] Nguyen, V. (1993). Processing networks with parallel and sequential tasks: heavy traffic analysis and Brownian limits. Ann. Appl. Prob. 3, 28–55. CrossRefGoogle Scholar
[22] Nguyen, V. (1994). The trouble with diversity: fork-join networks with heterogeneous customer population. Ann. Appl. Prob. 4, 1–25. Google Scholar
[23] Padala, P. et al. (2009). Automated control of multiple virtualized resources. In Proc. 4th ACM Europ. Conf. on Computer Systems, ACM, New York, pp. 13–26. Google Scholar
[24] Pedarsani, R., Walrand, J. and Zhong, Y. (2014). Robust scheduling in a flexible fork-join network. In IEEE 53rd Ann. Conf. on Decision and Control, IEEE, pp. 3669–3676. Google Scholar
[25] Pedarsani, R., Walrand, J. and Zhong, Y. (2014). Scheduling tasks with precedence constraints on multiple servers. In 52nd Ann. Allerton Conf. on Communication, Control, and Computing, IEEE, pp. 1196–1203. Google Scholar
[26] Pedarsani, R., Walrand, J. and Zhong, Y. (2016). Robust scheduling for flexible processing networks. Preprint. Available at https://arXiv.org/abs/1610.03803vl. Google Scholar
[27] Stolyar, A. L. and Yudovina, E. (2012). Tightness of invariant distributions of a large-scale flexible service system under a priority discipline. Stoch. Systems 2, 381–408. Google Scholar