Abstract
We consider the tail behavior of the response time distribution in an M/G/1 queue with heavy-tailed job sizes, specifically those with intermediately regularly varying tails. In this setting, the response time tail of many individual policies has been characterized, and it is known that policies such as Shortest Remaining Processing Time (SRPT) and Foreground-Background (FB) have response time tails of the same order as the job size tail, and thus such policies are tail-optimal. Our goal in this work is to move beyond individual policies and characterize the set of policies that are tail-optimal. Toward that end, we use the recently introduced SOAP framework to derive sufficient conditions on the form of prioritization used by a scheduling policy that ensure the policy is tail-optimal. These conditions are general and lead to new results for important policies that have previously resisted analysis, including the Gittins policy, which minimizes mean response time among policies that do not have access to job size information. As a by-product of our analysis, we derive a general upper bound for fractional moments of M/G/1 busy periods, which is of independent interest.
- S. Aalto, U. Ayesta, S. Borst, V. Misra, and R. Nú nez-Queija. 2007. Beyond processor sharing. ACM SIGMETRICS Performance Evaluation Review , Vol. 34 (2007), 36--43.Google Scholar
Digital Library
- S. Aalto, U. Ayesta, and R. Righter. 2009. On the Gittins index in the M/G/1 queue. Queueing Systems , Vol. 63 (2009), 437--458.Google Scholar
Digital Library
- E. Altman, K. Avrachenkov, and U. Ayesta. 2006. A survey on discriminatory processor sharing. Queueing Systems , Vol. 53 (2006), 53--63.Google Scholar
Digital Library
- N. Bansal, B. Kamphorst, and B. Zwart. 2018. Achievable performance of blind policies in heavy traffic. Mathematics of Operations Research , Vol. 43, 3 (2018), 949--964.Google Scholar
Cross Ref
- L. Becchetti and S. Leonardi. 2004. Nonclairvoyant scheduling to minimize the total flow time on single and parallel machines. Journal of the ACM (JACM) , Vol. 51, 4 (2004), 517--539.Google Scholar
Digital Library
- N. Bingham, C. Goldie, and J. Teugels. 1987. Regular Variation .Cambridge University Press.Google Scholar
- S. Borst, O. Boxma, R. Nú nez Queija, and B. Zwart. 2003. The impact of the service discipline on delay asymptotics. Performance Evaluation , Vol. 54 (2003), 175--206.Google Scholar
Digital Library
- S.C. Borst, R. Nú n ez-Queija, and B. Zwart. 2006. Sojourn-time asymptotics in processor-sharing queues. Queueing Systems 53 (2006), 31--51.Google Scholar
Digital Library
- O. Boxma and D. Denisov. 2011. Sojourn time tails in the single server queue with heavy-tailed service times. Queueing Systems , Vol. 69, 2 (2011), 101--119.Google Scholar
Digital Library
- O. Boxma and B. Zwart. 2007. Tails in scheduling. ACM SIGMETRICS Performance Evaluation Review , Vol. 34, 4 (2007), 13--20.Google Scholar
Digital Library
- J. Cohen. 1973. Some results on regular variation for distributions in queueing and fluctuation theory. Journal of Applied Probability , Vol. 10, 2 (1973), 343--353.Google Scholar
Cross Ref
- M. Crovella and A. Bestavros. 1996. Self-similarity in World Wide Web traffic: evidence and possible causes. Proceedings of ACM Sigmetrics '96 (1996), 160--169.Google Scholar
- S. Foss, D. Korshunov, and S. Zachary. 2013. An Introduction to Heavy-tailed and Subexponential Distributions .Springer, New York, NY.Google Scholar
- F. Guillemin, Ph. Robert, and B. Zwart. 2004. Tail asymptotics for processor-sharing queues. Advances in Applied Probability 36 (2004), 525--543.Google Scholar
Cross Ref
- W. Johnson. 2002. The curious history of Faà di Bruno's formula. The American Mathematical Monthly , Vol. 109, 3 (2002), 217--234.Google Scholar
- B. Kalyanasundaram and K. R. Pruhs. 1997. Minimizing flow time nonclairvoyantly. In Proceedings 38th Annual Symposium on Foundations of Computer Science. IEEE, 345--352.Google Scholar
- N. Morris, C. Stewart, L. Chen, R. Birke, and J. Kelley. 2018. Model-driven computational sprinting. In Proceedings of the Thirteenth EuroSys Conference . 1--13.Google Scholar
- J. Nair, A. Wierman, and B. Zwart. 2010. Tail-robust scheduling via limited processor sharing. Performance Evaluation , Vol. 67, 11 (2010), 978--995.Google Scholar
Digital Library
- R. Nú n ez-Queija. 2002. Queues with equally heavy sojourn time and service requirement distributions. Annals of Operations Research , Vol. 113, 1 (01 Jul 2002), 101--117. https://doi.org/10.1023/A:1020905810996Google Scholar
- M. Nuyens and A. Wierman. 2008. The foreground-background queue: a survey. Performance Evaluation , Vol. 65 (2008), 286--307.Google Scholar
Digital Library
- M. Nuyens, A. Wierman, and B. Zwart. 2008. Preventing large sojourn times using SMART scheduling. Operations Research , Vol. 56 (2008), 88--101.Google Scholar
Digital Library
- K. Park and W. Willinger. 2000. Self-similar Network Traffic and Performance Evaluation .Wiley.Google Scholar
- D. Peterson. 1996. Data center I/O patterns and power laws. CMG Proceedings (1996).Google Scholar
- A. Raghavan, Y. Luo, A. Chandawalla, M. Papaefthymiou, K. Pipe, T. Wenisch, and M. Martin. 2012. Computational sprinting. In Proceedings of the 18th Symposium on High Performance Computer Architecture. IEEE, 1--12.Google Scholar
- M. Remerova, S. Foss, and B. Zwart. 2014. Random fluid limits of an overloaded polling model. Advances in Applied Probability 46 (2014), 76--101.Google Scholar
Cross Ref
- L. Schrage and L. Miller. 1966. The queue M/G/1 with the shortest remaining processing time discipline. Operations Research , Vol. 14 (1966), 670--684.Google Scholar
Digital Library
- Z. Scully and M. Harchol-Balter. 2018. SOAP Bubbles: Robust Scheduling Under Adversarial Noise. In 2018 56th Annual Allerton Conference on Communication, Control, and Computing (Allerton) . 144--154. https://doi.org/10.1109/ALLERTON.2018.8635963Google Scholar
- Z. Scully, M. Harchol-Balter, and A. Scheller-Wolf. 2018. SOAP: one clean analysis of all age-based scheduling policies. Proceedings of the ACM on Measurement and Analysis of Computing Systems , Vol. 2, 1, Article 16 (April 2018), bibinfonumpages30 pages. https://doi.org/10.1145/3179419Google Scholar
- Z. Scully, M. Harchol-Balter, and A. Scheller-Wolf. 2020. Simple Near-Optimal Scheduling for the M/G/1. Proceedings of the ACM on Measurement and Analysis of Computing Systems , Vol. 4, 1, Article 11 (March 2020), bibinfonumpages29 pages. https://doi.org/10.1145/3379477Google Scholar
- A. Stolyar and K. Ramanan. 2001. Largest weighted delay first scheduling: large deviations and optimality. Annals of Applied Probability , Vol. 11 (2001), 1--48.Google Scholar
Cross Ref
- A. Wierman and B. Zwart. 2012. Is tail-optimal scheduling possible? Operations Research , Vol. 60, 5 (2012), 1249--1257.Google Scholar
Digital Library
- B. Zwart and O. Boxma. 2000. Sojourn time asymptotics in the M/G/1 processor sharing queue. Queueing Systems , Vol. 35, 1--4 (2000), 141--166.Google Scholar
Digital Library
Index Terms
Characterizing Policies with Optimal Response Time Tails under Heavy-Tailed Job Sizes
Recommendations
Simple Near-Optimal Scheduling for the M/G/1
SIGMETRICSWe consider the problem of preemptively scheduling jobs to minimize mean response time of an M/G/1 queue. When we know each job's size, the shortest remaining processing time (SRPT) policy is optimal. Unfortunately, in many settings we do not have ...
The Gittins Policy is Nearly Optimal in the M/G/k under Extremely General Conditions
POMACSThe Gittins scheduling policy minimizes the mean response in the single-server M/G/1 queue in a wide variety of settings. Most famously, Gittins is optimal when preemption is allowed and service requirements are unknown but drawn from a known ...
Characterizing Policies with Optimal Response Time Tails under Heavy-Tailed Job Sizes
SIGMETRICS '20: Abstracts of the 2020 SIGMETRICS/Performance Joint International Conference on Measurement and Modeling of Computer SystemsWe consider the tail behavior of the response time distribution in an M/G/1 queue with heavy-tailed job sizes, specifically those with intermediately regularly varying tails. In this setting, the response time tail of many individual policies has been ...






Comments